exploratory

.gitattributes

@@ -0,0 +1 @@
+cpp/* linguist-vendored

.github/FUNDING.yml

@@ -1,12 +0,0 @@
-# These are supported funding model platforms
-
-github: fulmicoton
-patreon: # Replace with a single Patreon username
-open_collective: # Replace with a single Open Collective username
-ko_fi: # Replace with a single Ko-fi username
-tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
-community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
-liberapay: # Replace with a single Liberapay username
-issuehunt: # Replace with a single IssueHunt username
-otechie: # Replace with a single Otechie username
-custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']

.github/ISSUE_TEMPLATE/actions.md

@@ -1,13 +0,0 @@
----
-name: Actions
-about: Actions not directly related to producing code.
-
----
-
-# Actions title
-
-Action description. 
-e.g. 
-- benchmark
-- investigate and report
-- etc.

.github/dependabot.yml

@@ -1,15 +0,0 @@
-version: 2
-updates:
-- package-ecosystem: cargo
-  directory: "/"
-  schedule:
-    interval: daily
-    time: "20:00"
-  open-pull-requests-limit: 10
-
-- package-ecosystem: "github-actions"
-  directory: "/"
-  schedule:
-    interval: daily
-    time: "20:00"
-  open-pull-requests-limit: 10

.github/workflows/coverage.yml

@@ -1,26 +0,0 @@
-name: Coverage
-
-on:
-  push:
-    branches: [ main ]
-  pull_request:
-    branches: [ main ]
-
-jobs:
-  coverage:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v3
-      - name: Install Rust
-        run: rustup toolchain install nightly --profile minimal --component llvm-tools-preview
-      - uses: Swatinem/rust-cache@v2
-      - uses: taiki-e/install-action@cargo-llvm-cov
-      - name: Generate code coverage
-        run: cargo +nightly llvm-cov --all-features --workspace --lcov --output-path lcov.info
-      - name: Upload coverage to Codecov
-        uses: codecov/codecov-action@v3
-        continue-on-error: true
-        with:
-          token: ${{ secrets.CODECOV_TOKEN }} # not required for public repos
-          files: lcov.info
-          fail_ci_if_error: true

.github/workflows/long_running.yml

@@ -1,28 +0,0 @@
-name: Long running tests
-
-on:
-  push:
-    branches: [ main ]
-
-env:
-  CARGO_TERM_COLOR: always
-  NUM_FUNCTIONAL_TEST_ITERATIONS: 20000
-
-jobs:
-  test:
-
-    runs-on: ubuntu-latest
-
-    steps:
-    - uses: actions/checkout@v3
-    - name: Install stable
-      uses: actions-rs/toolchain@v1
-      with:
-          toolchain: stable
-          profile: minimal
-          override: true
-
-    - name: Run indexing_unsorted
-      run: cargo test indexing_unsorted -- --ignored
-    - name: Run indexing_sorted
-      run: cargo test indexing_sorted -- --ignored

.github/workflows/test.yml

@@ -1,74 +0,0 @@
-name: Unit tests
-
-on:
-  push:
-    branches: [ main ]
-  pull_request:
-    branches: [ main ]
-
-env:
-  CARGO_TERM_COLOR: always
-
-jobs:
-  check:
-
-    runs-on: ubuntu-latest
-
-    steps:
-    - uses: actions/checkout@v3
-
-    - name: Install nightly
-      uses: actions-rs/toolchain@v1
-      with:
-            toolchain: nightly
-            profile: minimal
-            components: rustfmt
-    - name: Install stable
-      uses: actions-rs/toolchain@v1
-      with:
-            toolchain: stable
-            profile: minimal
-            components: clippy
-
-    - uses: Swatinem/rust-cache@v2
-
-    - name: Check Formatting
-      run: cargo +nightly fmt --all -- --check
-
-    - uses: actions-rs/clippy-check@v1
-      with:
-        toolchain: stable
-        token: ${{ secrets.GITHUB_TOKEN }}
-        args: --tests
-
-  test:
-
-    runs-on: ubuntu-latest
-
-    strategy:
-      matrix:
-        features: [
-            { label: "all", flags: "mmap,stopwords,brotli-compression,lz4-compression,snappy-compression,zstd-compression,failpoints" },
-            { label: "quickwit", flags: "mmap,quickwit,failpoints" }
-        ]
-
-    name: test-${{ matrix.features.label}}
-
-    steps:
-    - uses: actions/checkout@v3
-
-    - name: Install stable
-      uses: actions-rs/toolchain@v1
-      with:
-            toolchain: stable
-            profile: minimal
-            override: true
-
-    - uses: taiki-e/install-action@nextest
-    - uses: Swatinem/rust-cache@v2
-
-    - name: Run tests
-      run: cargo +stable nextest run --features ${{ matrix.features.flags }} --verbose --workspace
-
-    - name: Run doctests
-      run: cargo +stable test --doc --features ${{ matrix.features.flags }} --verbose --workspace

.gitignore

@@ -1,6 +1,4 @@
 tantivy.iml
-.cargo
-proptest-regressions
 *.swp
 target
 target/debug
@@ -9,7 +7,7 @@ target/release
 Cargo.lock
 benchmark
 .DS_Store
+cpp/simdcomp/bitpackingbenchmark
 *.bk
 .idea
 trace.dat
-cargo-timing*

.travis.yml

@@ -0,0 +1,85 @@
+# Based on the "trust" template v0.1.2
+# https://github.com/japaric/trust/tree/v0.1.2
+
+dist: trusty
+language: rust
+services: docker
+sudo: required
+
+env:
+  global:
+    - CRATE_NAME=tantivy
+    - TRAVIS_CARGO_NIGHTLY_FEATURE=""
+    - secure: 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
+
+addons:
+  apt:
+    sources:
+      - ubuntu-toolchain-r-test
+      - kalakris-cmake
+    packages:
+      - gcc-4.8
+      - g++-4.8
+      - libcurl4-openssl-dev
+      - libelf-dev
+      - libdw-dev
+      - binutils-dev
+      - cmake
+
+matrix:
+  include:
+    # Android
+    - env: TARGET=aarch64-linux-android DISABLE_TESTS=1
+    #- env: TARGET=arm-linux-androideabi DISABLE_TESTS=1
+    #- env: TARGET=armv7-linux-androideabi DISABLE_TESTS=1
+    #- env: TARGET=i686-linux-android DISABLE_TESTS=1
+    #- env: TARGET=x86_64-linux-android DISABLE_TESTS=1
+
+    # Linux
+    #- env: TARGET=aarch64-unknown-linux-gnu
+    #- env: TARGET=i686-unknown-linux-gnu
+    - env: TARGET=x86_64-unknown-linux-gnu CODECOV=1
+    # - env: TARGET=x86_64-unknown-linux-musl CODECOV=1
+
+    # OSX
+    - env: TARGET=x86_64-apple-darwin
+      os: osx
+
+before_install:
+  - set -e
+  - rustup self update
+
+install:
+  - sh ci/install.sh
+  - source ~/.cargo/env || true
+
+before_script:
+  - export PATH=$HOME/.cargo/bin:$PATH
+  - cargo install cargo-update || echo "cargo-update already installed"
+  - cargo install cargo-travis || echo "cargo-travis already installed"
+
+script:
+  - bash ci/script.sh
+
+before_deploy:
+  - sh ci/before_deploy.sh
+
+cache: cargo
+before_cache:
+  # Travis can't cache files that are not readable by "others"
+  - chmod -R a+r $HOME/.cargo
+  - find ./target/debug -type f -maxdepth 1 -delete
+  - rm -f  ./target/.rustc_info.json
+  - rm -fr ./target/debug/{deps,.fingerprint}/tantivy*
+  - rm -r target/debug/examples/
+  - ls -1 examples/ | sed -e 's/\.rs$//' | xargs -I "{}" find target/* -name "*{}*" -type f -delete
+
+#branches:
+#  only:
+#    # release tags
+#    - /^v\d+\.\d+\.\d+.*$/
+#    - master
+
+notifications:
+  email:
+    on_success: never

ARCHITECTURE.md

@@ -1,295 +0,0 @@
-# Tantivy
-
-## What is tantivy?
-
-Tantivy is a library that is meant to build search engines. Although it is by no means a port of Lucene, its architecture is strongly inspired by it. If you are familiar with Lucene, you may be struck by the overlapping vocabulary.
-This is not fortuitous.
-
-Tantivy's bread and butter is to address the problem of full-text search :
-
-Given a large set of textual documents, and a text query, return the K-most relevant documents in a very efficient way. To execute these queries rapidly, the tantivy needs to build an index beforehand. The relevance score implemented in the tantivy is not configurable. Tantivy uses the same score as the default similarity used in Lucene / Elasticsearch, called [BM25](https://en.wikipedia.org/wiki/Okapi_BM25).
-
-But tantivy's scope does not stop there. Numerous features are required to power rich-search applications. For instance, one may want to:
-
-- compute the count of documents matching a query in the different section of an e-commerce website,
-- display an average price per meter square for a real estate search engine,
-- take into account historical user data to rank documents in a specific way,
-- or even use tantivy to power an OLAP database.
-
-A more abstract description of the problem space tantivy is trying to address is the following.
-
-Ingest a large set of documents, create an index that makes it possible to
-rapidly select all documents matching a given predicate (also known as a query) and
-collect some information about them ([See collector](#collector-define-what-to-do-with-matched-documents)).
-
-Roughly speaking the design is following these guiding principles:
-
-- Search should be O(1) in memory.
-- Indexing should be O(1) in memory. (In practice it is just sublinear)
-- Search should be as fast as possible
-
-This comes at the cost of the dynamicity of the index: while it is possible to add, and delete documents from our corpus, the tantivy is designed to handle these updates in large batches.
-
-## [core/](src/core): Index, segments, searchers
-
-Core contains all of the high-level code to make it possible to create an index, add documents, delete documents and commit.
-
-This is both the most high-level part of tantivy, the least performance-sensitive one, the seemingly most mundane code... And paradoxically the most complicated part.
-
-### Index and Segments
-
-A tantivy index is a collection of smaller independent immutable segments.
-Each segment contains its own independent set of data structures.
-
-A segment is identified by a segment id that is in fact a UUID.
-The file of a segment has the format
-
- ```segment-id . ext```
-
-The extension signals which data structure (or [`SegmentComponent`](src/core/segment_component.rs)) is stored in the file.
-
-A small `meta.json` file is in charge of keeping track of the list of segments, as well as the schema.
-
-On commit, one segment per indexing thread is written to disk, and the `meta.json` is then updated atomically.
-
-For a better idea of how indexing works, you may read the [following blog post](https://fulmicoton.com/posts/behold-tantivy-part2/).
-
-### Deletes
-
-Deletes happen by deleting a "term". Tantivy does not offer any notion of primary id, so it is up to the user to use a field in their schema as if it was a primary id, and delete the associated term if they want to delete only one specific document.
-
-On commit, tantivy will find all of the segments with documents matching this existing term and remove from [alive bitset file](src/fastfield/alive_bitset.rs) that represents the bitset of the alive document ids.
-Like all segment files, this file is immutable. Because it is possible to have more than one alive bitset file at a given instant, the alive bitset filename has the format ```segment_id . commit_opstamp . del```.
-
-An opstamp is simply an incremental id that identifies any operation applied to the index. For instance, performing a commit or adding a document.
-
-### DocId
-
-Within a segment, all documents are identified by a DocId that ranges within `[0, max_doc)`.
-where `max_doc` is the number of documents in the segment, (deleted or not). Having such a compact `DocId` space is key to the compression of our data structures.
-
-The DocIds are simply allocated in the order documents are added to the index.
-
-### Merges
-
-In separate threads, tantivy's index writer search for opportunities to merge segments.
-The point of segment merge is to:
-
-- eventually get rid of tombstoned documents
-- reduce the otherwise ever-growing number of segments.
-
-Indeed, while having several segments instead of one does not hurt search too much, having hundreds can have a measurable impact on the search performance.
-
-### Searcher
-
-The user of the library usually does not need to know about the existence of Segments.
-Searching is done through an object called a [`Searcher`](src/core/searcher.rs), that captures a
-snapshot of the index at one point of time, by holding a list of [SegmentReader](src/core/segment_reader.rs).
-
-In other words, regardless of commits, file garbage collection, or segment merge that might happen, as long as the user holds and reuse the same [Searcher](src/core/searcher.rs), search will happen on an immutable snapshot of the index.
-
-## [directory/](src/directory): Where should the data be stored?
-
-Tantivy, like Lucene, abstracts the place where the data should be stored in a key-trait
-called [`Directory`](src/directory/directory.rs).
-Contrary to Lucene however, "files" are quite different from some kind of `io::Read` object.
-Check out [`src/directory/directory.rs`](src/directory/directory.rs) trait for more details.
-
-Tantivy ships two main directory implementation: the `MmapDirectory` and the `RamDirectory`,
-but users can extend tantivy with their own implementation.
-
-## [schema/](src/schema): What are documents?
-
-Tantivy's document follows a very strict schema, decided before building any index.
-
-The schema defines all of the fields that the indexes [`Document`](src/schema/document.rs) may and should contain, their types (`text`, `i64`, `u64`, `Date`, ...) as well as how it should be indexed / represented in tantivy.
-
-Depending on the type of the field, you can decide to
-
-- put it in the docstore
-- store it as a fast field
-- index it
-
-Practically, tantivy will push values associated with this type to up to 3 respective
-data structures.
-
-*Limitations*
-
-As of today, tantivy's schema imposes a 1:1 relationship between a field that is being ingested and a field represented in the search index. In sophisticated search application, it is fairly common to want to index a field twice using different tokenizers, or to index the concatenation of several fields together into one field.
-
-This is not something tantivy supports, and it is up to the user to duplicate field / concatenate fields before feeding them to tantivy.
-
-## General information about these data structures
-
-All data structures in tantivy, have:
-
-- a writer
-- a serializer
-- a reader
-
-The writer builds an in-memory representation of a batch of documents. This representation is not searchable. It is just meant as an intermediary mutable representation, to which we can sequentially add
-the document of a batch. At the end of the batch (or if a memory limit is reached), this representation
-is then converted into an on-disk immutable representation, that is extremely compact.
-This conversion is done by the serializer.
-
-Finally, the reader is in charge of offering an API to read on this on-disk read-only representation.
-In tantivy, readers are designed to require very little anonymous memory. The data is read straight from an mmapped file, and loading an index is as fast as mmapping its files.
-
-## [store/](src/store): Here is my DocId, Gimme my document
-
-The docstore is a row-oriented storage that, for each document, stores a subset of the fields
-that are marked as stored in the schema. The docstore is compressed using a general-purpose algorithm
-like LZ4.
-
-**Useful for**
-
-In search engines, it is often used to display search results.
-Once the top 10 documents have been identified, we fetch them from the store, and display them or their snippet on the search result page (aka SERP).
-
-**Not useful for**
-
-Fetching a document from the store is typically a "slow" operation. It usually consists in
-
-- searching into a compact tree-like data structure to find the position of the right block.
-- decompressing a small block
-- returning the document from this block.
-
-It is NOT meant to be called for every document matching a query.
-
-As a rule of thumb, if you hit the docstore more than 100 times per search query, you are probably misusing tantivy.
-
-## [fastfield/](src/fastfield): Here is my DocId, Gimme my value
-
-Fast fields are stored in a column-oriented storage that allows for random access.
-The only compression applied is bitpacking. The column comes with two meta data.
-The minimum value in the column and the number of bits per doc.
-
-Fetching a value for a `DocId` is then as simple as computing
-
-```rust
-min_value + fetch_bits(num_bits * doc_id..num_bits * (doc_id+1))
-```
-
-This operation just requires one memory fetch.
-Because, DocSets are scanned through in order (DocId are iterated in a sorted manner) which
-also help locality.
-
-In Lucene's jargon, fast fields are called DocValues.
-
-**Useful for**
-
-They are typically integer values that are useful to either rank or compute aggregate over
-all of the documents matching a query (aka [DocSet](src/docset.rs)).
-
-For instance, one could define a function to combine upvotes with tantivy's internal relevancy score.
-This can be done by fetching a fast field during scoring.
-One could also compute the mean price of the items matching a query in an e-commerce website.
-This can be done by fetching a fast field in a collector.
-Finally one could decide to post-filter a docset to remove docset with a price within a specific range.
-If the ratio of filtered out documents is not too low, an efficient way to do this is to fetch the price and apply the filter on the collector side.
-
-Aside from integer values, it is also possible to store an actual byte payload.
-For advanced search engine, it is possible to store all of the features required for learning-to-rank in a byte payload, access it during search, and apply the learning-to-rank model.
-
-Finally facets are a specific kind of fast field, and the associated source code is in [`fastfield/facet_reader.rs`](src/fastfield/facet_reader.rs).
-
-# The inverted search index
-
-The inverted index is the core part of full-text search.
-When presented a new document with the text field "Hello, happy tax payer!", tantivy breaks it into a list of so-called tokens. In addition to just splitting these strings into tokens, it might also do different kinds of operations like dropping the punctuation, converting the character to lowercase, apply stemming, etc. Tantivy makes it possible to configure the operations to be applied in the schema (tokenizer/ is the place where these operations are implemented).
-
-For instance, the default tokenizer of tantivy would break our text into: `[hello, happy, tax, payer]`.
-The document will therefore be registered in the inverted index as containing the terms
-`[text:hello, text:happy, text:tax, text:payer]`.
-
-The role of the inverted index is, when given a term, gives us in return a very fast iterator over the sorted doc ids that match the term.
-
-Such an iterator is called a posting list. In addition to giving us `DocId`, they can also give us optionally the number of occurrence of the term for each document, also called term frequency or TF.
-
-These iterators being sorted by DocId, one can create an iterator over the document containing `text:tax AND text:payer`, `(text:tax AND text:payer) OR (text:contribuable)` or any boolean expression.
-
-In order to represent the function
-```Term ⟶ Posting```
-
-The inverted index actually consists of two data structures chained together.
-
-- [Term](src/schema/term.rs) ⟶ [TermInfo](src/postings/term_info.rs) is addressed by the term dictionary.
-- [TermInfo](src/postings/term_info.rs) ⟶ [Posting](src/postings/postings.rs) is addressed by the posting lists.
-
-Where [TermInfo](src/postings/term_info.rs) is an object containing some meta data about a term.
-
-## [termdict/](src/termdict): Here is a term, give me the [TermInfo](src/postings/term_info.rs)
-
-Tantivy's term dictionary is mainly in charge of supplying the function
-
-[Term](src/schema/term.rs) ⟶ [TermInfo](src/postings/term_info.rs)
-
-It is itself broken into two parts.
-
-- [Term](src/schema/term.rs) ⟶ [TermOrdinal](src/termdict/mod.rs) is addressed by a finite state transducer, implemented by the fst crate.
-- [TermOrdinal](src/termdict/mod.rs) ⟶ [TermInfo](src/postings/term_info.rs) is addressed by the term info store.
-
-## [postings/](src/postings): Iterate over documents... very fast
-
-A posting list makes it possible to store a sorted list of doc ids and for each doc store
-a term frequency as well.
-
-The posting lists are stored in a separate file. The [TermInfo](src/postings/term_info.rs) contains an offset into that file and a number of documents for the given posting list. Both are required and sufficient to read the posting list.
-
-The posting list is organized in block of 128 documents.
-One block of doc ids is followed by one block of term frequencies.
-
-The doc ids are delta encoded and bitpacked.
-The term frequencies are bitpacked.
-
-Because the number of docs is rarely a multiple of 128, the last block may contain an arbitrary number of docs between 1 and 127 documents. We then use variable int encoding instead of bitpacking.
-
-## [positions/](src/positions): Where are my terms within the documents?
-
-Phrase queries make it possible to search for documents containing a specific sequence of terms.
-For instance, when the phrase query "the art of war" does not match "the war of art".
-To make it possible, it is possible to specify in the schema that a field should store positions in addition to being indexed.
-
-The token positions of all of the terms are then stored in a separate file with the extension `.pos`.
-The [TermInfo](src/postings/term_info.rs) gives an offset (expressed in position this time) in this file. As we iterate through the docset,
-we advance the position reader by the number of term frequencies of the current document.
-
-## [fieldnorms/](src/fieldnorms): Here is my doc, how many tokens in this field?
-
-The [BM25](https://en.wikipedia.org/wiki/Okapi_BM25) formula also requires to know the number of tokens stored in a specific field for a given document. We store this information on one byte per document in the fieldnorm.
-The fieldnorm is therefore compressed. Values up to 40 are encoded unchanged.
-
-## [tokenizer/](src/tokenizer): How should we process text?
-
-Text processing is key to a good search experience.
-Splits or normalize your text too much, and the search results will have a less precision and a higher recall.
-Do not normalize, or under split your text, you will end up with a higher precision and a lesser recall.
-
-Text processing can be configured by selecting an off-the-shelf [`Tokenizer`](./src/tokenizer/tokenizer.rs) or implementing your own to first split the text into tokens, and then chain different [`TokenFilter`](src/tokenizer/tokenizer.rs)'s to it.
-
-Tantivy's comes with few tokenizers, but external crates are offering advanced tokenizers, such as [Lindera](https://crates.io/crates/lindera) for Japanese.
-
-## [query/](src/query): Define and compose queries
-
-The [Query](src/query/query.rs) trait defines what a query is.
-Due to the necessity for some queries to compute some statistics over the entire index, and because the
-index is composed of several `SegmentReader`, the path from transforming a `Query` to an iterator over documents is slightly convoluted, but fundamentally, this is what a Query is.
-
-The iterator over a document comes with some scoring function. The resulting trait is called a
-[Scorer](src/query/scorer.rs) and is specific to a segment.
-
-Different queries can be combined using the [BooleanQuery](src/query/boolean_query/).
-Tantivy comes with different types of queries and can be extended by implementing
-the `Query`, `Weight`, and `Scorer` traits.
-
-## [collector](src/collector): Define what to do with matched documents
-
-Collectors define how to aggregate the documents matching a query, in the broadest sense possible.
-The search will push matched documents one by one, calling their
-`fn collect(doc: DocId, score: Score);` method.
-
-Users may implement their own collectors by implementing the [Collector](src/collector/mod.rs) trait.
-
-## [query-grammar](query-grammar): Defines the grammar of the query parser
-
-While the [QueryParser](src/query/query_parser/query_parser.rs) struct is located in the `query/` directory, the actual parser combinator used to convert user queries into an AST is in an external crate called `query-grammar`. This part was externalized to lighten the work of the compiler.

CHANGELOG.md

@@ -1,320 +1,47 @@
-Tantivy 0.19
-================================
-#### Bugfixes
-- Fix missing fieldnorms for u64, i64, f64, bool, bytes and date [#1620](https://github.com/quickwit-oss/tantivy/pull/1620) (@PSeitz)
-- Fix interpolation overflow in linear interpolation fastfield codec [#1480](https://github.com/quickwit-oss/tantivy/pull/1480) (@PSeitz @fulmicoton)
-
-#### Features/Improvements
-- Add support for `IN` in queryparser , e.g. `field: IN [val1 val2 val3]` [#1683](https://github.com/quickwit-oss/tantivy/pull/1683) (@trinity-1686a)
-- Skip score calculation, when no scoring is required [#1646](https://github.com/quickwit-oss/tantivy/pull/1646) (@PSeitz)
-- Limit fast fields to u32 (`get_val(u32)`) [#1644](https://github.com/quickwit-oss/tantivy/pull/1644) (@PSeitz)
-- The `DateTime` type has been updated to hold timestamps with microseconds precision.
-  `DateOptions` and `DatePrecision` have been added to configure Date fields. The precision is used to hint on fast values compression. Otherwise, seconds precision is used everywhere else (i.e terms, indexing) [#1396](https://github.com/quickwit-oss/tantivy/pull/1396) (@evanxg852000)
-- Add IP address field type [#1553](https://github.com/quickwit-oss/tantivy/pull/1553) (@PSeitz)
-- Add boolean field type [#1382](https://github.com/quickwit-oss/tantivy/pull/1382) (@boraarslan)
-- Remove Searcher pool and make `Searcher` cloneable. (@PSeitz)
-- Validate settings on create [#1570](https://github.com/quickwit-oss/tantivy/pull/1570) (@PSeitz)
-- Detect and apply gcd on fastfield codecs [#1418](https://github.com/quickwit-oss/tantivy/pull/1418) (@PSeitz)
-- Doc store
-  - use separate thread to compress block store [#1389](https://github.com/quickwit-oss/tantivy/pull/1389) [#1510](https://github.com/quickwit-oss/tantivy/pull/1510) (@PSeitz @fulmicoton)
-  - Expose doc store cache size [#1403](https://github.com/quickwit-oss/tantivy/pull/1403) (@PSeitz)
-  - Enable compression levels for doc store [#1378](https://github.com/quickwit-oss/tantivy/pull/1378) (@PSeitz)
-  - Make block size configurable [#1374](https://github.com/quickwit-oss/tantivy/pull/1374) (@kryesh)
-- Make `tantivy::TantivyError` cloneable [#1402](https://github.com/quickwit-oss/tantivy/pull/1402) (@PSeitz)
-- Add support for phrase slop in query language [#1393](https://github.com/quickwit-oss/tantivy/pull/1393) (@saroh)
-- Aggregation
-  - Add aggregation support for date type [#1693](https://github.com/quickwit-oss/tantivy/pull/1693)(@PSeitz)
-  - Add support for keyed parameter in range and histgram aggregations [#1424](https://github.com/quickwit-oss/tantivy/pull/1424) (@k-yomo)
-  - Add aggregation bucket limit [#1363](https://github.com/quickwit-oss/tantivy/pull/1363) (@PSeitz)
-- Faster indexing
-  - [#1610](https://github.com/quickwit-oss/tantivy/pull/1610) (@PSeitz)
-  - [#1594](https://github.com/quickwit-oss/tantivy/pull/1594) (@PSeitz)
-  - [#1582](https://github.com/quickwit-oss/tantivy/pull/1582) (@PSeitz)
-  - [#1611](https://github.com/quickwit-oss/tantivy/pull/1611) (@PSeitz)
-  - Added a pre-configured stop word filter for various language [#1666](https://github.com/quickwit-oss/tantivy/pull/1666) (@adamreichold)
-
-Tantivy 0.18
-================================
-
-- For date values `chrono` has been replaced with `time` (@uklotzde) #1304 :
-  - The `time` crate is re-exported as `tantivy::time` instead of `tantivy::chrono`.
-  - The type alias `tantivy::DateTime` has been removed.
-  - `Value::Date` wraps `time::PrimitiveDateTime` without time zone information.
-  - Internally date/time values are stored as seconds since UNIX epoch in UTC.
-  - Converting a `time::OffsetDateTime` to `Value::Date` implicitly converts the value into UTC.
-    If this is not desired do the time zone conversion yourself and use `time::PrimitiveDateTime`
-    directly instead.
-- Add [histogram](https://github.com/quickwit-oss/tantivy/pull/1306) aggregation (@PSeitz)
-- Add support for fastfield on text fields (@PSeitz)
-- Add terms aggregation (@PSeitz)
-- Add support for zstd compression (@kryesh)
-
-Tantivy 0.18.1
-================================
-- Hotfix: positions computation.  #1629 (@fmassot, @fulmicoton, @PSeitz)
-
-Tantivy 0.17
-================================
-
-- LogMergePolicy now triggers merges if the ratio of deleted documents reaches a threshold (@shikhar @fulmicoton) [#115](https://github.com/quickwit-oss/tantivy/issues/115)
-- Adds a searcher Warmer API (@shikhar @fulmicoton)
-- Change to non-strict schema. Ignore fields in data which are not defined in schema. Previously this returned an error. #1211
-- Facets are necessarily indexed. Existing index with indexed facets should work out of the box. Index without facets that are marked with index: false should be broken (but they were already broken in a sense). (@fulmicoton) #1195 .
-- Bugfix that could in theory impact durability in theory on some filesystems [#1224](https://github.com/quickwit-oss/tantivy/issues/1224)
-- Schema now offers not indexing fieldnorms (@lpouget) [#922](https://github.com/quickwit-oss/tantivy/issues/922)
-- Reduce the number of fsync calls [#1225](https://github.com/quickwit-oss/tantivy/issues/1225)
-- Fix opening bytes index with dynamic codec (@PSeitz) [#1278](https://github.com/quickwit-oss/tantivy/issues/1278)
-- Added an aggregation collector for range, average and stats compatible with Elasticsearch. (@PSeitz)
-- Added a JSON schema type @fulmicoton [#1251](https://github.com/quickwit-oss/tantivy/issues/1251)
-- Added support for slop in phrase queries @halvorboe [#1068](https://github.com/quickwit-oss/tantivy/issues/1068)
-
-Tantivy 0.16.2
-================================
-
-- Bugfix in FuzzyTermQuery. (transposition_cost_one was not doing anything)
-
-Tantivy 0.16.1
-========================
-
-- Major Bugfix on multivalued fastfield.  #1151
-- Demux operation (@PSeitz)
-
-Tantivy 0.16.0
-=========================
-
-- Bugfix in the filesum check. (@evanxg852000) #1127
-- Bugfix in positions when the index is sorted by a field. (@appaquet) #1125
-
-Tantivy 0.15.3
-=========================
-
-- Major bugfix. Deleting documents was broken when the index was sorted by a field. (@appaquet, @fulmicoton) #1101
-
-Tantivy 0.15.2
-========================
-
-- Major bugfix. DocStore still panics when a deleted doc is at the beginning of a block. (@appaquet) #1088
-
-Tantivy 0.15.1
-=========================
-
-- Major bugfix. DocStore panics when first block is deleted. (@appaquet) #1077
-
-Tantivy 0.15.0
-=========================
-
-- API Changes. Using Range instead of (start, end) in the API and internals (`FileSlice`, `OwnedBytes`, `Snippets`, ...)
-  This change is breaking but migration is trivial.
-- Added an Histogram collector. (@fulmicoton) #994
-- Added support for Option<TCollector>.  (@fulmicoton)
-- DocAddress is now a struct (@scampi) #987
-- Bugfix consistent tie break handling in facet's topk (@hardikpnsp) #357
-- Date field support for range queries (@rihardsk) #516
-- Added lz4-flex as the default compression scheme in tantivy (@PSeitz) #1009
-- Renamed a lot of symbols to avoid all uppercasing on acronyms, as per new clippy recommendation. For instance, RAMDirectory -> RamDirectory. (@fulmicoton)
-- Simplified positions index format (@fulmicoton) #1022
-- Moved bitpacking to bitpacker subcrate and add BlockedBitpacker, which bitpacks blocks of 128 elements (@PSeitz) #1030
-- Added support for more-like-this query in tantivy (@evanxg852000) #1011
-- Added support for sorting an index, e.g presorting documents in an index by a timestamp field. This can heavily improve performance for certain scenarios, by utilizing the sorted data (Top-n optimizations)(@PSeitz). #1026
-- Add iterator over documents in doc store (@PSeitz). #1044
-- Fix log merge policy (@PSeitz). #1043
-- Add detection to avoid small doc store blocks on merge (@PSeitz). #1054
-- Make doc store compression dynamic (@PSeitz). #1060
-- Switch to json for footer version handling (@PSeitz). #1060
-- Updated TermMerger implementation to rely on the union feature of the FST (@scampi) #469
-- Add boolean marking whether position is required in the query_terms API call (@fulmicoton). #1070
-
-Tantivy 0.14.0
-=========================
-
-- Remove dependency to atomicwrites #833 .Implemented by @fulmicoton upon suggestion and research from @asafigan).
-- Migrated tantivy error from the now deprecated `failure` crate to `thiserror` #760. (@hirevo)
-- API Change. Accessing the typed value off a `Schema::Value` now returns an Option instead of panicking if the type does not match.
-- Large API Change in the Directory API. Tantivy used to assume that all files could be somehow memory mapped. After this change, Directory return a `FileSlice` that can be reduced and eventually read into an `OwnedBytes` object. Long and blocking io operation are still required by they do not span over the entire file.
-- Added support for Brotli compression in the DocStore. (@ppodolsky)
-- Added helper for building intersections and unions in BooleanQuery (@guilload)
-- Bugfix in `Query::explain`
-- Removed dependency on `notify` #924. Replaced with `FileWatcher` struct that polls meta file every 500ms in background thread. (@halvorboe @guilload)
-- Added `FilterCollector`, which wraps another collector and filters docs using a predicate over a fast field (@barrotsteindev)
-- Simplified the encoding of the skip reader struct. BlockWAND max tf is now encoded over a single byte. (@fulmicoton)
-- `FilterCollector` now supports all Fast Field value types (@barrotsteindev)
-- FastField are not all loaded when opening the segment reader. (@fulmicoton)
-- Added an API to merge segments, see `tantivy::merge_segments` #1005. (@evanxg852000)
-
-This version breaks compatibility and requires users to reindex everything.
-
-Tantivy 0.13.2
-===================
-
-Bugfix. Acquiring a facet reader on a segment that does not contain any
-doc with this facet returns `None`. (#896)
-
-Tantivy 0.13.1
-===================
-
-Made `Query` and `Collector` `Send + Sync`.
-Updated misc dependency versions.
-
-Tantivy 0.13.0
-======================
-
-Tantivy 0.13 introduce a change in the index format that will require
-you to reindex your index (BlockWAND information are added in the skiplist).
-The index size increase is minor as this information is only added for
-full blocks.
-If you have a massive index for which reindexing is not an option, please contact me
-so that we can discuss possible solutions.
-
-- Bugfix in `FuzzyTermQuery` not matching terms by prefix when it should (@Peachball)
-- Relaxed constraints on the custom/tweak score functions. At the segment level, they can be mut, and they are not required to be Sync + Send.
-- `MMapDirectory::open` does not return a `Result` anymore.
-- Change in the DocSet and Scorer API. (@fulmicoton).
-A freshly created DocSet point directly to their first doc. A sentinel value called TERMINATED marks the end of a DocSet.
-`.advance()` returns the new DocId. `Scorer::skip(target)` has been replaced by `Scorer::seek(target)` and returns the resulting DocId.
-As a result, iterating through DocSet now looks as follows
-
-```rust
-let mut doc = docset.doc();
-while doc != TERMINATED {
-   // ...
-   doc = docset.advance();
-}
-```
-
-The change made it possible to greatly simplify a lot of the docset's code.
-
-- Misc internal optimization and introduction of the `Scorer::for_each_pruning` function. (@fulmicoton)
-- Added an offset option to the Top(.*)Collectors. (@robyoung)
-- Added Block WAND. Performance on TOP-K on term-unions should be greatly increased. (@fulmicoton, and special thanks
-to the PISA team for answering all my questions!)
-
-Tantivy 0.12.0
-======================
-
-- Removing static dispatch in tokenizers for simplicity. (#762)
-- Added backward iteration for `TermDictionary` stream. (@halvorboe)
-- Fixed a performance issue when searching for the posting lists of a missing term (@audunhalland)
-- Added a configurable maximum number of docs (10M by default) for a segment to be considered for merge (@hntd187, landed by @halvorboe #713)
-- Important Bugfix #777, causing tantivy to retain memory mapping. (diagnosed by @poljar)
-- Added support for field boosting. (#547, @fulmicoton)
-
-## How to update?
-
-Crates relying on custom tokenizer, or registering tokenizer in the manager will require some
-minor changes. Check <https://github.com/quickwit-oss/tantivy/blob/main/examples/custom_tokenizer.rs>
-to check for some code sample.
-
-Tantivy 0.11.3
-=======================
-
-- Fixed DateTime as a fast field (#735)
-
-Tantivy 0.11.2
-=======================
-
-- The future returned by `IndexWriter::merge` does not borrow `self` mutably anymore (#732)
-- Exposing a constructor for `WatchHandle` (#731)
-
-Tantivy 0.11.1
-=====================
-
-- Bug fix #729
-
-Tantivy 0.11.0
-=====================
-
-- Added f64 field. Internally reuse u64 code the same way i64 does (@fdb-hiroshima)
-- Various bugfixes in the query parser.
-  - Better handling of hyphens in query parser. (#609)
-  - Better handling of whitespaces.
-- Closes #498 - add support for Elastic-style unbounded range queries for alphanumeric types eg. "title:>hello", "weight:>=70.5", "height:<200" (@petr-tik)
-- API change around `Box<BoxableTokenizer>`. See detail in #629
-- Avoid rebuilding Regex automaton whenever a regex query is reused. #639 (@brainlock)
-- Add footer with some metadata to index files. #605 (@fdb-hiroshima)
-- Add a method to check the compatibility of the footer in the index with the running version of tantivy (@petr-tik)
-- TopDocs collector: ensure stable sorting on equal score. #671 (@brainlock)
-- Added handling of pre-tokenized text fields (#642), which will enable users to
-  load tokens created outside tantivy. See usage in examples/pre_tokenized_text. (@kkoziara)
-- Fix crash when committing multiple times with deleted documents. #681 (@brainlock)
-
-## How to update?
-
-- The index format is changed. You are required to reindex your data to use tantivy 0.11.
-- `Box<dyn BoxableTokenizer>` has been replaced by a `BoxedTokenizer` struct.
-- Regex are now compiled when the `RegexQuery` instance is built. As a result, it can now return
-an error and handling the `Result` is required.
-- `tantivy::version()` now returns a `Version` object. This object implements `ToString()`
-
-Tantivy 0.10.2
-=====================
-
-- Closes #656. Solving memory leak.
-
-Tantivy 0.10.1
-=====================
-
-- Closes #544.  A few users experienced problems with the directory watching system.
-Avoid watching the mmap directory until someone effectively creates a reader that uses
-this functionality.
-
 Tantivy 0.10.0
 =====================
 
 *Tantivy 0.10.0 index format is compatible with the index format in 0.9.0.*
 
-- Added an API to easily tweak or entirely replace the
- default score. See `TopDocs::tweak_score`and `TopScore::custom_score` (@fulmicoton)
 - Added an ASCII folding filter (@drusellers)
-- Bugfix in `query.count` in presence of deletes (@fulmicoton)
-- Added `.explain(...)` in `Query` and `Weight` to (@fulmicoton)
-- Added an efficient way to `delete_all_documents` in `IndexWriter` (@petr-tik).
-  All segments are simply removed.
+- Bugfix in `query.count` in presence of deletes (@pmasurel)
 
 Minor
 ---------
-
-- Switched to Rust 2018 (@uvd)
-- Small simplification of the code.
-Calling .freq() or .doc() when .advance() has never been called
+- Small simplification of the code. 
+Calling .freq() or .doc() when .advance() has never 
 on segment postings should panic from now on.
 - Tokens exceeding `u16::max_value() - 4` chars are discarded silently instead of panicking.
 - Fast fields are now preloaded when the `SegmentReader` is created.
-- `IndexMeta` is now public.  (@hntd187)
-- `IndexWriter` `add_document`, `delete_term`. `IndexWriter` is `Sync`, making it possible to use it with a `Arc<RwLock<IndexWriter>>`. `add_document` and `delete_term` can
-only require a read lock. (@fulmicoton)
-- Introducing `Opstamp` as an expressive type alias for `u64`. (@petr-tik)
-- Stamper now relies on `AtomicU64` on all platforms (@petr-tik)
-- Bugfix - Files get deleted slightly earlier
-- Compilation resources improved (@fdb-hiroshima)
 
 ## How to update?
 
-Your program should be usable as is.
+Your existing indexes are usable as is. Your may or may need some 
+trivial updates.
 
 ### Fast fields
 
 Fast fields used to be accessed directly from the `SegmentReader`.
 The API changed, you are now required to acquire your fast field reader via the
-`segment_reader.fast_fields()`, and use one of the typed method:
-
+`segment_reader.fast_fields()`, and use one of the typed method: 
 - `.u64()`, `.i64()` if your field is single-valued ;
 - `.u64s()`, `.i64s()` if your field is multi-valued ;
 - `.bytes()` if your field is bytes fast field.
 
+
+
 Tantivy 0.9.0
 =====================
-
-*0.9.0 index format is not compatible with the
+*0.9.0 index format is not compatible with the 
 previous index format.*
-
-- MAJOR BUGFIX :
+- MAJOR BUGFIX : 
   Some `Mmap` objects were being leaked, and would never get released. (@fulmicoton)
 - Removed most unsafe (@fulmicoton)
 - Indexer memory footprint improved. (VInt comp, inlining the first block. (@fulmicoton)
 - Stemming in other language possible (@pentlander)
 - Segments with no docs are deleted earlier (@barrotsteindev)
-- Added grouped add and delete operations.
-  They are guaranteed to happen together (i.e. they cannot be split by a commit).
+- Added grouped add and delete operations. 
+  They are guaranteed to happen together (i.e. they cannot be split by a commit). 
   In addition, adds are guaranteed to happen on the same segment. (@elbow-jason)
 - Removed `INT_STORED` and `INT_INDEXED`. It is now possible to use `STORED` and `INDEXED`
   for int fields. (@fulmicoton)
@@ -328,62 +55,59 @@ tantivy 0.9 brought some API breaking change.
 To update from tantivy 0.8, you will need to go through the following steps.
 
 - `schema::INT_INDEXED` and `schema::INT_STORED`  should be replaced by `schema::INDEXED` and `schema::INT_STORED`.
-- The index now does not hold the pool of searcher anymore. You are required to create an intermediary object called
-`IndexReader` for this.
-
+- The index now does not hold the pool of searcher anymore. You are required to create an intermediary object called 
+`IndexReader` for this. 
+    
     ```rust
     // create the reader. You typically need to create 1 reader for the entire
     // lifetime of you program.
     let reader = index.reader()?;
-
+    
     // Acquire a searcher (previously `index.searcher()`) is now written:
     let searcher = reader.searcher();
-
-    // With the default setting of the reader, you are not required to
+    
+    // With the default setting of the reader, you are not required to 
     // call `index.load_searchers()` anymore.
     //
     // The IndexReader will pick up that change automatically, regardless
     // of whether the update was done in a different process or not.
-    // If this behavior is not wanted, you can create your reader with
+    // If this behavior is not wanted, you can create your reader with 
     // the `ReloadPolicy::Manual`, and manually decide when to reload the index
     // by calling `reader.reload()?`.
-
+  
     ```
 
+
 Tantivy 0.8.2
 =====================
-
 Fixing build for x86_64 platforms. (#496)
 No need to update from 0.8.1 if tantivy
 is building on your platform.
 
+
 Tantivy 0.8.1
 =====================
-
 Hotfix of #476.
 
-Merge was reflecting deletes before commit was passed.
+Merge was reflecting deletes before commit was passed. 
 Thanks @barrotsteindev  for reporting the bug.
 
+
 Tantivy 0.8.0
 =====================
-
 *No change in the index format*
-
 - API Breaking change in the collector API. (@jwolfe, @fulmicoton)
-- Multithreaded search (@jwolfe, @fulmicoton)
+- Multithreaded search (@jwolfe, @fulmicoton) 
+
 
 Tantivy 0.7.1
 =====================
-
 *No change in the index format*
-
 - Bugfix: NGramTokenizer panics on non ascii chars
 - Added a space usage API
 
 Tantivy 0.7
 =====================
-
 - Skip data for doc ids and positions (@fulmicoton),
   greatly improving performance
 - Tantivy error now rely on the failure crate (@drusellers)
@@ -393,15 +117,15 @@ Tantivy 0.7
 
 Tantivy 0.6.1
 =========================
-
 - Bugfix #324. GC removing was removing file that were still in useful
 - Added support for parsing AllQuery and RangeQuery via QueryParser
-  - AllQuery: `*`
-  - RangeQuery:
-    - Inclusive `field:[startIncl to endIncl]`
-    - Exclusive `field:{startExcl to endExcl}`
-    - Mixed `field:[startIncl to endExcl}` and vice versa
-    - Unbounded `field:[start to *]`, `field:[* to end]`
+    - AllQuery: `*`
+    - RangeQuery:
+        - Inclusive `field:[startIncl to endIncl]`
+        - Exclusive `field:{startExcl to endExcl}`
+        - Mixed `field:[startIncl to endExcl}` and vice versa
+        - Unbounded `field:[start to *]`, `field:[* to end]`
+ 
 
 Tantivy 0.6
 ==========================
@@ -409,58 +133,63 @@ Tantivy 0.6
 Special thanks to @drusellers and @jason-wolfe for their contributions
 to this release!
 
-- Removed C code. Tantivy is now pure Rust. (@fulmicoton)
-- BM25 (@fulmicoton)
-- Approximate field norms encoded over 1 byte. (@fulmicoton)
-- Compiles on stable rust (@fulmicoton)
+- Removed C code. Tantivy is now pure Rust. (@pmasurel)
+- BM25 (@pmasurel)
+- Approximate field norms encoded over 1 byte. (@pmasurel)
+- Compiles on stable rust (@pmasurel)
 - Add &[u8] fastfield for associating arbitrary bytes to each document (@jason-wolfe) (#270)
-  - Completely uncompressed
-  - Internally: One u64 fast field for indexes, one fast field for the bytes themselves.
+    - Completely uncompressed
+    - Internally: One u64 fast field for indexes, one fast field for the bytes themselves.
 - Add NGram token support (@drusellers)
 - Add Stopword Filter support (@drusellers)
 - Add a FuzzyTermQuery (@drusellers)
 - Add a RegexQuery (@drusellers)
-- Various performance improvements (@fulmicoton)_
+- Various performance improvements (@pmasurel)_
+
 
 Tantivy 0.5.2
 ===========================
-
 - bugfix #274
 - bugfix #280
 - bugfix #289
 
+
 Tantivy 0.5.1
 ==========================
-
 - bugfix #254 : tantivy failed if no documents in a segment contained a specific field.
 
+
 Tantivy 0.5
 ==========================
-
 - Faceting
 - RangeQuery
 - Configurable tokenization pipeline
 - Bugfix in PhraseQuery
 - Various query optimisation
 - Allowing very large indexes
-  - 64 bits file address
-  - Smarter encoding of the `TermInfo` objects
+    - 64 bits file address
+    - Smarter encoding of the `TermInfo` objects
+
+
 
 Tantivy 0.4.3
 ==========================
 
 - Bugfix race condition when deleting files. (#198)
 
+
 Tantivy 0.4.2
 ==========================
 
 - Prevent usage of AVX2 instructions (#201)
 
+
 Tantivy 0.4.1
 ==========================
 
 - Bugfix for non-indexed fields. (#199)
 
+
 Tantivy 0.4.0
 ==========================
 
@@ -475,31 +204,37 @@ Tantivy 0.4.0
   - Searching for a non-indexed field returns an explicit Error
   - Phrase query for non-tokenized field are not tokenized by the query parser.
 - Faster/Better indexing (@fulmicoton)
-  - using murmurhash2
-  - faster merging
-  - more memory efficient fast field writer (@lnicola )
-  - better handling of collisions
-  - lesser memory usage
+    - using murmurhash2
+    - faster merging
+    - more memory efficient fast field writer (@lnicola )
+    - better handling of collisions
+    - lesser memory usage
 - Added API, most notably to iterate over ranges of terms (@fulmicoton)
 - Bugfix that was preventing to unmap segment files, on index drop (@fulmicoton)
 - Made the doc! macro public (@fulmicoton)
 - Added an alternative implementation of the streaming dictionary (@fulmicoton)
 
+
+
 Tantivy 0.3.1
 ==========================
 
 - Expose a method to trigger files garbage collection
 
+
+
 Tantivy 0.3
 ==========================
 
+
 Special thanks to @Kodraus @lnicola @Ameobea @manuel-woelker @celaus
 for their contribution to this release.
 
 Thanks also to everyone in tantivy gitter chat
 for their advise and company :)
 
-<https://gitter.im/tantivy-search/tantivy>
+https://gitter.im/tantivy-search/tantivy
+
 
 Warning:
 
@@ -508,16 +243,19 @@ code and index format.
 You should not expect backward compatibility before
 tantivy 1.0.
 
+
+
 New Features
 ------------
 
 - Delete. You can now delete documents from an index.
 - Support for windows (Thanks to @lnicola)
 
+
 Various Bugfixes & small improvements
 ----------------------------------------
 
-- Added CI for Windows (<https://ci.appveyor.com/project/fulmicoton/tantivy>)
+- Added CI for Windows (https://ci.appveyor.com/project/fulmicoton/tantivy)
 Thanks to @KodrAus ! (#108)
 - Various dependy version update (Thanks to @Ameobea) #76
 - Fixed several race conditions in `Index.wait_merge_threads`
@@ -529,3 +267,7 @@ Thanks to @KodrAus ! (#108)
 - Building binary targets for tantivy-cli (Thanks to @KodrAus)
 - Misc invisible bug fixes, and code cleanup.
 - Use
+
+
+
+

Cargo.toml

@@ -1,87 +1,66 @@
 [package]
 name = "tantivy"
-version = "0.19.0"
+version = "0.10.0-dev"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]
 description = """Search engine library"""
-documentation = "https://docs.rs/tantivy/"
-homepage = "https://github.com/quickwit-oss/tantivy"
-repository = "https://github.com/quickwit-oss/tantivy"
+documentation = "https://tantivy-search.github.io/tantivy/tantivy/index.html"
+homepage = "https://github.com/tantivy-search/tantivy"
+repository = "https://github.com/tantivy-search/tantivy"
 readme = "README.md"
 keywords = ["search", "information", "retrieval"]
-edition = "2021"
-rust-version = "1.62"
 
 [dependencies]
-oneshot = "0.1.5"
-base64 = "0.20.0"
-byteorder = "1.4.3"
-crc32fast = "1.3.2"
-once_cell = "1.10.0"
-regex = { version = "1.5.5", default-features = false, features = ["std", "unicode"] }
-aho-corasick = "0.7"
-tantivy-fst = "0.4.0"
-memmap2 = { version = "0.5.3", optional = true }
-lz4_flex = { version = "0.9.2", default-features = false, features = ["checked-decode"], optional = true }
-brotli = { version = "3.3.4", optional = true }
-zstd = { version = "0.12", optional = true, default-features = false }
-snap = { version = "1.0.5", optional = true }
-tempfile = { version = "3.3.0", optional = true }
-log = "0.4.16"
-serde = { version = "1.0.136", features = ["derive"] }
-serde_json = "1.0.79"
-num_cpus = "1.13.1"
-fs2 = { version = "0.4.3", optional = true }
-levenshtein_automata = "0.2.1"
-uuid = { version = "1.0.0", features = ["v4", "serde"] }
-crossbeam-channel = "0.5.4"
-stable_deref_trait = "1.2.0"
-rust-stemmers = "1.2.0"
-downcast-rs = "1.2.0"
-bitpacking = { version = "0.8.4", default-features = false, features = ["bitpacker4x"] }
-census = "0.4.0"
-rustc-hash = "1.1.0"
-thiserror = "1.0.30"
+base64 = "0.10.0"
+byteorder = "1.0"
+lazy_static = "1"
+regex = "1.0"
+tantivy-fst = "0.1"
+memmap = {version = "0.7", optional=true}
+lz4 = {version="1.20", optional=true}
+snap = {version="0.2"}
+atomicwrites = {version="0.2.2", optional=true}
+tempfile = "3.0"
+log = "0.4"
+combine = ">=3.6.0,<4.0.0"
+tempdir = "0.3"
+serde = "1.0"
+serde_derive = "1.0"
+serde_json = "1.0"
+num_cpus = "1.2"
+fs2={version="0.4", optional=true}
+itertools = "0.8"
+levenshtein_automata = {version="0.1", features=["fst_automaton"]}
+notify = {version="4", optional=true}
+bit-set = "0.5"
+uuid = { version = "0.7.2", features = ["v4", "serde"] }
+crossbeam = "0.5"
+futures = "0.1"
+futures-cpupool = "0.1"
+owning_ref = "0.4"
+stable_deref_trait = "1.0.0"
+rust-stemmers = "1.1"
+downcast-rs = { version="1.0" }
+bitpacking = "0.6"
+census = "0.2"
+fnv = "1.0.6"
+owned-read = "0.4"
+failure = "0.1"
 htmlescape = "0.3.1"
-fail = "0.5.0"
-murmurhash32 = "0.2.0"
-time = { version = "0.3.10", features = ["serde-well-known"] }
-smallvec = "1.8.0"
-rayon = "1.5.2"
-lru = "0.7.5"
-fastdivide = "0.4.0"
-itertools = "0.10.3"
-measure_time = "0.8.2"
-async-trait = "0.1.53"
-arc-swap = "1.5.0"
-
-sstable = { version="0.1", path="./sstable", package ="tantivy-sstable", optional = true }
-stacker = { version="0.1", path="./stacker", package ="tantivy-stacker" }
-tantivy-query-grammar = { version= "0.19.0", path="./query-grammar" }
-tantivy-bitpacker = 		{ version= "0.3", path="./bitpacker" }
-common = 								{ version= "0.5", path = "./common/", package = "tantivy-common" }
-fastfield_codecs = 			{ version= "0.3", path="./fastfield_codecs", default-features = false }
-ownedbytes = 						{ version= "0.5", path="./ownedbytes" }
+fail = "0.2"
+scoped-pool = "1.0"
+murmurhash32 = "0.2"
+chrono = "0.4"
 
 [target.'cfg(windows)'.dependencies]
-winapi = "0.3.9"
+winapi = "0.2"
 
 [dev-dependencies]
-rand = "0.8.5"
-maplit = "1.0.2"
-matches = "0.1.9"
-pretty_assertions = "1.2.1"
-proptest = "1.0.0"
-criterion = "0.4"
-test-log = "0.2.10"
-env_logger = "0.10.0"
-pprof = { version = "0.11.0", features = ["flamegraph", "criterion"] }
-futures = "0.3.21"
-
-[dev-dependencies.fail]
-version = "0.5.0"
-features = ["failpoints"]
+rand = "0.6"
+maplit = "1"
+matches = "0.1.8"
+time = "0.1.42"
 
 [profile.release]
 opt-level = 3
@@ -93,40 +72,13 @@ debug-assertions = true
 overflow-checks = true
 
 [features]
-default = ["mmap", "stopwords", "lz4-compression"]
-mmap = ["fs2", "tempfile", "memmap2"]
-stopwords = []
-
-brotli-compression = ["brotli"]
-lz4-compression = ["lz4_flex"]
-snappy-compression = ["snap"]
-zstd-compression = ["zstd"]
-
-failpoints = ["fail/failpoints"]
+# by default no-fail is disabled. We manually enable it when running test.
+default = ["mmap", "no_fail"]
+mmap = ["atomicwrites", "fs2", "memmap", "notify"]
+lz4-compression = ["lz4"]
+no_fail = ["fail/no_fail"]
 unstable = [] # useful for benches.
+wasm-bindgen = ["uuid/wasm-bindgen"]
 
-quickwit = ["sstable"]
-
-[workspace]
-members = ["query-grammar", "bitpacker", "common", "fastfield_codecs", "ownedbytes", "stacker", "sstable", "columnar"]
-
-# Following the "fail" crate best practises, we isolate
-# tests that define specific behavior in fail check points
-# in a different binary.
-#
-# We do that because, fail rely on a global definition of
-# failpoints behavior and hence, it is incompatible with
-# multithreading.
-[[test]]
-name = "failpoints"
-path = "tests/failpoints/mod.rs"
-required-features = ["fail/failpoints"]
-
-[[bench]]
-name = "analyzer"
-harness = false
-
-[[bench]]
-name = "index-bench"
-harness = false
-
+[badges]
+travis-ci = { repository = "tantivy-search/tantivy" }

Makefile

@@ -1,6 +0,0 @@
-test:
-	echo "Run test only... No examples."
-	cargo test --tests --lib
-
-fmt:
-	cargo +nightly fmt --all

README.md

@@ -1,173 +1,109 @@
-[![Docs](https://docs.rs/tantivy/badge.svg)](https://docs.rs/crate/tantivy/)
-[![Build Status](https://github.com/quickwit-oss/tantivy/actions/workflows/test.yml/badge.svg)](https://github.com/quickwit-oss/tantivy/actions/workflows/test.yml)
-[![codecov](https://codecov.io/gh/quickwit-oss/tantivy/branch/main/graph/badge.svg)](https://codecov.io/gh/quickwit-oss/tantivy)
-[![Join the chat at https://discord.gg/MT27AG5EVE](https://shields.io/discord/908281611840282624?label=chat%20on%20discord)](https://discord.gg/MT27AG5EVE)
+
+[![Build Status](https://travis-ci.org/tantivy-search/tantivy.svg?branch=master)](https://travis-ci.org/tantivy-search/tantivy)
+[![codecov](https://codecov.io/gh/tantivy-search/tantivy/branch/master/graph/badge.svg)](https://codecov.io/gh/tantivy-search/tantivy)
+[![Join the chat at https://gitter.im/tantivy-search/tantivy](https://badges.gitter.im/tantivy-search/tantivy.svg)](https://gitter.im/tantivy-search/tantivy?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Crates.io](https://img.shields.io/crates/v/tantivy.svg)](https://crates.io/crates/tantivy)
+[![Build status](https://ci.appveyor.com/api/projects/status/r7nb13kj23u8m9pj/branch/master?svg=true)](https://ci.appveyor.com/project/fulmicoton/tantivy/branch/master)
+[![Say Thanks!](https://img.shields.io/badge/Say%20Thanks-!-1EAEDB.svg)](https://saythanks.io/to/fulmicoton)
 
 ![Tantivy](https://tantivy-search.github.io/logo/tantivy-logo.png)
 
-**Tantivy** is a **full-text search engine library** written in Rust.
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/0)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/0)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/1)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/1)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/2)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/2)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/3)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/3)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/4)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/4)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/5)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/5)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/6)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/6)
+[![](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/images/7)](https://sourcerer.io/fame/fulmicoton/tantivy-search/tantivy/links/7)
 
-It is closer to [Apache Lucene](https://lucene.apache.org/) than to [Elasticsearch](https://www.elastic.co/products/elasticsearch) or [Apache Solr](https://lucene.apache.org/solr/) in the sense it is not
+[![Become a patron](https://c5.patreon.com/external/logo/become_a_patron_button.png)](https://www.patreon.com/fulmicoton)
+
+
+**Tantivy** is a **full text search engine library** written in rust.
+
+It is closer to [Apache Lucene](https://lucene.apache.org/) than to [Elasticsearch](https://www.elastic.co/products/elasticsearch) and [Apache Solr](https://lucene.apache.org/solr/) in the sense it is not
 an off-the-shelf search engine server, but rather a crate that can be used
 to build such a search engine.
 
 Tantivy is, in fact, strongly inspired by Lucene's design.
 
-If you are looking for an alternative to Elasticsearch or Apache Solr, check out [Quickwit](https://github.com/quickwit-oss/quickwit), our search engine built on top of Tantivy.
-
 # Benchmark
 
-The following [benchmark](https://tantivy-search.github.io/bench/) breakdowns
-performance for different types of queries/collections.
+Tantivy is typically faster than Lucene, but the results will depend on 
+the nature of the queries in your workload.
 
-Your mileage WILL vary depending on the nature of queries and their load.
-
-<img src="doc/assets/images/searchbenchmark.png">
+The following [benchmark](https://tantivy-search.github.io/bench/) break downs 
+performance for different type of queries / collection.
 
 # Features
 
 - Full-text search
-- Configurable tokenizer (stemming available for 17 Latin languages with third party support for Chinese ([tantivy-jieba](https://crates.io/crates/tantivy-jieba) and [cang-jie](https://crates.io/crates/cang-jie)), Japanese ([lindera](https://github.com/lindera-morphology/lindera-tantivy), [Vaporetto](https://crates.io/crates/vaporetto_tantivy), and [tantivy-tokenizer-tiny-segmenter](https://crates.io/crates/tantivy-tokenizer-tiny-segmenter)) and Korean ([lindera](https://github.com/lindera-morphology/lindera-tantivy) + [lindera-ko-dic-builder](https://github.com/lindera-morphology/lindera-ko-dic-builder))
+- Configurable tokenizer. (stemming available for 17 latin languages. Third party support for Chinese ([tantivy-jieba](https://crates.io/crates/tantivy-jieba) and [cang-jie](https://crates.io/crates/cang-jie)) and [Japanese](https://crates.io/crates/tantivy-tokenizer-tiny-segmenter)
 - Fast (check out the :racehorse: :sparkles: [benchmark](https://tantivy-search.github.io/bench/) :sparkles: :racehorse:)
-- Tiny startup time (<10ms), perfect for command-line tools
-- BM25 scoring (the same as Lucene)
-- Natural query language (e.g. `(michael AND jackson) OR "king of pop"`)
-- Phrase queries search (e.g. `"michael jackson"`)
+- Tiny startup time (<10ms), perfect for command line tools
+- BM25 scoring (the same as lucene)
+- Natural query language `(michael AND jackson) OR "king of pop"`
+- Phrase queries search (`"michael jackson"`)
 - Incremental indexing
 - Multithreaded indexing (indexing English Wikipedia takes < 3 minutes on my desktop)
 - Mmap directory
-- SIMD integer compression when the platform/CPU includes the SSE2 instruction set
-- Single valued and multivalued u64, i64, and f64 fast fields (equivalent of doc values in Lucene)
+- SIMD integer compression when the platform/CPU includes the SSE2 instruction set.
+- Single valued and multivalued u64 and i64 fast fields (equivalent of doc values in Lucene)
 - `&[u8]` fast fields
-- Text, i64, u64, f64, dates, and hierarchical facet fields
+- Text, i64, u64, dates and hierarchical facet fields
 - LZ4 compressed document store
 - Range queries
 - Faceted search
 - Configurable indexing (optional term frequency and position indexing)
-- JSON Field
-- Aggregation Collector: range buckets, average, and stats metrics
-- LogMergePolicy with deletes
-- Searcher Warmer API
 - Cheesy logo with a horse
 
-## Non-features
+# Non-features
 
-Distributed search is out of the scope of Tantivy, but if you are looking for this feature, check out [Quickwit](https://github.com/quickwit-oss/quickwit/).
+- Distributed search is out of the scope of tantivy. That being said, tantivy is meant as a
+library upon which one could build a distributed search. Serializable/mergeable collector state for instance, 
+are within the scope of tantivy.
+
+
+# Supported OS and compiler
+
+Tantivy works on stable rust (>= 1.27) and supports Linux, MacOS and Windows.
 
 # Getting started
 
-Tantivy works on stable Rust and supports Linux, macOS, and Windows.
+- [tantivy's simple search example](http://fulmicoton.com/tantivy-examples/simple_search.html)
+- [tantivy-cli and its tutorial](https://github.com/tantivy-search/tantivy-cli).
+`tantivy-cli` is an actual command line interface that makes it easy for you to create a search engine,
+index documents and search via the CLI or a small server with a REST API.
+It will walk you through getting a wikipedia search engine up and running in a few minutes.
+- [reference doc]
+    - [For the last released version](https://docs.rs/tantivy/)
+    - [For the last master branch](https://tantivy-search.github.io/tantivy/tantivy/index.html)
 
-- [Tantivy's simple search example](https://tantivy-search.github.io/examples/basic_search.html)
-- [tantivy-cli and its tutorial](https://github.com/quickwit-oss/tantivy-cli) - `tantivy-cli` is an actual command-line interface that makes it easy for you to create a search engine,
-index documents, and search via the CLI or a small server with a REST API.
-It walks you through getting a Wikipedia search engine up and running in a few minutes.
-- [Reference doc for the last released version](https://docs.rs/tantivy/)
+# Compiling
 
-# How can I support this project?
+## Development
 
-There are many ways to support this project.
+Tantivy compiles on stable rust but requires `Rust >= 1.27`.
+To check out and run tests, you can simply run :
 
-- Use Tantivy and tell us about your experience on [Discord](https://discord.gg/MT27AG5EVE) or by email (paul.masurel@gmail.com)
-- Report bugs
-- Write a blog post
-- Help with documentation by asking questions or submitting PRs
-- Contribute code (you can join [our Discord server](https://discord.gg/MT27AG5EVE))
-- Talk about Tantivy around you
-
-# Contributing code
-
-We use the GitHub Pull Request workflow: reference a GitHub ticket and/or include a comprehensive commit message when opening a PR.
-
-## Minimum supported Rust version
-
-Tantivy currently requires at least Rust 1.62 or later to compile.
-
-## Clone and build locally
-
-Tantivy compiles on stable Rust.
-To check out and run tests, you can simply run:
-
-```bash
-    git clone https://github.com/quickwit-oss/tantivy.git
+    git clone https://github.com/tantivy-search/tantivy.git
     cd tantivy
     cargo build
-```
 
-## Run tests
+## Running tests
 
 Some tests will not run with just `cargo test` because of `fail-rs`.
-To run the tests exhaustively, run `./run-tests.sh`.
+To run the tests exhaustively, run `./run-tests.sh`. 
 
-## Debug
+# How can I support this project ?
 
-You might find it useful to step through the programme with a debugger.
+There are many ways to support this project. 
 
-### A failing test
-
-Make sure you haven't run `cargo clean` after the most recent `cargo test` or `cargo build` to guarantee that the `target/` directory exists. Use this bash script to find the name of the most recent debug build of Tantivy and run it under `rust-gdb`:
-
-```bash
-find target/debug/ -maxdepth 1 -executable -type f -name "tantivy*" -printf '%TY-%Tm-%Td %TT %p\n' | sort -r | cut -d " " -f 3 | xargs -I RECENT_DBG_TANTIVY rust-gdb RECENT_DBG_TANTIVY
-```
-
-Now that you are in `rust-gdb`, you can set breakpoints on lines and methods that match your source code and run the debug executable with flags that you normally pass to `cargo test` like this:
-
-```bash
-$gdb run --test-threads 1 --test $NAME_OF_TEST
-```
-
-### An example
-
-By default, `rustc` compiles everything in the `examples/` directory in debug mode. This makes it easy for you to make examples to reproduce bugs:
-
-```bash
-rust-gdb target/debug/examples/$EXAMPLE_NAME
-$ gdb run
-```
-
-# Companies Using Tantivy
-
-<p align="left">
-<img align="center" src="doc/assets/images/etsy.png" alt="Etsy" height="25" width="auto" />&nbsp;
-<img align="center" src="doc/assets/images/Nuclia.png#gh-light-mode-only" alt="Nuclia" height="25" width="auto" /> &nbsp;
-<img align="center" src="doc/assets/images/humanfirst.png#gh-light-mode-only" alt="Humanfirst.ai" height="30" width="auto" />
-<img align="center" src="doc/assets/images/element.io.svg#gh-light-mode-only" alt="Element.io" height="25" width="auto" />
-<img align="center" src="doc/assets/images/nuclia-dark-theme.png#gh-dark-mode-only" alt="Nuclia" height="35" width="auto" /> &nbsp;
-<img align="center" src="doc/assets/images/humanfirst.ai-dark-theme.png#gh-dark-mode-only" alt="Humanfirst.ai" height="25" width="auto" />&nbsp; &nbsp;
-<img align="center" src="doc/assets/images/element-dark-theme.png#gh-dark-mode-only" alt="Element.io" height="25" width="auto" />
-</p>
-
-# FAQ
-
-### Can I use Tantivy in other languages?
-
-- Python → [tantivy-py](https://github.com/quickwit-oss/tantivy-py)
-- Ruby → [tantiny](https://github.com/baygeldin/tantiny)
-
-You can also find other bindings on [GitHub](https://github.com/search?q=tantivy) but they may be less maintained.
-
-### What are some examples of Tantivy use?
-
-- [seshat](https://github.com/matrix-org/seshat/): A matrix message database/indexer
-- [tantiny](https://github.com/baygeldin/tantiny): Tiny full-text search for Ruby
-- [lnx](https://github.com/lnx-search/lnx): adaptable, typo tolerant search engine with a REST API
-- and [more](https://github.com/search?q=tantivy)!
-
-### On average, how much faster is Tantivy compared to Lucene?
-
-- According to our [search latency benchmark](https://tantivy-search.github.io/bench/), Tantivy is approximately 2x faster than Lucene.
-
-### Does tantivy support incremental indexing?
-
-- Yes.
-
-### How can I edit documents?
-
-- Data in tantivy is immutable. To edit a document, the document needs to be deleted and reindexed.
-
-### When will my documents be searchable during indexing?
-
-- Documents will be searchable after a `commit` is called on an `IndexWriter`. Existing `IndexReader`s will also need to be reloaded in order to reflect the changes. Finally, changes are only visible to newly acquired `Searcher`.
+- If you use tantivy, tell us about your experience on [gitter](https://gitter.im/tantivy-search/tantivy) or by email (paul.masurel@gmail.com)
+- Report bugs
+- Write a blog post
+- Complete documentation
+- Contribute code (you can join [our gitter](https://gitter.im/tantivy-search/tantivy) )
+- Talk about tantivy around you
+- Drop a word on on [![Say Thanks!](https://img.shields.io/badge/Say%20Thanks-!-1EAEDB.svg)](https://saythanks.io/to/fulmicoton) or even [![Become a patron](https://c5.patreon.com/external/logo/become_a_patron_button.png)](https://www.patreon.com/fulmicoton)

appveyor.yml

@@ -18,6 +18,5 @@ install:
 build: false
 
 test_script:
-  - REM SET RUST_LOG=tantivy,test & cargo test --all --verbose --no-default-features --features lz4-compression --features mmap
-  - REM SET RUST_LOG=tantivy,test & cargo test test_store --verbose --no-default-features --features lz4-compression --features snappy-compression --features brotli-compression --features mmap
+  - REM SET RUST_LOG=tantivy,test & cargo test --verbose --no-default-features --features mmap -- --test-threads 1
   - REM SET RUST_BACKTRACE=1 & cargo build --examples

benches/analyzer.rs

@@ -1,22 +0,0 @@
-use criterion::{criterion_group, criterion_main, Criterion};
-use tantivy::tokenizer::TokenizerManager;
-
-const ALICE_TXT: &str = include_str!("alice.txt");
-
-pub fn criterion_benchmark(c: &mut Criterion) {
-    let tokenizer_manager = TokenizerManager::default();
-    let tokenizer = tokenizer_manager.get("default").unwrap();
-    c.bench_function("default-tokenize-alice", |b| {
-        b.iter(|| {
-            let mut word_count = 0;
-            let mut token_stream = tokenizer.token_stream(ALICE_TXT);
-            while token_stream.advance() {
-                word_count += 1;
-            }
-            assert_eq!(word_count, 30_731);
-        })
-    });
-}
-
-criterion_group!(benches, criterion_benchmark);
-criterion_main!(benches);

benches/index-bench.rs

@@ -1,121 +0,0 @@
-use criterion::{criterion_group, criterion_main, Criterion};
-use pprof::criterion::{Output, PProfProfiler};
-use tantivy::schema::{INDEXED, STORED, STRING, TEXT};
-use tantivy::Index;
-
-const HDFS_LOGS: &str = include_str!("hdfs.json");
-const NUM_REPEATS: usize = 2;
-
-pub fn hdfs_index_benchmark(c: &mut Criterion) {
-    let schema = {
-        let mut schema_builder = tantivy::schema::SchemaBuilder::new();
-        schema_builder.add_u64_field("timestamp", INDEXED);
-        schema_builder.add_text_field("body", TEXT);
-        schema_builder.add_text_field("severity", STRING);
-        schema_builder.build()
-    };
-    let schema_with_store = {
-        let mut schema_builder = tantivy::schema::SchemaBuilder::new();
-        schema_builder.add_u64_field("timestamp", INDEXED | STORED);
-        schema_builder.add_text_field("body", TEXT | STORED);
-        schema_builder.add_text_field("severity", STRING | STORED);
-        schema_builder.build()
-    };
-    let dynamic_schema = {
-        let mut schema_builder = tantivy::schema::SchemaBuilder::new();
-        schema_builder.add_json_field("json", TEXT);
-        schema_builder.build()
-    };
-
-    let mut group = c.benchmark_group("index-hdfs");
-    group.sample_size(20);
-    group.bench_function("index-hdfs-no-commit", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(schema.clone());
-            let index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let doc = schema.parse_document(doc_json).unwrap();
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-        })
-    });
-    group.bench_function("index-hdfs-with-commit", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(schema.clone());
-            let mut index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let doc = schema.parse_document(doc_json).unwrap();
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-            index_writer.commit().unwrap();
-        })
-    });
-    group.bench_function("index-hdfs-no-commit-with-docstore", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(schema_with_store.clone());
-            let index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let doc = schema.parse_document(doc_json).unwrap();
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-        })
-    });
-    group.bench_function("index-hdfs-with-commit-with-docstore", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(schema_with_store.clone());
-            let mut index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let doc = schema.parse_document(doc_json).unwrap();
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-            index_writer.commit().unwrap();
-        })
-    });
-    group.bench_function("index-hdfs-no-commit-json-without-docstore", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(dynamic_schema.clone());
-            let json_field = dynamic_schema.get_field("json").unwrap();
-            let mut index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let json_val: serde_json::Map<String, serde_json::Value> =
-                        serde_json::from_str(doc_json).unwrap();
-                    let doc = tantivy::doc!(json_field=>json_val);
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-            index_writer.commit().unwrap();
-        })
-    });
-    group.bench_function("index-hdfs-with-commit-json-without-docstore", |b| {
-        b.iter(|| {
-            let index = Index::create_in_ram(dynamic_schema.clone());
-            let json_field = dynamic_schema.get_field("json").unwrap();
-            let mut index_writer = index.writer_with_num_threads(1, 100_000_000).unwrap();
-            for _ in 0..NUM_REPEATS {
-                for doc_json in HDFS_LOGS.trim().split("\n") {
-                    let json_val: serde_json::Map<String, serde_json::Value> =
-                        serde_json::from_str(doc_json).unwrap();
-                    let doc = tantivy::doc!(json_field=>json_val);
-                    index_writer.add_document(doc).unwrap();
-                }
-            }
-            index_writer.commit().unwrap();
-        })
-    });
-}
-
-criterion_group! {
-    name = benches;
-    config = Criterion::default().with_profiler(PProfProfiler::new(100, Output::Flamegraph(None)));
-    targets = hdfs_index_benchmark
-}
-criterion_main!(benches);

bitpacker/Cargo.toml

@@ -1,17 +0,0 @@
-[package]
-name = "tantivy-bitpacker"
-version = "0.3.0"
-edition = "2021"
-authors = ["Paul Masurel <paul.masurel@gmail.com>"]
-license = "MIT"
-categories = []
-description = """Tantivy-sub crate: bitpacking"""
-repository = "https://github.com/quickwit-oss/tantivy"
-keywords = []
-documentation = "https://docs.rs/tantivy-bitpacker/latest/tantivy_bitpacker"
-homepage = "https://github.com/quickwit-oss/tantivy"
-
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]

bitpacker/benches/bench.rs

@@ -1,35 +0,0 @@
-#![feature(test)]
-
-extern crate test;
-
-#[cfg(test)]
-mod tests {
-    use tantivy_bitpacker::BlockedBitpacker;
-    use test::Bencher;
-
-    #[bench]
-    fn bench_blockedbitp_read(b: &mut Bencher) {
-        let mut blocked_bitpacker = BlockedBitpacker::new();
-        for val in 0..=21500 {
-            blocked_bitpacker.add(val * val);
-        }
-        b.iter(|| {
-            let mut out = 0;
-            for val in 0..=21500 {
-                out = blocked_bitpacker.get(val);
-            }
-            out
-        });
-    }
-
-    #[bench]
-    fn bench_blockedbitp_create(b: &mut Bencher) {
-        b.iter(|| {
-            let mut blocked_bitpacker = BlockedBitpacker::new();
-            for val in 0..=21500 {
-                blocked_bitpacker.add(val * val);
-            }
-            blocked_bitpacker
-        });
-    }
-}

bitpacker/src/blocked_bitpacker.rs

@@ -1,179 +0,0 @@
-use super::bitpacker::BitPacker;
-use super::compute_num_bits;
-use crate::{minmax, BitUnpacker};
-
-const BLOCK_SIZE: usize = 128;
-
-/// `BlockedBitpacker` compresses data in blocks of
-/// 128 elements, while keeping an index on it
-#[derive(Debug, Clone)]
-pub struct BlockedBitpacker {
-    // bitpacked blocks
-    compressed_blocks: Vec<u8>,
-    // uncompressed data, collected until BLOCK_SIZE
-    buffer: Vec<u64>,
-    offset_and_bits: Vec<BlockedBitpackerEntryMetaData>,
-}
-impl Default for BlockedBitpacker {
-    fn default() -> Self {
-        BlockedBitpacker::new()
-    }
-}
-
-/// `BlockedBitpackerEntryMetaData` encodes the
-/// offset and bit_width into a u64 bit field
-///
-/// This saves some space, since 7byte is more
-/// than enough and also keeps the access fast
-/// because of alignment
-#[derive(Debug, Clone, Default)]
-struct BlockedBitpackerEntryMetaData {
-    encoded: u64,
-    base_value: u64,
-}
-
-impl BlockedBitpackerEntryMetaData {
-    fn new(offset: u64, num_bits: u8, base_value: u64) -> Self {
-        let encoded = offset | (num_bits as u64) << (64 - 8);
-        Self {
-            encoded,
-            base_value,
-        }
-    }
-    fn offset(&self) -> u64 {
-        (self.encoded << 8) >> 8
-    }
-    fn num_bits(&self) -> u8 {
-        (self.encoded >> 56) as u8
-    }
-    fn base_value(&self) -> u64 {
-        self.base_value
-    }
-}
-
-#[test]
-fn metadata_test() {
-    let meta = BlockedBitpackerEntryMetaData::new(50000, 6, 40000);
-    assert_eq!(meta.offset(), 50000);
-    assert_eq!(meta.num_bits(), 6);
-}
-
-fn mem_usage<T>(items: &Vec<T>) -> usize {
-    items.capacity() * std::mem::size_of::<T>()
-}
-
-impl BlockedBitpacker {
-    pub fn new() -> Self {
-        let mut compressed_blocks = vec![];
-        compressed_blocks.resize(8, 0);
-        Self {
-            compressed_blocks,
-            buffer: vec![],
-            offset_and_bits: vec![],
-        }
-    }
-
-    /// The memory used (inclusive childs)
-    pub fn mem_usage(&self) -> usize {
-        std::mem::size_of::<BlockedBitpacker>()
-            + self.compressed_blocks.capacity()
-            + mem_usage(&self.offset_and_bits)
-            + mem_usage(&self.buffer)
-    }
-
-    #[inline]
-    pub fn add(&mut self, val: u64) {
-        self.buffer.push(val);
-        if self.buffer.len() == BLOCK_SIZE {
-            self.flush();
-        }
-    }
-
-    pub fn flush(&mut self) {
-        if let Some((min_value, max_value)) = minmax(self.buffer.iter()) {
-            let mut bit_packer = BitPacker::new();
-            let num_bits_block = compute_num_bits(*max_value - min_value);
-            // todo performance: the padding handling could be done better, e.g. use a slice and
-            // return num_bytes written from bitpacker
-            self.compressed_blocks
-                .resize(self.compressed_blocks.len() - 8, 0); // remove padding for bitpacker
-            let offset = self.compressed_blocks.len() as u64;
-            // todo performance: for some bit_width we
-            // can encode multiple vals into the
-            // mini_buffer before checking to flush
-            // (to be done in BitPacker)
-            for val in self.buffer.iter() {
-                bit_packer
-                    .write(
-                        *val - min_value,
-                        num_bits_block,
-                        &mut self.compressed_blocks,
-                    )
-                    .expect("cannot write bitpacking to output"); // write to in memory can't fail
-            }
-            bit_packer.flush(&mut self.compressed_blocks).unwrap();
-            self.offset_and_bits
-                .push(BlockedBitpackerEntryMetaData::new(
-                    offset,
-                    num_bits_block,
-                    *min_value,
-                ));
-
-            self.buffer.clear();
-            self.compressed_blocks
-                .resize(self.compressed_blocks.len() + 8, 0); // add padding for bitpacker
-        }
-    }
-    #[inline]
-    pub fn get(&self, idx: usize) -> u64 {
-        let metadata_pos = idx / BLOCK_SIZE;
-        let pos_in_block = idx % BLOCK_SIZE;
-        if let Some(metadata) = self.offset_and_bits.get(metadata_pos) {
-            let unpacked = BitUnpacker::new(metadata.num_bits()).get(
-                pos_in_block as u32,
-                &self.compressed_blocks[metadata.offset() as usize..],
-            );
-            unpacked + metadata.base_value()
-        } else {
-            self.buffer[pos_in_block]
-        }
-    }
-
-    pub fn iter(&self) -> impl Iterator<Item = u64> + '_ {
-        // todo performance: we could decompress a whole block and cache it instead
-        let bitpacked_elems = self.offset_and_bits.len() * BLOCK_SIZE;
-        let iter = (0..bitpacked_elems)
-            .map(move |idx| self.get(idx))
-            .chain(self.buffer.iter().cloned());
-        iter
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    #[test]
-    fn blocked_bitpacker_empty() {
-        let blocked_bitpacker = BlockedBitpacker::new();
-        assert_eq!(blocked_bitpacker.iter().collect::<Vec<u64>>(), vec![]);
-    }
-    #[test]
-    fn blocked_bitpacker_one() {
-        let mut blocked_bitpacker = BlockedBitpacker::new();
-        blocked_bitpacker.add(50000);
-        assert_eq!(blocked_bitpacker.get(0), 50000);
-        assert_eq!(blocked_bitpacker.iter().collect::<Vec<u64>>(), vec![50000]);
-    }
-    #[test]
-    fn blocked_bitpacker_test() {
-        let mut blocked_bitpacker = BlockedBitpacker::new();
-        for val in 0..21500 {
-            blocked_bitpacker.add(val);
-        }
-        for val in 0..21500 {
-            assert_eq!(blocked_bitpacker.get(val as usize), val);
-        }
-        assert_eq!(blocked_bitpacker.iter().count(), 21500);
-        assert_eq!(blocked_bitpacker.iter().last().unwrap(), 21499);
-    }
-}

bitpacker/src/lib.rs

@@ -1,80 +0,0 @@
-mod bitpacker;
-mod blocked_bitpacker;
-
-pub use crate::bitpacker::{BitPacker, BitUnpacker};
-pub use crate::blocked_bitpacker::BlockedBitpacker;
-
-/// Computes the number of bits that will be used for bitpacking.
-///
-/// In general the target is the minimum number of bits
-/// required to express the amplitude given in argument.
-///
-/// e.g. If the amplitude is 10, we can store all ints on simply 4bits.
-///
-/// The logic is slightly more convoluted here as for optimization
-/// reasons, we want to ensure that a value spawns over at most 8 bytes
-/// of aligned bytes.
-///
-/// Spanning over 9 bytes is possible for instance, if we do
-/// bitpacking with an amplitude of 63 bits.
-/// In this case, the second int will start on bit
-/// 63 (which belongs to byte 7) and ends at byte 15;
-/// Hence 9 bytes (from byte 7 to byte 15 included).
-///
-/// To avoid this, we force the number of bits to 64bits
-/// when the result is greater than `64-8 = 56 bits`.
-///
-/// Note that this only affects rare use cases spawning over
-/// a very large range of values. Even in this case, it results
-/// in an extra cost of at most 12% compared to the optimal
-/// number of bits.
-pub fn compute_num_bits(n: u64) -> u8 {
-    let amplitude = (64u32 - n.leading_zeros()) as u8;
-    if amplitude <= 64 - 8 {
-        amplitude
-    } else {
-        64
-    }
-}
-
-pub fn minmax<I, T>(mut vals: I) -> Option<(T, T)>
-where
-    I: Iterator<Item = T>,
-    T: Copy + Ord,
-{
-    if let Some(first_el) = vals.next() {
-        return Some(vals.fold((first_el, first_el), |(min_val, max_val), el| {
-            (min_val.min(el), max_val.max(el))
-        }));
-    }
-    None
-}
-
-#[test]
-fn test_compute_num_bits() {
-    assert_eq!(compute_num_bits(1), 1u8);
-    assert_eq!(compute_num_bits(0), 0u8);
-    assert_eq!(compute_num_bits(2), 2u8);
-    assert_eq!(compute_num_bits(3), 2u8);
-    assert_eq!(compute_num_bits(4), 3u8);
-    assert_eq!(compute_num_bits(255), 8u8);
-    assert_eq!(compute_num_bits(256), 9u8);
-    assert_eq!(compute_num_bits(5_000_000_000), 33u8);
-}
-
-#[test]
-fn test_minmax_empty() {
-    let vals: Vec<u32> = vec![];
-    assert_eq!(minmax(vals.into_iter()), None);
-}
-
-#[test]
-fn test_minmax_one() {
-    assert_eq!(minmax(vec![1].into_iter()), Some((1, 1)));
-}
-
-#[test]
-fn test_minmax_two() {
-    assert_eq!(minmax(vec![1, 2].into_iter()), Some((1, 2)));
-    assert_eq!(minmax(vec![2, 1].into_iter()), Some((1, 2)));
-}

ci/script.sh

@@ -7,7 +7,7 @@ set -ex
 main() {
     if [ ! -z $CODECOV ]; then
         echo "Codecov"
-        cargo build --verbose && cargo coverage --verbose --all && bash <(curl -s https://codecov.io/bash) -s target/kcov
+        cargo build --verbose && cargo coverage --verbose && bash <(curl -s https://codecov.io/bash) -s target/kcov
     else
         echo "Build"
         cross build --target $TARGET
@@ -15,8 +15,7 @@ main() {
             return
         fi
         echo "Test"
-        cross test --target $TARGET --no-default-features --features mmap
-        cross test --target $TARGET --no-default-features --features mmap query-grammar
+        cross test --target $TARGET --no-default-features --features mmap -- --test-threads 1
     fi
     for example in $(ls examples/*.rs)
     do

columnar/Cargo.toml

@@ -1,26 +0,0 @@
-[package]
-name = "tantivy-columnar"
-version = "0.1.0"
-edition = "2021"
-
-[dependencies]
-stacker = { path = "../stacker", package="tantivy-stacker"}
-serde_json = "1"
-thiserror = "1"
-fnv = "1"
-tantivy-fst =  "0.4.0"
-sstable = { path = "../sstable", package = "tantivy-sstable" }
-common = { path = "../common", package = "tantivy-common" }
-fastfield_codecs = { path = "../fastfield_codecs"}
-ordered-float = "3.4"
-itertools = "0.10"
-
-[features]
-# default = ["quickwit"]
-# quickwit = ["common/quickwit"]
-
-
-
-
-[dev-dependencies]
-proptest = "1"

columnar/README.md

@@ -1,33 +0,0 @@
-# Columnar format
-
-This crate describes columnar format used in tantivy.
-
-
-## Goals
-
-This format is special in the following way.
-- it needs to be compact
-- it does not required to be loaded in memory.
-- it is designed to fit well with quickwit's strange constraint:
-we need to be able to load columns rapidly.
-- columns of several types can be associated with the same column name.
-- it needs to support columns with different types `(str, u64, i64, f64)`
-and different cardinality `(required, optional, multivalued)`.
-- columns, once loaded, offer cheap random access.
-
-# Format
-
-A quickwit/tantivy style sstable associated
-`(column names, column_cardinality, column_type) to range of bytes.
-
-The format of the key is:
-`[column_name][ZERO_BYTE][column_type_header: u8]`
-
-Column name may not contain the zero byte.
-
-Listing all columns associated to `column_name` can therefore
-be done by listing all keys prefixed by
-`[column_name][ZERO_BYTE]`
-
-The associated range of bytes refer to a range of bytes
-

columnar/src/column_type_header.rs

@@ -1,154 +0,0 @@
-use crate::value::NumericalType;
-
-#[derive(Clone, Copy, Hash, Default, Debug, PartialEq, Eq, PartialOrd, Ord)]
-#[repr(u8)]
-pub enum Cardinality {
-    #[default]
-    Required = 0,
-    Optional = 1,
-    Multivalued = 2,
-}
-
-impl Cardinality {
-    pub fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub fn try_from_code(code: u8) -> Option<Cardinality> {
-        match code {
-            0 => Some(Cardinality::Required),
-            1 => Some(Cardinality::Optional),
-            2 => Some(Cardinality::Multivalued),
-            _ => None,
-        }
-    }
-}
-
-#[derive(Hash, Eq, PartialEq, Debug, Clone, Copy)]
-pub enum ColumnType {
-    Bytes,
-    Numerical(NumericalType),
-}
-
-impl ColumnType {
-    pub fn to_code(self) -> u8 {
-        match self {
-            ColumnType::Bytes => 0u8,
-            ColumnType::Numerical(numerical_type) => 1u8 | (numerical_type.to_code() << 1),
-        }
-    }
-
-    pub fn try_from_code(code: u8) -> Option<ColumnType> {
-        if code == 0u8 {
-            return Some(ColumnType::Bytes);
-        }
-        if code & 1u8 == 0u8 {
-            return None;
-        }
-        let numerical_type = NumericalType::try_from_code(code >> 1)?;
-        Some(ColumnType::Numerical(numerical_type))
-    }
-}
-
-/// Represents the type and cardinality of a column.
-/// This is encoded over one-byte and added to a column key in the
-/// columnar sstable.
-///
-/// Cardinality is encoded as the first two highest two bits.
-/// The low 6 bits encode the column type.
-#[derive(Eq, Hash, PartialEq, Debug, Copy, Clone)]
-pub struct ColumnTypeAndCardinality {
-    pub cardinality: Cardinality,
-    pub typ: ColumnType,
-}
-
-#[inline]
-const fn compute_mask(num_bits: u8) -> u8 {
-    if num_bits == 8 {
-        u8::MAX
-    } else {
-        (1u8 << num_bits) - 1
-    }
-}
-
-#[inline]
-fn select_bits<const START: u8, const END: u8>(code: u8) -> u8 {
-    assert!(START <= END);
-    assert!(END <= 8);
-    let num_bits: u8 = END - START;
-    let mask: u8 = compute_mask(num_bits);
-    (code >> START) & mask
-}
-
-#[inline]
-fn place_bits<const START: u8, const END: u8>(code: u8) -> u8 {
-    assert!(START <= END);
-    assert!(END <= 8);
-    let num_bits: u8 = END - START;
-    let mask: u8 = compute_mask(num_bits);
-    assert!(code <= mask);
-    code << START
-}
-
-impl ColumnTypeAndCardinality {
-    pub fn to_code(self) -> u8 {
-        place_bits::<6, 8>(self.cardinality.to_code()) | place_bits::<0, 6>(self.typ.to_code())
-    }
-
-    pub fn try_from_code(code: u8) -> Option<ColumnTypeAndCardinality> {
-        let typ_code = select_bits::<0, 6>(code);
-        let cardinality_code = select_bits::<6, 8>(code);
-        let cardinality = Cardinality::try_from_code(cardinality_code)?;
-        let typ = ColumnType::try_from_code(typ_code)?;
-        assert_eq!(typ.to_code(), typ_code);
-        Some(ColumnTypeAndCardinality { cardinality, typ })
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use std::collections::HashSet;
-
-    use super::ColumnTypeAndCardinality;
-    use crate::column_type_header::{Cardinality, ColumnType};
-
-    #[test]
-    fn test_column_type_header_to_code() {
-        let mut column_type_header_set: HashSet<ColumnTypeAndCardinality> = HashSet::new();
-        for code in u8::MIN..=u8::MAX {
-            if let Some(column_type_header) = ColumnTypeAndCardinality::try_from_code(code) {
-                assert_eq!(column_type_header.to_code(), code);
-                assert!(column_type_header_set.insert(column_type_header));
-            }
-        }
-        assert_eq!(
-            column_type_header_set.len(),
-            3 /* cardinality */ * (1 + 3) // column_types
-        );
-    }
-
-    #[test]
-    fn test_column_type_to_code() {
-        let mut column_type_set: HashSet<ColumnType> = HashSet::new();
-        for code in u8::MIN..=u8::MAX {
-            if let Some(column_type) = ColumnType::try_from_code(code) {
-                assert_eq!(column_type.to_code(), code);
-                assert!(column_type_set.insert(column_type));
-            }
-        }
-        assert_eq!(column_type_set.len(), 1 + 3);
-    }
-
-    #[test]
-    fn test_cardinality_to_code() {
-        let mut num_cardinality = 0;
-        for code in u8::MIN..=u8::MAX {
-            let cardinality_opt = Cardinality::try_from_code(code);
-            if let Some(cardinality) = cardinality_opt {
-                assert_eq!(cardinality.to_code(), code);
-                num_cardinality += 1;
-            }
-        }
-        assert_eq!(num_cardinality, 3);
-    }
-}

columnar/src/dictionary.rs

@@ -1,78 +0,0 @@
-use std::io;
-
-use fnv::FnvHashMap;
-
-fn fst_err_into_io_err(fst_err: tantivy_fst::Error) -> io::Error {
-    match fst_err {
-        tantivy_fst::Error::Fst(fst_err) => {
-            io::Error::new(io::ErrorKind::Other, format!("FST Error: {:?}", fst_err))
-        }
-        tantivy_fst::Error::Io(io_err) => io_err,
-    }
-}
-
-/// `DictionaryBuilder` for dictionary encoding.
-///
-/// It stores the different terms encounterred and assigns them a temporary value
-/// we call unordered id.
-///
-/// Upon serialization, we will sort the ids and hence build a `UnorderedId -> Term ordinal`
-/// mapping.
-#[derive(Default)]
-pub struct DictionaryBuilder {
-    dict: FnvHashMap<Vec<u8>, UnorderedId>,
-}
-
-pub struct IdMapping {
-    unordered_to_ord: Vec<OrderedId>,
-}
-
-impl IdMapping {
-    pub fn to_ord(&self, unordered: UnorderedId) -> OrderedId {
-        self.unordered_to_ord[unordered.0 as usize]
-    }
-}
-
-impl DictionaryBuilder {
-    /// Get or allocate an unordered id.
-    /// (This ID is simply an auto-incremented id.)
-    pub fn get_or_allocate_id(&mut self, term: &[u8]) -> UnorderedId {
-        if let Some(term_id) = self.dict.get(term) {
-            return *term_id;
-        }
-        let new_id = UnorderedId(self.dict.len() as u32);
-        self.dict.insert(term.to_vec(), new_id);
-        new_id
-    }
-
-    /// Serialize the dictionary into an fst, and returns the
-    /// `UnorderedId -> TermOrdinal` map.
-    pub fn serialize<'a, W: io::Write + 'a>(&self, wrt: &mut W) -> io::Result<IdMapping> {
-        serialize_inner(&self.dict, wrt).map_err(fst_err_into_io_err)
-    }
-}
-
-/// Helper function just there for error conversion.
-fn serialize_inner<'a, W: io::Write + 'a>(
-    dict: &FnvHashMap<Vec<u8>, UnorderedId>,
-    wrt: &mut W,
-) -> tantivy_fst::Result<IdMapping> {
-    let mut terms: Vec<(&[u8], UnorderedId)> =
-        dict.iter().map(|(k, v)| (k.as_slice(), *v)).collect();
-    terms.sort_unstable_by_key(|(key, _)| *key);
-    let mut unordered_to_ord: Vec<OrderedId> = vec![OrderedId(0u32); terms.len()];
-    let mut fst_builder = tantivy_fst::MapBuilder::new(wrt)?;
-    for (ord, (key, unordered_id)) in terms.into_iter().enumerate() {
-        let ordered_id = OrderedId(ord as u32);
-        fst_builder.insert(key, ord as u64)?;
-        unordered_to_ord[unordered_id.0 as usize] = ordered_id;
-    }
-    fst_builder.finish()?;
-    Ok(IdMapping { unordered_to_ord })
-}
-
-#[derive(Clone, Copy, Debug)]
-pub struct UnorderedId(pub u32);
-
-#[derive(Clone, Copy)]
-pub struct OrderedId(pub u32);

columnar/src/lib.rs

@@ -1,69 +0,0 @@
-// Copyright (C) 2022 Quickwit, Inc.
-//
-// Quickwit is offered under the AGPL v3.0 and as commercial software.
-// For commercial licensing, contact us at hello@quickwit.io.
-//
-// AGPL:
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as
-// published by the Free Software Foundation, either version 3 of the
-// License, or (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program. If not, see <http://www.gnu.org/licenses/>.
-
-mod column_type_header;
-mod dictionary;
-mod reader;
-mod serializer;
-mod value;
-mod writer;
-
-pub use column_type_header::Cardinality;
-pub use reader::ColumnarReader;
-pub use serializer::ColumnarSerializer;
-pub use writer::ColumnarWriter;
-
-pub type DocId = u32;
-
-#[cfg(test)]
-mod tests {
-    use std::ops::Range;
-
-    use common::file_slice::FileSlice;
-
-    use crate::column_type_header::ColumnTypeAndCardinality;
-    use crate::reader::ColumnarReader;
-    use crate::serializer::ColumnarSerializer;
-    use crate::value::NumericalValue;
-    use crate::ColumnarWriter;
-
-    #[test]
-    fn test_dataframe_writer() {
-        let mut dataframe_writer = ColumnarWriter::default();
-        dataframe_writer.record_numerical(1u32, b"srical.value", NumericalValue::U64(1u64));
-        dataframe_writer.record_numerical(2u32, b"srical.value", NumericalValue::U64(2u64));
-        dataframe_writer.record_numerical(4u32, b"srical.value", NumericalValue::I64(2i64));
-        let mut buffer: Vec<u8> = Vec::new();
-        let serializer = ColumnarSerializer::new(&mut buffer);
-        dataframe_writer.serialize(5, serializer).unwrap();
-        let columnar_fileslice = FileSlice::from(buffer);
-        let columnar = ColumnarReader::open(columnar_fileslice).unwrap();
-        assert_eq!(columnar.num_columns(), 1);
-        let cols: Vec<(ColumnTypeAndCardinality, Range<u64>)> =
-            columnar.read_columns("srical.value").unwrap();
-        assert_eq!(cols.len(), 1);
-        // Right now this 31 bytes are spent as follows
-        //
-        // - header 14 bytes
-        // - vals  8 //< due to padding? could have been 1byte?.
-        // - null footer 6 bytes
-        // - version footer 3 bytes // Should be file-wide
-        assert_eq!(cols[0].1, 0..31);
-    }
-}

columnar/src/reader/mod.rs

@@ -1,66 +0,0 @@
-use std::ops::Range;
-use std::{io, mem};
-
-use common::file_slice::FileSlice;
-use common::BinarySerializable;
-use sstable::{Dictionary, SSTableRange};
-
-use crate::column_type_header::ColumnTypeAndCardinality;
-
-fn io_invalid_data(msg: String) -> io::Error {
-    io::Error::new(io::ErrorKind::InvalidData, msg) // format!("Invalid key found.
-                                                    // {key_bytes:?}")));
-}
-pub struct ColumnarReader {
-    column_dictionary: Dictionary<SSTableRange>,
-    column_data: FileSlice,
-}
-
-impl ColumnarReader {
-    pub fn num_columns(&self) -> usize {
-        self.column_dictionary.num_terms()
-    }
-
-    pub fn open(file_slice: FileSlice) -> io::Result<ColumnarReader> {
-        let (file_slice_without_sstable_len, sstable_len_bytes) =
-            file_slice.split_from_end(mem::size_of::<u64>());
-        let mut sstable_len_bytes = sstable_len_bytes.read_bytes()?;
-        let sstable_len = u64::deserialize(&mut sstable_len_bytes)?;
-        let (column_data, sstable) =
-            file_slice_without_sstable_len.split_from_end(sstable_len as usize);
-        let column_dictionary = Dictionary::open(sstable)?;
-        Ok(ColumnarReader {
-            column_dictionary,
-            column_data,
-        })
-    }
-
-    pub fn read_columns(
-        &self,
-        field_name: &str,
-    ) -> io::Result<Vec<(ColumnTypeAndCardinality, Range<u64>)>> {
-        let mut start_key = field_name.to_string();
-        start_key.push('\0');
-        let mut end_key = field_name.to_string();
-        end_key.push(1u8 as char);
-        let mut stream = self
-            .column_dictionary
-            .range()
-            .ge(start_key.as_bytes())
-            .lt(end_key.as_bytes())
-            .into_stream()?;
-        let mut results = Vec::new();
-        while stream.advance() {
-            let key_bytes: &[u8] = stream.key();
-            if !key_bytes.starts_with(start_key.as_bytes()) {
-                return Err(io_invalid_data(format!("Invalid key found. {key_bytes:?}")));
-            }
-            let column_code: u8 = key_bytes.last().cloned().unwrap();
-            let column_type_and_cardinality = ColumnTypeAndCardinality::try_from_code(column_code)
-                .ok_or_else(|| io_invalid_data(format!("Unknown column code `{column_code}`")))?;
-            let range = stream.value().clone();
-            results.push((column_type_and_cardinality, range));
-        }
-        Ok(results)
-    }
-}

columnar/src/serializer.rs

@@ -1,39 +0,0 @@
-use std::io;
-use std::io::Write;
-use std::ops::Range;
-
-use common::CountingWriter;
-use sstable::value::RangeWriter;
-use sstable::SSTableRange;
-
-pub struct ColumnarSerializer<W: io::Write> {
-    wrt: CountingWriter<W>,
-    sstable_range: sstable::Writer<Vec<u8>, RangeWriter>,
-}
-
-impl<W: io::Write> ColumnarSerializer<W> {
-    pub fn new(wrt: W) -> ColumnarSerializer<W> {
-        let sstable_range: sstable::Writer<Vec<u8>, RangeWriter> =
-            sstable::Dictionary::<SSTableRange>::builder(Vec::with_capacity(100_000)).unwrap();
-        ColumnarSerializer {
-            wrt: CountingWriter::wrap(wrt),
-            sstable_range,
-        }
-    }
-
-    pub fn record_column_offsets(&mut self, key: &[u8], byte_range: Range<u64>) -> io::Result<()> {
-        self.sstable_range.insert(key, &byte_range)
-    }
-
-    pub fn wrt(&mut self) -> &mut CountingWriter<W> {
-        &mut self.wrt
-    }
-
-    pub fn finalize(mut self) -> io::Result<()> {
-        let sstable_bytes: Vec<u8> = self.sstable_range.finish()?;
-        let sstable_num_bytes: u64 = sstable_bytes.len() as u64;
-        self.wrt.write_all(&sstable_bytes)?;
-        self.wrt.write_all(&sstable_num_bytes.to_le_bytes()[..])?;
-        Ok(())
-    }
-}

columnar/src/value.rs

@@ -1,123 +0,0 @@
-use ordered_float::NotNan;
-
-#[derive(Copy, Clone, Debug, PartialEq)]
-pub enum NumericalValue {
-    I64(i64),
-    U64(u64),
-    F64(NotNan<f64>),
-}
-
-impl From<u64> for NumericalValue {
-    fn from(val: u64) -> NumericalValue {
-        NumericalValue::U64(val)
-    }
-}
-
-impl From<i64> for NumericalValue {
-    fn from(val: i64) -> Self {
-        NumericalValue::I64(val)
-    }
-}
-
-impl From<NotNan<f64>> for NumericalValue {
-    fn from(val: NotNan<f64>) -> Self {
-        NumericalValue::F64(val)
-    }
-}
-
-impl NumericalValue {
-    pub fn numerical_type(&self) -> NumericalType {
-        match self {
-            NumericalValue::F64(_) => NumericalType::F64,
-            NumericalValue::I64(_) => NumericalType::I64,
-            NumericalValue::U64(_) => NumericalType::U64,
-        }
-    }
-}
-
-impl Eq for NumericalValue {}
-
-#[derive(Clone, Copy, Debug, Default, Hash, Eq, PartialEq)]
-#[repr(u8)]
-pub enum NumericalType {
-    #[default]
-    I64 = 0,
-    U64 = 1,
-    F64 = 2,
-}
-
-impl NumericalType {
-    pub fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub fn try_from_code(code: u8) -> Option<NumericalType> {
-        match code {
-            0 => Some(NumericalType::I64),
-            1 => Some(NumericalType::U64),
-            2 => Some(NumericalType::F64),
-            _ => None,
-        }
-    }
-}
-
-/// We voluntarily avoid using `Into` here to keep this
-/// implementation quirk as private as possible.
-///
-/// This coercion trait actually panics if it is used
-/// to convert a loose types to a stricter type.
-///
-/// The level is strictness is somewhat arbitrary.
-/// - i64
-/// - u64
-/// - f64.
-pub(crate) trait Coerce {
-    fn coerce(numerical_value: NumericalValue) -> Self;
-}
-
-impl Coerce for i64 {
-    fn coerce(value: NumericalValue) -> Self {
-        match value {
-            NumericalValue::I64(val) => val,
-            NumericalValue::U64(val) => val as i64,
-            NumericalValue::F64(_) => unreachable!(),
-        }
-    }
-}
-
-impl Coerce for u64 {
-    fn coerce(value: NumericalValue) -> Self {
-        match value {
-            NumericalValue::I64(val) => val as u64,
-            NumericalValue::U64(val) => val,
-            NumericalValue::F64(_) => unreachable!(),
-        }
-    }
-}
-
-impl Coerce for NotNan<f64> {
-    fn coerce(value: NumericalValue) -> Self {
-        match value {
-            NumericalValue::I64(val) => unsafe { NotNan::new_unchecked(val as f64) },
-            NumericalValue::U64(val) => unsafe { NotNan::new_unchecked(val as f64) },
-            NumericalValue::F64(val) => val,
-        }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::NumericalType;
-
-    #[test]
-    fn test_numerical_type_code() {
-        let mut num_numerical_type = 0;
-        for code in u8::MIN..=u8::MAX {
-            if let Some(numerical_type) = NumericalType::try_from_code(code) {
-                assert_eq!(numerical_type.to_code(), code);
-                num_numerical_type += 1;
-            }
-        }
-        assert_eq!(num_numerical_type, 3);
-    }
-}

columnar/src/writer/column_operation.rs

@@ -1,321 +0,0 @@
-use std::fmt;
-use std::num::NonZeroU8;
-
-use ordered_float::NotNan;
-use thiserror::Error;
-
-use crate::dictionary::UnorderedId;
-use crate::value::NumericalValue;
-use crate::DocId;
-
-/// When we build a columnar dataframe, we first just group
-/// all mutations per column, and append them in append-only object.
-///
-/// We represents all of these operations as `ColumnOperation`.
-#[derive(Eq, PartialEq, Debug, Clone, Copy)]
-pub(crate) enum ColumnOperation<T> {
-    NewDoc(DocId),
-    Value(T),
-}
-
-impl<T> From<T> for ColumnOperation<T> {
-    fn from(value: T) -> Self {
-        ColumnOperation::Value(value)
-    }
-}
-
-#[allow(clippy::from_over_into)]
-pub(crate) trait SymbolValue: Into<MiniBuffer> + Clone + Copy + fmt::Debug {
-    fn deserialize(header: NonZeroU8, bytes: &mut &[u8]) -> Result<Self, ParseError>;
-}
-
-pub(crate) struct MiniBuffer {
-    pub bytes: [u8; 9],
-    pub len: usize,
-}
-
-impl MiniBuffer {
-    pub fn as_slice(&self) -> &[u8] {
-        &self.bytes[..self.len]
-    }
-}
-
-fn compute_header_byte(typ: SymbolType, len: usize) -> u8 {
-    assert!(len <= 9);
-    (len << 4) as u8 | typ as u8
-}
-
-impl SymbolValue for NumericalValue {
-    fn deserialize(header_byte: NonZeroU8, bytes: &mut &[u8]) -> Result<Self, ParseError> {
-        let (typ, len) = parse_header_byte(header_byte)?;
-        let value_bytes: &[u8];
-        (value_bytes, *bytes) = bytes.split_at(len);
-        let symbol: NumericalValue = match typ {
-            SymbolType::U64 => {
-                let mut octet: [u8; 8] = [0u8; 8];
-                octet[..value_bytes.len()].copy_from_slice(value_bytes);
-                let val: u64 = u64::from_le_bytes(octet);
-                NumericalValue::U64(val)
-            }
-            SymbolType::I64 => {
-                let mut octet: [u8; 8] = [0u8; 8];
-                octet[..value_bytes.len()].copy_from_slice(value_bytes);
-                let encoded: u64 = u64::from_le_bytes(octet);
-                let val: i64 = decode_zig_zag(encoded);
-                NumericalValue::I64(val)
-            }
-            SymbolType::Float => {
-                let octet: [u8; 8] =
-                    value_bytes.try_into().map_err(|_| ParseError::InvalidLen {
-                        typ: SymbolType::Float,
-                        len,
-                    })?;
-                let val_possibly_nan = f64::from_le_bytes(octet);
-                let val_not_nan = NotNan::new(val_possibly_nan)
-                    .map_err(|_| ParseError::NaN)?;
-                NumericalValue::F64(val_not_nan)
-            }
-        };
-        Ok(symbol)
-    }
-}
-
-#[allow(clippy::from_over_into)]
-impl Into<MiniBuffer> for NumericalValue {
-    fn into(self) -> MiniBuffer {
-        let mut bytes = [0u8; 9];
-        match self {
-            NumericalValue::F64(val) => {
-                let len = 8;
-                let header_byte = compute_header_byte(SymbolType::Float, len);
-                bytes[0] = header_byte;
-                bytes[1..].copy_from_slice(&val.to_le_bytes());
-                MiniBuffer {
-                    bytes,
-                    len: len + 1,
-                }
-            }
-            NumericalValue::U64(val) => {
-                let len = compute_num_bytes_for_u64(val);
-                let header_byte = compute_header_byte(SymbolType::U64, len);
-                bytes[0] = header_byte;
-                bytes[1..].copy_from_slice(&val.to_le_bytes());
-                MiniBuffer {
-                    bytes,
-                    len: len + 1,
-                }
-            }
-            NumericalValue::I64(val) => {
-                let encoded = encode_zig_zag(val);
-                let len = compute_num_bytes_for_u64(encoded);
-                let header_byte = compute_header_byte(SymbolType::I64, len);
-                bytes[0] = header_byte;
-                bytes[1..].copy_from_slice(&encoded.to_le_bytes());
-                MiniBuffer {
-                    bytes,
-                    len: len + 1,
-                }
-            }
-        }
-    }
-}
-
-#[allow(clippy::from_over_into)]
-impl Into<MiniBuffer> for UnorderedId {
-    fn into(self) -> MiniBuffer {
-        let mut bytes = [0u8; 9];
-        let val = self.0 as u64;
-        let len = compute_num_bytes_for_u64(val) + 1;
-        bytes[0] = len as u8;
-        bytes[1..].copy_from_slice(&val.to_le_bytes());
-        MiniBuffer { bytes, len }
-    }
-}
-
-impl SymbolValue for UnorderedId {
-    fn deserialize(header: NonZeroU8, bytes: &mut &[u8]) -> Result<UnorderedId, ParseError> {
-        let len = header.get() as usize;
-        let symbol_bytes: &[u8];
-        (symbol_bytes, *bytes) = bytes.split_at(len);
-        let mut value_bytes = [0u8; 4];
-        value_bytes[..len - 1].copy_from_slice(&symbol_bytes[1..]);
-        let value = u32::from_le_bytes(value_bytes);
-        Ok(UnorderedId(value))
-    }
-}
-
-const HEADER_MASK: u8 = (1u8 << 4) - 1u8;
-
-fn compute_num_bytes_for_u64(val: u64) -> usize {
-    let msb = (64u32 - val.leading_zeros()) as usize;
-    (msb + 7) / 8
-}
-
-fn parse_header_byte(byte: NonZeroU8) -> Result<(SymbolType, usize), ParseError> {
-    let len = (byte.get() as usize) >> 4;
-    let typ_code = byte.get() & HEADER_MASK;
-    let typ = SymbolType::try_from(typ_code)?;
-    Ok((typ, len))
-}
-
-#[derive(Error, Debug)]
-pub enum ParseError {
-    #[error("Type byte unknown `{0}`")]
-    UnknownType(u8),
-    #[error("Invalid len for type `{len}` for type `{typ:?}`.")]
-    InvalidLen { typ: SymbolType, len: usize },
-    #[error("Missing bytes.")]
-    MissingBytes,
-    #[error("Not a number value.")]
-    NaN,
-}
-
-impl<V: SymbolValue> ColumnOperation<V> {
-    pub fn serialize(self) -> MiniBuffer {
-        match self {
-            ColumnOperation::NewDoc(doc) => {
-                let mut minibuf: [u8; 9] = [0u8; 9];
-                minibuf[0] = 0u8;
-                minibuf[1..5].copy_from_slice(&doc.to_le_bytes());
-                MiniBuffer {
-                    bytes: minibuf,
-                    len: 5,
-                }
-            }
-            ColumnOperation::Value(val) => val.into(),
-        }
-    }
-
-    pub fn deserialize(bytes: &mut &[u8]) -> Result<Self, ParseError> {
-        if bytes.is_empty() {
-            return Err(ParseError::MissingBytes);
-        }
-        let header_byte = bytes[0];
-        *bytes = &bytes[1..];
-        if let Some(header_byte) = NonZeroU8::new(header_byte) {
-            let value = V::deserialize(header_byte, bytes)?;
-            Ok(ColumnOperation::Value(value))
-        } else {
-            let doc_bytes: &[u8];
-            (doc_bytes, *bytes) = bytes.split_at(4);
-            let doc: u32 =
-                u32::from_le_bytes(doc_bytes.try_into().map_err(|_| ParseError::MissingBytes)?);
-            Ok(ColumnOperation::NewDoc(doc))
-        }
-    }
-}
-
-#[derive(Copy, Clone, Debug, Eq, PartialEq)]
-#[repr(u8)]
-pub enum SymbolType {
-    U64 = 1u8,
-    I64 = 2u8,
-    Float = 3u8,
-}
-
-impl TryFrom<u8> for SymbolType {
-    type Error = ParseError;
-
-    fn try_from(byte: u8) -> Result<Self, ParseError> {
-        match byte {
-            1u8 => Ok(SymbolType::U64),
-            2u8 => Ok(SymbolType::I64),
-            3u8 => Ok(SymbolType::Float),
-            _ => Err(ParseError::UnknownType(byte)),
-        }
-    }
-}
-
-fn encode_zig_zag(n: i64) -> u64 {
-    ((n << 1) ^ (n >> 63)) as u64
-}
-
-fn decode_zig_zag(n: u64) -> i64 {
-    ((n >> 1) as i64) ^ (-((n & 1) as i64))
-}
-
-#[cfg(test)]
-mod tests {
-    use super::{SymbolType, *};
-
-    #[track_caller]
-    fn test_zig_zag_aux(val: i64) {
-        let encoded = super::encode_zig_zag(val);
-        assert_eq!(decode_zig_zag(encoded), val);
-        if let Some(abs_val) = val.checked_abs() {
-            let abs_val = abs_val as u64;
-            assert!(encoded <= abs_val * 2);
-        }
-    }
-
-    #[test]
-    fn test_zig_zag() {
-        assert_eq!(encode_zig_zag(0i64), 0u64);
-        assert_eq!(encode_zig_zag(-1i64), 1u64);
-        assert_eq!(encode_zig_zag(1i64), 2u64);
-        test_zig_zag_aux(0i64);
-        test_zig_zag_aux(i64::MIN);
-        test_zig_zag_aux(i64::MAX);
-    }
-
-    use proptest::prelude::any;
-    use proptest::proptest;
-
-    proptest! {
-        #[test]
-        fn test_proptest_zig_zag(val in any::<i64>()) {
-            test_zig_zag_aux(val);
-        }
-    }
-
-    #[track_caller]
-    fn ser_deser_header_byte_aux(symbol_type: SymbolType, len: usize) {
-        let header_byte = compute_header_byte(symbol_type, len);
-        let (serdeser_numerical_type, serdeser_len) =
-            parse_header_byte(NonZeroU8::new(header_byte).unwrap()).unwrap();
-        assert_eq!(symbol_type, serdeser_numerical_type);
-        assert_eq!(len, serdeser_len);
-    }
-
-    #[test]
-    fn test_header_byte_serialization() {
-        for len in 1..9 {
-            ser_deser_header_byte_aux(SymbolType::Float, len);
-            ser_deser_header_byte_aux(SymbolType::I64, len);
-            ser_deser_header_byte_aux(SymbolType::U64, len);
-        }
-    }
-
-    #[track_caller]
-    fn ser_deser_symbol(symbol: ColumnOperation<NumericalValue>) {
-        let buf = symbol.serialize();
-        let mut bytes = &buf.bytes[..];
-        let serdeser_symbol = ColumnOperation::deserialize(&mut bytes).unwrap();
-        assert_eq!(bytes.len() + buf.len, buf.bytes.len());
-        assert_eq!(symbol, serdeser_symbol);
-    }
-
-    #[test]
-    fn test_compute_num_bytes_for_u64() {
-        assert_eq!(compute_num_bytes_for_u64(0), 0);
-        assert_eq!(compute_num_bytes_for_u64(1), 1);
-        assert_eq!(compute_num_bytes_for_u64(255), 1);
-        assert_eq!(compute_num_bytes_for_u64(256), 2);
-        assert_eq!(compute_num_bytes_for_u64((1 << 16) - 1), 2);
-        assert_eq!(compute_num_bytes_for_u64(1 << 16), 3);
-    }
-
-    #[test]
-    fn test_symbol_serialization() {
-        ser_deser_symbol(ColumnOperation::NewDoc(0));
-        ser_deser_symbol(ColumnOperation::NewDoc(3));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::I64(0i64)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::I64(1i64)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::U64(257u64)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::I64(-257i64)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::I64(i64::MIN)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::U64(0u64)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::U64(u64::MIN)));
-        ser_deser_symbol(ColumnOperation::Value(NumericalValue::U64(u64::MAX)));
-    }
-}

columnar/src/writer/mod.rs

@@ -1,675 +0,0 @@
-mod column_operation;
-mod value_index;
-
-use std::io::{self, Write};
-
-use column_operation::ColumnOperation;
-use common::CountingWriter;
-use fastfield_codecs::serialize::ValueIndexInfo;
-use fastfield_codecs::{Column, MonotonicallyMappableToU64, VecColumn};
-use ordered_float::NotNan;
-use stacker::{Addr, ArenaHashMap, ExpUnrolledLinkedList, MemoryArena};
-
-use crate::column_type_header::{ColumnType, ColumnTypeAndCardinality};
-use crate::dictionary::{DictionaryBuilder, IdMapping, UnorderedId};
-use crate::value::{Coerce, NumericalType, NumericalValue};
-use crate::writer::column_operation::SymbolValue;
-use crate::writer::value_index::{IndexBuilder, SpareIndexBuilders};
-use crate::{Cardinality, ColumnarSerializer, DocId};
-
-#[derive(Copy, Clone, Default)]
-struct ColumnWriter {
-    // Detected cardinality of the column so far.
-    cardinality: Cardinality,
-    // Last document inserted.
-    // None if no doc has been added yet.
-    last_doc_opt: Option<u32>,
-    // Buffer containing the serialized values.
-    values: ExpUnrolledLinkedList,
-}
-
-#[derive(Clone, Copy, Default)]
-pub struct NumericalColumnWriter {
-    compatible_numerical_types: CompatibleNumericalTypes,
-    column_writer: ColumnWriter,
-}
-
-#[derive(Clone, Copy)]
-struct CompatibleNumericalTypes {
-    all_values_within_i64_range: bool,
-    all_values_within_u64_range: bool,
-}
-
-impl Default for CompatibleNumericalTypes {
-    fn default() -> CompatibleNumericalTypes {
-        CompatibleNumericalTypes {
-            all_values_within_i64_range: true,
-            all_values_within_u64_range: true,
-        }
-    }
-}
-
-impl CompatibleNumericalTypes {
-    pub fn accept_value(&mut self, numerical_value: NumericalValue) {
-        match numerical_value {
-            NumericalValue::I64(val_i64) => {
-                let value_within_u64_range = val_i64 >= 0i64;
-                self.all_values_within_u64_range &= value_within_u64_range;
-            }
-            NumericalValue::U64(val_u64) => {
-                let value_within_i64_range = val_u64 < i64::MAX as u64;
-                self.all_values_within_i64_range &= value_within_i64_range;
-            }
-            NumericalValue::F64(_) => {
-                self.all_values_within_i64_range = false;
-                self.all_values_within_u64_range = false;
-            }
-        }
-    }
-
-    pub fn to_numerical_type(self) -> NumericalType {
-        if self.all_values_within_i64_range {
-            NumericalType::I64
-        } else if self.all_values_within_u64_range {
-            NumericalType::U64
-        } else {
-            NumericalType::F64
-        }
-    }
-}
-
-impl NumericalColumnWriter {
-    pub fn record_numerical_value(
-        &mut self,
-        doc: DocId,
-        value: NumericalValue,
-        arena: &mut MemoryArena,
-    ) {
-        self.compatible_numerical_types.accept_value(value);
-        self.column_writer.record(doc, value, arena);
-    }
-}
-
-impl ColumnWriter {
-    fn symbol_iterator<'a, V: SymbolValue>(
-        &self,
-        arena: &MemoryArena,
-        buffer: &'a mut Vec<u8>,
-    ) -> impl Iterator<Item = ColumnOperation<V>> + 'a {
-        buffer.clear();
-        self.values.read_to_end(arena, buffer);
-        let mut cursor: &[u8] = &buffer[..];
-        std::iter::from_fn(move || {
-            if cursor.is_empty() {
-                return None;
-            }
-            let symbol = ColumnOperation::deserialize(&mut cursor)
-                .expect("Failed to deserialize symbol from in-memory. This should never happen.");
-            Some(symbol)
-        })
-    }
-
-    fn delta_with_last_doc(&self, doc: DocId) -> u32 {
-        self.last_doc_opt
-            .map(|last_doc| doc - last_doc)
-            .unwrap_or(doc + 1u32)
-    }
-
-    /// Records a change of the document being recorded.
-    ///
-    /// This function will also update the cardinality of the column
-    /// if necessary.
-    fn record(&mut self, doc: DocId, value: NumericalValue, arena: &mut MemoryArena) {
-        // Difference between `doc` and the last doc.
-        match self.delta_with_last_doc(doc) {
-            0 => {
-                // This is the last encounterred document.
-                self.cardinality = Cardinality::Multivalued;
-            }
-            1 => {
-                self.last_doc_opt = Some(doc);
-                self.write_symbol::<NumericalValue>(ColumnOperation::NewDoc(doc), arena);
-            }
-            _ => {
-                self.cardinality = self.cardinality.max(Cardinality::Optional);
-                self.last_doc_opt = Some(doc);
-                self.write_symbol::<NumericalValue>(ColumnOperation::NewDoc(doc), arena);
-            }
-        }
-        self.write_symbol(ColumnOperation::Value(value), arena);
-    }
-
-    // Get the cardinality.
-    // The overall number of docs in the column is necessary to
-    // deal with the case where the all docs contain 1 value, except some documents
-    // at the end of the column.
-    fn get_cardinality(&self, num_docs: DocId) -> Cardinality {
-        if self.delta_with_last_doc(num_docs) > 1 {
-            self.cardinality.max(Cardinality::Optional)
-        } else {
-            self.cardinality
-        }
-    }
-
-    fn write_symbol<V: SymbolValue>(
-        &mut self,
-        symbol: ColumnOperation<V>,
-        arena: &mut MemoryArena,
-    ) {
-        self.values
-            .writer(arena)
-            .extend_from_slice(symbol.serialize().as_slice());
-    }
-}
-
-#[derive(Copy, Clone, Default)]
-pub struct BytesColumnWriter {
-    dictionary_id: u32,
-    column_writer: ColumnWriter,
-}
-
-impl BytesColumnWriter {
-    pub fn with_dictionary_id(dictionary_id: u32) -> BytesColumnWriter {
-        BytesColumnWriter {
-            dictionary_id,
-            column_writer: Default::default(),
-        }
-    }
-
-    pub fn record_bytes(
-        &mut self,
-        doc: DocId,
-        bytes: &[u8],
-        dictionaries: &mut [DictionaryBuilder],
-        arena: &mut MemoryArena,
-    ) {
-        let unordered_id = dictionaries[self.dictionary_id as usize].get_or_allocate_id(bytes);
-        let numerical_value = NumericalValue::U64(unordered_id.0 as u64);
-        self.column_writer.record(doc, numerical_value, arena);
-    }
-}
-
-pub struct ColumnarWriter {
-    numerical_field_hash_map: ArenaHashMap,
-    bytes_field_hash_map: ArenaHashMap,
-    arena: MemoryArena,
-    // Dictionaries used to store dictionary-encoded values.
-    dictionaries: Vec<DictionaryBuilder>,
-    buffers: SpareBuffers,
-}
-
-#[derive(Default)]
-struct SpareBuffers {
-    byte_buffer: Vec<u8>,
-    value_index_builders: SpareIndexBuilders,
-    i64_values: Vec<i64>,
-    u64_values: Vec<u64>,
-    f64_values: Vec<ordered_float::NotNan<f64>>,
-}
-
-impl Default for ColumnarWriter {
-    fn default() -> Self {
-        ColumnarWriter {
-            numerical_field_hash_map: ArenaHashMap::new(10_000),
-            bytes_field_hash_map: ArenaHashMap::new(10_000),
-            dictionaries: Vec::new(),
-            arena: MemoryArena::default(),
-            buffers: SpareBuffers::default(),
-        }
-    }
-}
-
-#[derive(Copy, Clone, Ord, PartialOrd, Eq, PartialEq, Debug)]
-enum BytesOrNumerical {
-    Bytes,
-    Numerical,
-}
-
-impl ColumnarWriter {
-    pub fn record_numerical(&mut self, doc: DocId, key: &[u8], numerical_value: NumericalValue) {
-        let (hash_map, arena) = (&mut self.numerical_field_hash_map, &mut self.arena);
-        hash_map.mutate_or_create(key, |column_opt: Option<NumericalColumnWriter>| {
-            let mut column: NumericalColumnWriter = column_opt.unwrap_or_default();
-            column.record_numerical_value(doc, numerical_value, arena);
-            column
-        });
-    }
-
-    pub fn record_bytes(&mut self, doc: DocId, key: &[u8], value: &[u8]) {
-        let (hash_map, arena, dictionaries) = (
-            &mut self.bytes_field_hash_map,
-            &mut self.arena,
-            &mut self.dictionaries,
-        );
-        hash_map.mutate_or_create(key, |column_opt: Option<BytesColumnWriter>| {
-            let mut column: BytesColumnWriter = column_opt.unwrap_or_else(|| {
-                let dictionary_id = dictionaries.len() as u32;
-                dictionaries.push(DictionaryBuilder::default());
-                BytesColumnWriter::with_dictionary_id(dictionary_id)
-            });
-            column.record_bytes(doc, value, dictionaries, arena);
-            column
-        });
-    }
-
-    pub fn serialize<W: io::Write>(
-        &mut self,
-        num_docs: DocId,
-        mut serializer: ColumnarSerializer<W>,
-    ) -> io::Result<()> {
-        let mut field_columns: Vec<(&[u8], BytesOrNumerical, Addr)> = self
-            .numerical_field_hash_map
-            .iter()
-            .map(|(term, addr, _)| (term, BytesOrNumerical::Numerical, addr))
-            .collect();
-        field_columns.extend(
-            self.bytes_field_hash_map
-                .iter()
-                .map(|(term, addr, _)| (term, BytesOrNumerical::Bytes, addr)),
-        );
-        let mut key_buffer = Vec::new();
-        field_columns.sort_unstable_by_key(|(key, col_type, _)| (*key, *col_type));
-        let (arena, buffers, dictionaries) = (&self.arena, &mut self.buffers, &self.dictionaries);
-        for (key, bytes_or_numerical, addr) in field_columns {
-            let wrt = serializer.wrt();
-            let start_offset = wrt.written_bytes();
-            let column_type_and_cardinality: ColumnTypeAndCardinality =
-                match bytes_or_numerical {
-                BytesOrNumerical::Bytes => {
-                    let BytesColumnWriter { dictionary_id, column_writer } =
-                        self.bytes_field_hash_map.read(addr);
-                    let dictionary_builder =
-                        &dictionaries[dictionary_id as usize];
-                    serialize_bytes_column(
-                        &column_writer,
-                        num_docs,
-                        dictionary_builder,
-                        arena,
-                        buffers,
-                        wrt,
-                    )?;
-                    ColumnTypeAndCardinality {
-                        cardinality: column_writer.get_cardinality(num_docs),
-                        typ: ColumnType::Bytes,
-                    }
-                }
-                BytesOrNumerical::Numerical => {
-                    let NumericalColumnWriter { compatible_numerical_types, column_writer  } =
-                        self.numerical_field_hash_map.read(addr);
-                    let cardinality = column_writer.get_cardinality(num_docs);
-                    let numerical_type = compatible_numerical_types.to_numerical_type();
-                    serialize_numerical_column(
-                        cardinality,
-                        numerical_type,
-                        &column_writer,
-                        num_docs,
-                        arena,
-                        buffers,
-                        wrt,
-                    )?;
-                    ColumnTypeAndCardinality {
-                        cardinality,
-                        typ: ColumnType::Numerical(numerical_type),
-                    }
-                }
-            };
-            let end_offset = wrt.written_bytes();
-            let key_with_type = prepare_key(key, column_type_and_cardinality, &mut key_buffer);
-            serializer.record_column_offsets(key_with_type, start_offset..end_offset)?;
-        }
-        serializer.finalize()?;
-        Ok(())
-    }
-}
-
-/// Returns a key consisting of the concatenation of the key and the column_type_and_cardinality
-/// code.
-fn prepare_key<'a>(
-    key: &[u8],
-    column_type_cardinality: ColumnTypeAndCardinality,
-    buffer: &'a mut Vec<u8>,
-) -> &'a [u8] {
-    buffer.clear();
-    buffer.extend_from_slice(key);
-    buffer.push(0u8);
-    buffer.push(column_type_cardinality.to_code());
-    &buffer[..]
-}
-
-fn serialize_bytes_column<W: io::Write>(
-    column_writer: &ColumnWriter,
-    num_docs: DocId,
-    dictionary_builder: &DictionaryBuilder,
-    arena: &MemoryArena,
-    buffers: &mut SpareBuffers,
-    wrt: &mut CountingWriter<W>,
-) -> io::Result<()> {
-    let start_offset = wrt.written_bytes();
-    let id_mapping: IdMapping = dictionary_builder.serialize(wrt)?;
-    let dictionary_num_bytes: u32 = (wrt.written_bytes() - start_offset) as u32;
-    let cardinality = column_writer.get_cardinality(num_docs);
-    let SpareBuffers {
-        byte_buffer,
-        value_index_builders,
-        u64_values,
-        ..
-    } = buffers;
-    let symbol_iterator = column_writer
-        .symbol_iterator(arena, byte_buffer)
-        .map(|symbol: ColumnOperation<UnorderedId>| {
-            // We map unordered ids to ordered ids.
-            match symbol {
-                ColumnOperation::Value(unordered_id) => {
-                    let ordered_id = id_mapping.to_ord(unordered_id);
-                    ColumnOperation::Value(ordered_id.0 as u64)
-                }
-                ColumnOperation::NewDoc(doc) => ColumnOperation::NewDoc(doc),
-            }
-        });
-    serialize_column(
-        symbol_iterator,
-        cardinality,
-        num_docs,
-        value_index_builders,
-        u64_values,
-        wrt,
-    )?;
-    wrt.write_all(&dictionary_num_bytes.to_le_bytes()[..])?;
-    Ok(())
-}
-
-fn serialize_numerical_column<W: io::Write>(
-    cardinality: Cardinality,
-    numerical_type: NumericalType,
-    column_writer: &ColumnWriter,
-    num_docs: DocId,
-    arena: &MemoryArena,
-    buffers: &mut SpareBuffers,
-    wrt: &mut W,
-) -> io::Result<()> {
-    let SpareBuffers {
-        byte_buffer,
-        value_index_builders,
-        u64_values,
-        i64_values,
-        f64_values,
-    } = buffers;
-    let symbol_iterator = column_writer.symbol_iterator(arena, byte_buffer);
-    match numerical_type {
-        NumericalType::I64 => {
-            serialize_column(
-                coerce_numerical_symbol::<i64>(symbol_iterator),
-                cardinality,
-                num_docs,
-                value_index_builders,
-                i64_values,
-                wrt,
-            )?;
-        }
-        NumericalType::U64 => {
-            serialize_column(
-                coerce_numerical_symbol::<u64>(symbol_iterator),
-                cardinality,
-                num_docs,
-                value_index_builders,
-                u64_values,
-                wrt,
-            )?;
-        }
-        NumericalType::F64 => {
-            serialize_column(
-                coerce_numerical_symbol::<NotNan<f64>>(symbol_iterator),
-                cardinality,
-                num_docs,
-                value_index_builders,
-                f64_values,
-                wrt,
-            )?;
-        }
-    };
-    Ok(())
-}
-
-fn serialize_column<
-    T: Copy + Ord + Default + Send + Sync + MonotonicallyMappableToU64,
-    W: io::Write,
->(
-    symbol_iterator: impl Iterator<Item = ColumnOperation<T>>,
-    cardinality: Cardinality,
-    num_docs: DocId,
-    value_index_builders: &mut SpareIndexBuilders,
-    values: &mut Vec<T>,
-    wrt: &mut W,
-) -> io::Result<()>
-where
-    for<'a> VecColumn<'a, T>: Column<T>,
-{
-    match cardinality {
-        Cardinality::Required => {
-            consume_symbol_iterator(
-                symbol_iterator,
-                value_index_builders.borrow_required_index_builder(),
-                values,
-            );
-            fastfield_codecs::serialize(
-                VecColumn::from(&values[..]),
-                wrt,
-                &fastfield_codecs::ALL_CODEC_TYPES[..],
-            )?;
-        }
-        Cardinality::Optional => {
-            let optional_index_builder = value_index_builders.borrow_optional_index_builder();
-            consume_symbol_iterator(symbol_iterator, optional_index_builder, values);
-            let optional_index = optional_index_builder.finish(num_docs);
-            fastfield_codecs::serialize::serialize_new(
-                ValueIndexInfo::SingleValue(Box::new(optional_index)),
-                VecColumn::from(&values[..]),
-                wrt,
-                &fastfield_codecs::ALL_CODEC_TYPES[..],
-            )?;
-        }
-        Cardinality::Multivalued => {
-            let multivalued_index_builder = value_index_builders.borrow_multivalued_index_builder();
-            consume_symbol_iterator(symbol_iterator, multivalued_index_builder, values);
-            let multivalued_index = multivalued_index_builder.finish(num_docs);
-            fastfield_codecs::serialize::serialize_new(
-                ValueIndexInfo::MultiValue(Box::new(multivalued_index)),
-                VecColumn::from(&values[..]),
-                wrt,
-                &fastfield_codecs::ALL_CODEC_TYPES[..],
-            )?;
-        }
-    }
-    Ok(())
-}
-
-fn coerce_numerical_symbol<T>(
-    symbol_iterator: impl Iterator<Item = ColumnOperation<NumericalValue>>,
-) -> impl Iterator<Item = ColumnOperation<T>>
-where T: Coerce {
-    symbol_iterator.map(|symbol| match symbol {
-        ColumnOperation::NewDoc(doc) => ColumnOperation::NewDoc(doc),
-        ColumnOperation::Value(numerical_value) => {
-            ColumnOperation::Value(Coerce::coerce(numerical_value))
-        }
-    })
-}
-
-fn consume_symbol_iterator<T, TIndexBuilder: IndexBuilder>(
-    symbol_iterator: impl Iterator<Item = ColumnOperation<T>>,
-    index_builder: &mut TIndexBuilder,
-    values: &mut Vec<T>,
-) {
-    for symbol in symbol_iterator {
-        match symbol {
-            ColumnOperation::NewDoc(doc) => {
-                index_builder.record_doc(doc);
-            }
-            ColumnOperation::Value(value) => {
-                index_builder.record_value();
-                values.push(value);
-            }
-        }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use ordered_float::NotNan;
-    use stacker::MemoryArena;
-
-    use super::prepare_key;
-    use crate::column_type_header::{ColumnType, ColumnTypeAndCardinality};
-    use crate::value::{NumericalType, NumericalValue};
-    use crate::writer::column_operation::ColumnOperation;
-    use crate::writer::CompatibleNumericalTypes;
-    use crate::Cardinality;
-
-    #[test]
-    fn test_prepare_key_bytes() {
-        let mut buffer: Vec<u8> = b"somegarbage".to_vec();
-        let column_type_and_cardinality = ColumnTypeAndCardinality {
-            typ: ColumnType::Bytes,
-            cardinality: Cardinality::Optional,
-        };
-        let prepared_key = prepare_key(b"root\0child", column_type_and_cardinality, &mut buffer);
-        assert_eq!(prepared_key.len(), 12);
-        assert_eq!(&prepared_key[..10], b"root\0child");
-        assert_eq!(prepared_key[10], 0u8);
-        assert_eq!(prepared_key[11], column_type_and_cardinality.to_code());
-    }
-
-    #[test]
-    fn test_column_writer_required_simple() {
-        let mut arena = MemoryArena::default();
-        let mut column_writer = super::ColumnWriter::default();
-        column_writer.record(0u32, 14i64.into(), &mut arena);
-        column_writer.record(1u32, 15i64.into(), &mut arena);
-        column_writer.record(2u32, (-16i64).into(), &mut arena);
-        assert_eq!(column_writer.get_cardinality(3), Cardinality::Required);
-        let mut buffer = Vec::new();
-        let symbols: Vec<ColumnOperation<NumericalValue>> = column_writer
-            .symbol_iterator(&mut arena, &mut buffer)
-            .collect();
-        assert_eq!(symbols.len(), 6);
-        assert!(matches!(symbols[0], ColumnOperation::NewDoc(0u32)));
-        assert!(matches!(
-            symbols[1],
-            ColumnOperation::Value(NumericalValue::I64(14i64))
-        ));
-        assert!(matches!(symbols[2], ColumnOperation::NewDoc(1u32)));
-        assert!(matches!(
-            symbols[3],
-            ColumnOperation::Value(NumericalValue::I64(15i64))
-        ));
-        assert!(matches!(symbols[4], ColumnOperation::NewDoc(2u32)));
-        assert!(matches!(
-            symbols[5],
-            ColumnOperation::Value(NumericalValue::I64(-16i64))
-        ));
-    }
-
-    #[test]
-    fn test_column_writer_optional_cardinality_missing_first() {
-        let mut arena = MemoryArena::default();
-        let mut column_writer = super::ColumnWriter::default();
-        column_writer.record(1u32, 15i64.into(), &mut arena);
-        column_writer.record(2u32, (-16i64).into(), &mut arena);
-        assert_eq!(column_writer.get_cardinality(3), Cardinality::Optional);
-        let mut buffer = Vec::new();
-        let symbols: Vec<ColumnOperation<NumericalValue>> = column_writer
-            .symbol_iterator(&mut arena, &mut buffer)
-            .collect();
-        assert_eq!(symbols.len(), 4);
-        assert!(matches!(symbols[0], ColumnOperation::NewDoc(1u32)));
-        assert!(matches!(
-            symbols[1],
-            ColumnOperation::Value(NumericalValue::I64(15i64))
-        ));
-        assert!(matches!(symbols[2], ColumnOperation::NewDoc(2u32)));
-        assert!(matches!(
-            symbols[3],
-            ColumnOperation::Value(NumericalValue::I64(-16i64))
-        ));
-    }
-
-    #[test]
-    fn test_column_writer_optional_cardinality_missing_last() {
-        let mut arena = MemoryArena::default();
-        let mut column_writer = super::ColumnWriter::default();
-        column_writer.record(0u32, 15i64.into(), &mut arena);
-        assert_eq!(column_writer.get_cardinality(2), Cardinality::Optional);
-        let mut buffer = Vec::new();
-        let symbols: Vec<ColumnOperation<NumericalValue>> = column_writer
-            .symbol_iterator(&mut arena, &mut buffer)
-            .collect();
-        assert_eq!(symbols.len(), 2);
-        assert!(matches!(symbols[0], ColumnOperation::NewDoc(0u32)));
-        assert!(matches!(
-            symbols[1],
-            ColumnOperation::Value(NumericalValue::I64(15i64))
-        ));
-    }
-
-    #[test]
-    fn test_column_writer_multivalued() {
-        let mut arena = MemoryArena::default();
-        let mut column_writer = super::ColumnWriter::default();
-        column_writer.record(0u32, 16i64.into(), &mut arena);
-        column_writer.record(0u32, 17i64.into(), &mut arena);
-        assert_eq!(column_writer.get_cardinality(1), Cardinality::Multivalued);
-        let mut buffer = Vec::new();
-        let symbols: Vec<ColumnOperation<NumericalValue>> = column_writer
-            .symbol_iterator(&mut arena, &mut buffer)
-            .collect();
-        assert_eq!(symbols.len(), 3);
-        assert!(matches!(symbols[0], ColumnOperation::NewDoc(0u32)));
-        assert!(matches!(
-            symbols[1],
-            ColumnOperation::Value(NumericalValue::I64(16i64))
-        ));
-        assert!(matches!(
-            symbols[2],
-            ColumnOperation::Value(NumericalValue::I64(17i64))
-        ));
-    }
-
-    #[track_caller]
-    fn test_column_writer_coercion_iter_aux(
-        values: impl Iterator<Item = NumericalValue>,
-        expected_numerical_type: NumericalType,
-    ) {
-        let mut compatible_numerical_types = CompatibleNumericalTypes::default();
-        for value in values {
-            compatible_numerical_types.accept_value(value);
-        }
-        assert_eq!(
-            compatible_numerical_types.to_numerical_type(),
-            expected_numerical_type
-        );
-    }
-
-    #[track_caller]
-    fn test_column_writer_coercion_aux(
-        values: &[NumericalValue],
-        expected_numerical_type: NumericalType,
-    ) {
-        test_column_writer_coercion_iter_aux(values.iter().copied(), expected_numerical_type);
-        test_column_writer_coercion_iter_aux(values.iter().rev().copied(), expected_numerical_type);
-    }
-
-    #[test]
-    fn test_column_writer_coercion() {
-        test_column_writer_coercion_aux(&[], NumericalType::I64);
-        test_column_writer_coercion_aux(&[1i64.into()], NumericalType::I64);
-        test_column_writer_coercion_aux(&[1u64.into()], NumericalType::I64);
-        // We don't detect exact integer at the moment. We could!
-        test_column_writer_coercion_aux(&[NotNan::new(1f64).unwrap().into()], NumericalType::F64);
-        test_column_writer_coercion_aux(&[u64::MAX.into()], NumericalType::U64);
-        test_column_writer_coercion_aux(&[(i64::MAX as u64).into()], NumericalType::U64);
-        test_column_writer_coercion_aux(&[(1u64 << 63).into()], NumericalType::U64);
-        test_column_writer_coercion_aux(&[1i64.into(), 1u64.into()], NumericalType::I64);
-        test_column_writer_coercion_aux(&[u64::MAX.into(), (-1i64).into()], NumericalType::F64);
-    }
-}

columnar/src/writer/value_index.rs

@@ -1,218 +0,0 @@
-use fastfield_codecs::serialize::{MultiValueIndexInfo, SingleValueIndexInfo};
-
-use crate::DocId;
-
-/// The `IndexBuilder` interprets a sequence of
-/// calls of the form:
-/// (record_doc,record_value+)*
-/// and can then serialize the results into an index.
-///
-/// It has different implementation depending on whether the
-/// cardinality is required, optional, or multivalued.
-pub(crate) trait IndexBuilder {
-    fn record_doc(&mut self, doc: DocId);
-    #[inline]
-    fn record_value(&mut self) {}
-}
-
-/// The RequiredIndexBuilder does nothing.
-#[derive(Default)]
-pub struct RequiredIndexBuilder;
-
-impl IndexBuilder for RequiredIndexBuilder {
-    #[inline(always)]
-    fn record_doc(&mut self, _doc: DocId) {}
-}
-
-#[derive(Default)]
-pub struct OptionalIndexBuilder {
-    docs: Vec<DocId>,
-}
-
-struct SingleValueArrayIndex<'a> {
-    docs: &'a [DocId],
-    num_docs: DocId,
-}
-
-impl<'a> SingleValueIndexInfo for SingleValueArrayIndex<'a> {
-    fn num_vals(&self) -> u32 {
-        self.num_docs as u32
-    }
-
-    fn num_non_nulls(&self) -> u32 {
-        self.docs.len() as u32
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = u32> + '_> {
-        Box::new(self.docs.iter().copied())
-    }
-}
-
-impl OptionalIndexBuilder {
-    pub fn finish(&mut self, num_docs: DocId) -> impl SingleValueIndexInfo + '_ {
-        debug_assert!(self
-            .docs
-            .last()
-            .copied()
-            .map(|last_doc| last_doc < num_docs)
-            .unwrap_or(true));
-        SingleValueArrayIndex {
-            docs: &self.docs[..],
-            num_docs,
-        }
-    }
-
-    fn reset(&mut self) {
-        self.docs.clear();
-    }
-}
-
-impl IndexBuilder for OptionalIndexBuilder {
-    #[inline(always)]
-    fn record_doc(&mut self, doc: DocId) {
-        debug_assert!(self
-            .docs
-            .last()
-            .copied()
-            .map(|prev_doc| doc > prev_doc)
-            .unwrap_or(true));
-        self.docs.push(doc);
-    }
-}
-
-#[derive(Default)]
-pub struct MultivaluedIndexBuilder {
-    // TODO should we switch to `start_offset`?
-    end_values: Vec<DocId>,
-    total_num_vals_seen: u32,
-}
-
-pub struct MultivaluedValueArrayIndex<'a> {
-    end_offsets: &'a [DocId],
-}
-
-impl<'a> MultiValueIndexInfo for MultivaluedValueArrayIndex<'a> {
-    fn num_docs(&self) -> u32 {
-        self.end_offsets.len() as u32
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.end_offsets.last().copied().unwrap_or(0u32)
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = u32> + '_> {
-        if self.end_offsets.is_empty() {
-            return Box::new(std::iter::empty());
-        }
-        let n = self.end_offsets.len();
-        Box::new(std::iter::once(0u32).chain(self.end_offsets[..n - 1].iter().copied()))
-    }
-}
-
-impl MultivaluedIndexBuilder {
-    pub fn finish(&mut self, num_docs: DocId) -> impl MultiValueIndexInfo + '_ {
-        self.end_values
-            .resize(num_docs as usize, self.total_num_vals_seen);
-        MultivaluedValueArrayIndex {
-            end_offsets: &self.end_values[..],
-        }
-    }
-
-    fn reset(&mut self) {
-        self.end_values.clear();
-        self.total_num_vals_seen = 0;
-    }
-}
-
-impl IndexBuilder for MultivaluedIndexBuilder {
-    fn record_doc(&mut self, doc: DocId) {
-        self.end_values
-            .resize(doc as usize, self.total_num_vals_seen);
-    }
-
-    fn record_value(&mut self) {
-        self.total_num_vals_seen += 1;
-    }
-}
-
-/// The `SpareIndexBuilders` is there to avoid allocating a
-/// new index builder for every single column.
-#[derive(Default)]
-pub struct SpareIndexBuilders {
-    required_index_builder: RequiredIndexBuilder,
-    optional_index_builder: OptionalIndexBuilder,
-    multivalued_index_builder: MultivaluedIndexBuilder,
-}
-
-impl SpareIndexBuilders {
-    pub fn borrow_required_index_builder(&mut self) -> &mut RequiredIndexBuilder {
-        &mut self.required_index_builder
-    }
-
-    pub fn borrow_optional_index_builder(&mut self) -> &mut OptionalIndexBuilder {
-        self.optional_index_builder.reset();
-        &mut self.optional_index_builder
-    }
-
-    pub fn borrow_multivalued_index_builder(&mut self) -> &mut MultivaluedIndexBuilder {
-        self.multivalued_index_builder.reset();
-        &mut self.multivalued_index_builder
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_optional_value_index_builder() {
-        let mut opt_value_index_builder = OptionalIndexBuilder::default();
-        opt_value_index_builder.record_doc(0u32);
-        opt_value_index_builder.record_value();
-        assert_eq!(
-            &opt_value_index_builder
-                .finish(1u32)
-                .iter()
-                .collect::<Vec<u32>>(),
-            &[0]
-        );
-        opt_value_index_builder.reset();
-        opt_value_index_builder.record_doc(1u32);
-        opt_value_index_builder.record_value();
-        assert_eq!(
-            &opt_value_index_builder
-                .finish(2u32)
-                .iter()
-                .collect::<Vec<u32>>(),
-            &[1]
-        );
-    }
-
-    #[test]
-    fn test_multivalued_value_index_builder() {
-        let mut multivalued_value_index_builder = MultivaluedIndexBuilder::default();
-        multivalued_value_index_builder.record_doc(1u32);
-        multivalued_value_index_builder.record_value();
-        multivalued_value_index_builder.record_value();
-        multivalued_value_index_builder.record_doc(2u32);
-        multivalued_value_index_builder.record_value();
-        assert_eq!(
-            multivalued_value_index_builder
-                .finish(4u32)
-                .iter()
-                .collect::<Vec<u32>>(),
-            vec![0, 0, 2, 3]
-        );
-        multivalued_value_index_builder.reset();
-        multivalued_value_index_builder.record_doc(2u32);
-        multivalued_value_index_builder.record_value();
-        multivalued_value_index_builder.record_value();
-        assert_eq!(
-            multivalued_value_index_builder
-                .finish(4u32)
-                .iter()
-                .collect::<Vec<u32>>(),
-            vec![0, 0, 0, 2]
-        );
-    }
-}

common/Cargo.toml

@@ -1,22 +0,0 @@
-[package]
-name = "tantivy-common"
-version = "0.5.0"
-authors = ["Paul Masurel <paul@quickwit.io>", "Pascal Seitz <pascal@quickwit.io>"]
-license = "MIT"
-edition = "2021"
-description = "common traits and utility functions used by multiple tantivy subcrates"
-documentation = "https://docs.rs/tantivy_common/"
-homepage = "https://github.com/quickwit-oss/tantivy"
-repository = "https://github.com/quickwit-oss/tantivy"
-
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-byteorder = "1.4.3"
-ownedbytes = { version= "0.5", path="../ownedbytes" }
-async-trait = "0.1"
-
-[dev-dependencies]
-proptest = "1.0.0"
-rand = "0.8.4"

common/src/bitset.rs

@@ -1,737 +0,0 @@
-use std::convert::TryInto;
-use std::io::Write;
-use std::{fmt, io, u64};
-
-use ownedbytes::OwnedBytes;
-
-#[derive(Clone, Copy, Eq, PartialEq)]
-pub struct TinySet(u64);
-
-impl fmt::Debug for TinySet {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        self.into_iter().collect::<Vec<u32>>().fmt(f)
-    }
-}
-
-pub struct TinySetIterator(TinySet);
-impl Iterator for TinySetIterator {
-    type Item = u32;
-
-    #[inline]
-    fn next(&mut self) -> Option<Self::Item> {
-        self.0.pop_lowest()
-    }
-}
-
-impl IntoIterator for TinySet {
-    type Item = u32;
-    type IntoIter = TinySetIterator;
-    fn into_iter(self) -> Self::IntoIter {
-        TinySetIterator(self)
-    }
-}
-
-impl TinySet {
-    pub fn serialize<T: Write>(&self, writer: &mut T) -> io::Result<()> {
-        writer.write_all(self.0.to_le_bytes().as_ref())
-    }
-
-    pub fn into_bytes(self) -> [u8; 8] {
-        self.0.to_le_bytes()
-    }
-
-    #[inline]
-    pub fn deserialize(data: [u8; 8]) -> Self {
-        let val: u64 = u64::from_le_bytes(data);
-        TinySet(val)
-    }
-
-    /// Returns an empty `TinySet`.
-    #[inline]
-    pub fn empty() -> TinySet {
-        TinySet(0u64)
-    }
-
-    /// Returns a full `TinySet`.
-    #[inline]
-    pub fn full() -> TinySet {
-        TinySet::empty().complement()
-    }
-
-    pub fn clear(&mut self) {
-        self.0 = 0u64;
-    }
-
-    /// Returns the complement of the set in `[0, 64[`.
-    ///
-    /// Careful on making this function public, as it will break the padding handling in the last
-    /// bucket.
-    #[inline]
-    fn complement(self) -> TinySet {
-        TinySet(!self.0)
-    }
-
-    /// Returns true iff the `TinySet` contains the element `el`.
-    #[inline]
-    pub fn contains(self, el: u32) -> bool {
-        !self.intersect(TinySet::singleton(el)).is_empty()
-    }
-
-    /// Returns the number of elements in the TinySet.
-    #[inline]
-    pub fn len(self) -> u32 {
-        self.0.count_ones()
-    }
-
-    /// Returns the intersection of `self` and `other`
-    #[inline]
-    #[must_use]
-    pub fn intersect(self, other: TinySet) -> TinySet {
-        TinySet(self.0 & other.0)
-    }
-
-    /// Creates a new `TinySet` containing only one element
-    /// within `[0; 64[`
-    #[inline]
-    pub fn singleton(el: u32) -> TinySet {
-        TinySet(1u64 << u64::from(el))
-    }
-
-    /// Insert a new element within [0..64)
-    #[inline]
-    #[must_use]
-    pub fn insert(self, el: u32) -> TinySet {
-        self.union(TinySet::singleton(el))
-    }
-
-    /// Removes an element within [0..64)
-    #[inline]
-    #[must_use]
-    pub fn remove(self, el: u32) -> TinySet {
-        self.intersect(TinySet::singleton(el).complement())
-    }
-
-    /// Insert a new element within [0..64)
-    ///
-    /// returns true if the set changed
-    #[inline]
-    pub fn insert_mut(&mut self, el: u32) -> bool {
-        let old = *self;
-        *self = old.insert(el);
-        old != *self
-    }
-
-    /// Remove a element within [0..64)
-    ///
-    /// returns true if the set changed
-    #[inline]
-    pub fn remove_mut(&mut self, el: u32) -> bool {
-        let old = *self;
-        *self = old.remove(el);
-        old != *self
-    }
-
-    /// Returns the union of two tinysets
-    #[inline]
-    #[must_use]
-    pub fn union(self, other: TinySet) -> TinySet {
-        TinySet(self.0 | other.0)
-    }
-
-    /// Returns true iff the `TinySet` is empty.
-    #[inline]
-    pub fn is_empty(self) -> bool {
-        self.0 == 0u64
-    }
-
-    /// Returns the lowest element in the `TinySet`
-    /// and removes it.
-    #[inline]
-    pub fn pop_lowest(&mut self) -> Option<u32> {
-        if self.is_empty() {
-            None
-        } else {
-            let lowest = self.0.trailing_zeros();
-            self.0 ^= TinySet::singleton(lowest).0;
-            Some(lowest)
-        }
-    }
-
-    /// Returns a `TinySet` than contains all values up
-    /// to limit excluded.
-    ///
-    /// The limit is assumed to be strictly lower than 64.
-    pub fn range_lower(upper_bound: u32) -> TinySet {
-        TinySet((1u64 << u64::from(upper_bound % 64u32)) - 1u64)
-    }
-
-    /// Returns a `TinySet` that contains all values greater
-    /// or equal to the given limit, included. (and up to 63)
-    ///
-    /// The limit is assumed to be strictly lower than 64.
-    pub fn range_greater_or_equal(from_included: u32) -> TinySet {
-        TinySet::range_lower(from_included).complement()
-    }
-}
-
-#[derive(Clone)]
-pub struct BitSet {
-    tinysets: Box<[TinySet]>,
-    len: u64,
-    max_value: u32,
-}
-
-fn num_buckets(max_val: u32) -> u32 {
-    (max_val + 63u32) / 64u32
-}
-
-impl BitSet {
-    /// serialize a `BitSet`.
-    pub fn serialize<T: Write>(&self, writer: &mut T) -> io::Result<()> {
-        writer.write_all(self.max_value.to_le_bytes().as_ref())?;
-        for tinyset in self.tinysets.iter().cloned() {
-            writer.write_all(&tinyset.into_bytes())?;
-        }
-        writer.flush()?;
-        Ok(())
-    }
-
-    /// Create a new `BitSet` that may contain elements
-    /// within `[0, max_val)`.
-    pub fn with_max_value(max_value: u32) -> BitSet {
-        let num_buckets = num_buckets(max_value);
-        let tinybitsets = vec![TinySet::empty(); num_buckets as usize].into_boxed_slice();
-        BitSet {
-            tinysets: tinybitsets,
-            len: 0,
-            max_value,
-        }
-    }
-
-    /// Create a new `BitSet` that may contain elements. Initially all values will be set.
-    /// within `[0, max_val)`.
-    pub fn with_max_value_and_full(max_value: u32) -> BitSet {
-        let num_buckets = num_buckets(max_value);
-        let mut tinybitsets = vec![TinySet::full(); num_buckets as usize].into_boxed_slice();
-
-        // Fix padding
-        let lower = max_value % 64u32;
-        if lower != 0 {
-            tinybitsets[tinybitsets.len() - 1] = TinySet::range_lower(lower);
-        }
-        BitSet {
-            tinysets: tinybitsets,
-            len: max_value as u64,
-            max_value,
-        }
-    }
-
-    /// Removes all elements from the `BitSet`.
-    pub fn clear(&mut self) {
-        for tinyset in self.tinysets.iter_mut() {
-            *tinyset = TinySet::empty();
-        }
-    }
-
-    /// Intersect with serialized bitset
-    pub fn intersect_update(&mut self, other: &ReadOnlyBitSet) {
-        self.intersect_update_with_iter(other.iter_tinysets());
-    }
-
-    /// Intersect with tinysets
-    fn intersect_update_with_iter(&mut self, other: impl Iterator<Item = TinySet>) {
-        self.len = 0;
-        for (left, right) in self.tinysets.iter_mut().zip(other) {
-            *left = left.intersect(right);
-            self.len += left.len() as u64;
-        }
-    }
-
-    /// Returns the number of elements in the `BitSet`.
-    #[inline]
-    pub fn len(&self) -> usize {
-        self.len as usize
-    }
-
-    /// Inserts an element in the `BitSet`
-    #[inline]
-    pub fn insert(&mut self, el: u32) {
-        // we do not check saturated els.
-        let higher = el / 64u32;
-        let lower = el % 64u32;
-        self.len += u64::from(self.tinysets[higher as usize].insert_mut(lower));
-    }
-
-    /// Inserts an element in the `BitSet`
-    #[inline]
-    pub fn remove(&mut self, el: u32) {
-        // we do not check saturated els.
-        let higher = el / 64u32;
-        let lower = el % 64u32;
-        self.len -= u64::from(self.tinysets[higher as usize].remove_mut(lower));
-    }
-
-    /// Returns true iff the elements is in the `BitSet`.
-    #[inline]
-    pub fn contains(&self, el: u32) -> bool {
-        self.tinyset(el / 64u32).contains(el % 64)
-    }
-
-    /// Returns the first non-empty `TinySet` associated with a bucket lower
-    /// or greater than bucket.
-    ///
-    /// Reminder: the tiny set with the bucket `bucket`, represents the
-    /// elements from `bucket * 64` to `(bucket+1) * 64`.
-    pub fn first_non_empty_bucket(&self, bucket: u32) -> Option<u32> {
-        self.tinysets[bucket as usize..]
-            .iter()
-            .cloned()
-            .position(|tinyset| !tinyset.is_empty())
-            .map(|delta_bucket| bucket + delta_bucket as u32)
-    }
-
-    #[inline]
-    pub fn max_value(&self) -> u32 {
-        self.max_value
-    }
-
-    /// Returns the tiny bitset representing the
-    /// the set restricted to the number range from
-    /// `bucket * 64` to `(bucket + 1) * 64`.
-    pub fn tinyset(&self, bucket: u32) -> TinySet {
-        self.tinysets[bucket as usize]
-    }
-}
-
-/// Serialized BitSet.
-#[derive(Clone)]
-pub struct ReadOnlyBitSet {
-    data: OwnedBytes,
-    max_value: u32,
-}
-
-pub fn intersect_bitsets(left: &ReadOnlyBitSet, other: &ReadOnlyBitSet) -> ReadOnlyBitSet {
-    assert_eq!(left.max_value(), other.max_value());
-    assert_eq!(left.data.len(), other.data.len());
-    let union_tinyset_it = left
-        .iter_tinysets()
-        .zip(other.iter_tinysets())
-        .map(|(left_tinyset, right_tinyset)| left_tinyset.intersect(right_tinyset));
-    let mut output_dataset: Vec<u8> = Vec::with_capacity(left.data.len());
-    for tinyset in union_tinyset_it {
-        output_dataset.extend_from_slice(&tinyset.into_bytes());
-    }
-    ReadOnlyBitSet {
-        data: OwnedBytes::new(output_dataset),
-        max_value: left.max_value(),
-    }
-}
-
-impl ReadOnlyBitSet {
-    pub fn open(data: OwnedBytes) -> Self {
-        let (max_value_data, data) = data.split(4);
-        assert_eq!(data.len() % 8, 0);
-        let max_value: u32 = u32::from_le_bytes(max_value_data.as_ref().try_into().unwrap());
-        ReadOnlyBitSet { data, max_value }
-    }
-
-    /// Number of elements in the bitset.
-    #[inline]
-    pub fn len(&self) -> usize {
-        self.iter_tinysets()
-            .map(|tinyset| tinyset.len() as usize)
-            .sum()
-    }
-
-    /// Iterate the tinyset on the fly from serialized data.
-    #[inline]
-    fn iter_tinysets(&self) -> impl Iterator<Item = TinySet> + '_ {
-        self.data.chunks_exact(8).map(move |chunk| {
-            let tinyset: TinySet = TinySet::deserialize(chunk.try_into().unwrap());
-            tinyset
-        })
-    }
-
-    /// Iterate over the positions of the elements.
-    #[inline]
-    pub fn iter(&self) -> impl Iterator<Item = u32> + '_ {
-        self.iter_tinysets()
-            .enumerate()
-            .flat_map(move |(chunk_num, tinyset)| {
-                let chunk_base_val = chunk_num as u32 * 64;
-                tinyset
-                    .into_iter()
-                    .map(move |val| val + chunk_base_val)
-                    .take_while(move |doc| *doc < self.max_value)
-            })
-    }
-
-    /// Returns true iff the elements is in the `BitSet`.
-    #[inline]
-    pub fn contains(&self, el: u32) -> bool {
-        let byte_offset = el / 8u32;
-        let b: u8 = self.data[byte_offset as usize];
-        let shift = (el % 8) as u8;
-        b & (1u8 << shift) != 0
-    }
-
-    /// Maximum value the bitset may contain.
-    /// (Note this is not the maximum value contained in the set.)
-    ///
-    /// A bitset has an intrinsic capacity.
-    /// It only stores elements within [0..max_value).
-    #[inline]
-    pub fn max_value(&self) -> u32 {
-        self.max_value
-    }
-
-    /// Number of bytes used in the bitset representation.
-    pub fn num_bytes(&self) -> usize {
-        self.data.len()
-    }
-}
-
-impl<'a> From<&'a BitSet> for ReadOnlyBitSet {
-    fn from(bitset: &'a BitSet) -> ReadOnlyBitSet {
-        let mut buffer = Vec::with_capacity(bitset.tinysets.len() * 8 + 4);
-        bitset
-            .serialize(&mut buffer)
-            .expect("serializing into a buffer should never fail");
-        ReadOnlyBitSet::open(OwnedBytes::new(buffer))
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use std::collections::HashSet;
-
-    use ownedbytes::OwnedBytes;
-    use rand::distributions::Bernoulli;
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
-
-    use super::{BitSet, ReadOnlyBitSet, TinySet};
-
-    #[test]
-    fn test_read_serialized_bitset_full_multi() {
-        for i in 0..1000 {
-            let bitset = BitSet::with_max_value_and_full(i);
-            let mut out = vec![];
-            bitset.serialize(&mut out).unwrap();
-
-            let bitset = ReadOnlyBitSet::open(OwnedBytes::new(out));
-            assert_eq!(bitset.len(), i as usize);
-        }
-    }
-
-    #[test]
-    fn test_read_serialized_bitset_full_block() {
-        let bitset = BitSet::with_max_value_and_full(64);
-        let mut out = vec![];
-        bitset.serialize(&mut out).unwrap();
-
-        let bitset = ReadOnlyBitSet::open(OwnedBytes::new(out));
-        assert_eq!(bitset.len(), 64);
-    }
-
-    #[test]
-    fn test_read_serialized_bitset_full() {
-        let mut bitset = BitSet::with_max_value_and_full(5);
-        bitset.remove(3);
-        let mut out = vec![];
-        bitset.serialize(&mut out).unwrap();
-
-        let bitset = ReadOnlyBitSet::open(OwnedBytes::new(out));
-        assert_eq!(bitset.len(), 4);
-    }
-
-    #[test]
-    fn test_bitset_intersect() {
-        let bitset_serialized = {
-            let mut bitset = BitSet::with_max_value_and_full(5);
-            bitset.remove(1);
-            bitset.remove(3);
-            let mut out = vec![];
-            bitset.serialize(&mut out).unwrap();
-
-            ReadOnlyBitSet::open(OwnedBytes::new(out))
-        };
-
-        let mut bitset = BitSet::with_max_value_and_full(5);
-        bitset.remove(1);
-        bitset.intersect_update(&bitset_serialized);
-
-        assert!(bitset.contains(0));
-        assert!(!bitset.contains(1));
-        assert!(bitset.contains(2));
-        assert!(!bitset.contains(3));
-        assert!(bitset.contains(4));
-
-        bitset.intersect_update_with_iter(vec![TinySet::singleton(0)].into_iter());
-
-        assert!(bitset.contains(0));
-        assert!(!bitset.contains(1));
-        assert!(!bitset.contains(2));
-        assert!(!bitset.contains(3));
-        assert!(!bitset.contains(4));
-        assert_eq!(bitset.len(), 1);
-
-        bitset.intersect_update_with_iter(vec![TinySet::singleton(1)].into_iter());
-        assert!(!bitset.contains(0));
-        assert!(!bitset.contains(1));
-        assert!(!bitset.contains(2));
-        assert!(!bitset.contains(3));
-        assert!(!bitset.contains(4));
-        assert_eq!(bitset.len(), 0);
-    }
-
-    #[test]
-    fn test_read_serialized_bitset_empty() {
-        let mut bitset = BitSet::with_max_value(5);
-        bitset.insert(3);
-        let mut out = vec![];
-        bitset.serialize(&mut out).unwrap();
-
-        let bitset = ReadOnlyBitSet::open(OwnedBytes::new(out));
-        assert_eq!(bitset.len(), 1);
-
-        {
-            let bitset = BitSet::with_max_value(5);
-            let mut out = vec![];
-            bitset.serialize(&mut out).unwrap();
-            let bitset = ReadOnlyBitSet::open(OwnedBytes::new(out));
-            assert_eq!(bitset.len(), 0);
-        }
-    }
-
-    #[test]
-    fn test_tiny_set_remove() {
-        {
-            let mut u = TinySet::empty().insert(63u32).insert(5).remove(63u32);
-            assert_eq!(u.pop_lowest(), Some(5u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let mut u = TinySet::empty()
-                .insert(63u32)
-                .insert(1)
-                .insert(5)
-                .remove(63u32);
-            assert_eq!(u.pop_lowest(), Some(1u32));
-            assert_eq!(u.pop_lowest(), Some(5u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let mut u = TinySet::empty().insert(1).remove(63u32);
-            assert_eq!(u.pop_lowest(), Some(1u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let mut u = TinySet::empty().insert(1).remove(1u32);
-            assert!(u.pop_lowest().is_none());
-        }
-    }
-    #[test]
-    fn test_tiny_set() {
-        assert!(TinySet::empty().is_empty());
-        {
-            let mut u = TinySet::empty().insert(1u32);
-            assert_eq!(u.pop_lowest(), Some(1u32));
-            assert!(u.pop_lowest().is_none())
-        }
-        {
-            let mut u = TinySet::empty().insert(1u32).insert(1u32);
-            assert_eq!(u.pop_lowest(), Some(1u32));
-            assert!(u.pop_lowest().is_none())
-        }
-        {
-            let mut u = TinySet::empty().insert(2u32);
-            assert_eq!(u.pop_lowest(), Some(2u32));
-            u.insert_mut(1u32);
-            assert_eq!(u.pop_lowest(), Some(1u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let mut u = TinySet::empty().insert(63u32);
-            assert_eq!(u.pop_lowest(), Some(63u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let mut u = TinySet::empty().insert(63u32).insert(5);
-            assert_eq!(u.pop_lowest(), Some(5u32));
-            assert_eq!(u.pop_lowest(), Some(63u32));
-            assert!(u.pop_lowest().is_none());
-        }
-        {
-            let original = TinySet::empty().insert(63u32).insert(5);
-            let after_serialize_deserialize = TinySet::deserialize(original.into_bytes());
-            assert_eq!(original, after_serialize_deserialize);
-        }
-    }
-
-    #[test]
-    fn test_bitset() {
-        let test_against_hashset = |els: &[u32], max_value: u32| {
-            let mut hashset: HashSet<u32> = HashSet::new();
-            let mut bitset = BitSet::with_max_value(max_value);
-            for &el in els {
-                assert!(el < max_value);
-                hashset.insert(el);
-                bitset.insert(el);
-            }
-            for el in 0..max_value {
-                assert_eq!(hashset.contains(&el), bitset.contains(el));
-            }
-            assert_eq!(bitset.max_value(), max_value);
-
-            // test deser
-            let mut data = vec![];
-            bitset.serialize(&mut data).unwrap();
-            let ro_bitset = ReadOnlyBitSet::open(OwnedBytes::new(data));
-            for el in 0..max_value {
-                assert_eq!(hashset.contains(&el), ro_bitset.contains(el));
-            }
-            assert_eq!(ro_bitset.max_value(), max_value);
-            assert_eq!(ro_bitset.len(), els.len());
-        };
-
-        test_against_hashset(&[], 0);
-        test_against_hashset(&[], 1);
-        test_against_hashset(&[0u32], 1);
-        test_against_hashset(&[0u32], 100);
-        test_against_hashset(&[1u32, 2u32], 4);
-        test_against_hashset(&[99u32], 100);
-        test_against_hashset(&[63u32], 64);
-        test_against_hashset(&[62u32, 63u32], 64);
-    }
-
-    #[test]
-    fn test_bitset_num_buckets() {
-        use super::num_buckets;
-        assert_eq!(num_buckets(0u32), 0);
-        assert_eq!(num_buckets(1u32), 1);
-        assert_eq!(num_buckets(64u32), 1);
-        assert_eq!(num_buckets(65u32), 2);
-        assert_eq!(num_buckets(128u32), 2);
-        assert_eq!(num_buckets(129u32), 3);
-    }
-
-    #[test]
-    fn test_tinyset_range() {
-        assert_eq!(
-            TinySet::range_lower(3).into_iter().collect::<Vec<u32>>(),
-            [0, 1, 2]
-        );
-        assert!(TinySet::range_lower(0).is_empty());
-        assert_eq!(
-            TinySet::range_lower(63).into_iter().collect::<Vec<u32>>(),
-            (0u32..63u32).collect::<Vec<_>>()
-        );
-        assert_eq!(
-            TinySet::range_lower(1).into_iter().collect::<Vec<u32>>(),
-            [0]
-        );
-        assert_eq!(
-            TinySet::range_lower(2).into_iter().collect::<Vec<u32>>(),
-            [0, 1]
-        );
-        assert_eq!(
-            TinySet::range_greater_or_equal(3)
-                .into_iter()
-                .collect::<Vec<u32>>(),
-            (3u32..64u32).collect::<Vec<_>>()
-        );
-    }
-
-    #[test]
-    fn test_bitset_len() {
-        let mut bitset = BitSet::with_max_value(1_000);
-        assert_eq!(bitset.len(), 0);
-        bitset.insert(3u32);
-        assert_eq!(bitset.len(), 1);
-        bitset.insert(103u32);
-        assert_eq!(bitset.len(), 2);
-        bitset.insert(3u32);
-        assert_eq!(bitset.len(), 2);
-        bitset.insert(103u32);
-        assert_eq!(bitset.len(), 2);
-        bitset.insert(104u32);
-        assert_eq!(bitset.len(), 3);
-        bitset.remove(105u32);
-        assert_eq!(bitset.len(), 3);
-        bitset.remove(104u32);
-        assert_eq!(bitset.len(), 2);
-        bitset.remove(3u32);
-        assert_eq!(bitset.len(), 1);
-        bitset.remove(103u32);
-        assert_eq!(bitset.len(), 0);
-    }
-
-    pub fn sample_with_seed(n: u32, ratio: f64, seed_val: u8) -> Vec<u32> {
-        StdRng::from_seed([seed_val; 32])
-            .sample_iter(&Bernoulli::new(ratio).unwrap())
-            .take(n as usize)
-            .enumerate()
-            .filter_map(|(val, keep)| if keep { Some(val as u32) } else { None })
-            .collect()
-    }
-
-    pub fn sample(n: u32, ratio: f64) -> Vec<u32> {
-        sample_with_seed(n, ratio, 4)
-    }
-
-    #[test]
-    fn test_bitset_clear() {
-        let mut bitset = BitSet::with_max_value(1_000);
-        let els = sample(1_000, 0.01f64);
-        for &el in &els {
-            bitset.insert(el);
-        }
-        assert!(els.iter().all(|el| bitset.contains(*el)));
-        bitset.clear();
-        for el in 0u32..1000u32 {
-            assert!(!bitset.contains(el));
-        }
-    }
-}
-
-#[cfg(all(test, feature = "unstable"))]
-mod bench {
-
-    use test;
-
-    use super::{BitSet, TinySet};
-
-    #[bench]
-    fn bench_tinyset_pop(b: &mut test::Bencher) {
-        b.iter(|| {
-            let mut tinyset = TinySet::singleton(test::black_box(31u32));
-            tinyset.pop_lowest();
-            tinyset.pop_lowest();
-            tinyset.pop_lowest();
-            tinyset.pop_lowest();
-            tinyset.pop_lowest();
-            tinyset.pop_lowest();
-        });
-    }
-
-    #[bench]
-    fn bench_tinyset_sum(b: &mut test::Bencher) {
-        let tiny_set = TinySet::empty().insert(10u32).insert(14u32).insert(21u32);
-        b.iter(|| {
-            assert_eq!(test::black_box(tiny_set).into_iter().sum::<u32>(), 45u32);
-        });
-    }
-
-    #[bench]
-    fn bench_tinyarr_sum(b: &mut test::Bencher) {
-        let v = [10u32, 14u32, 21u32];
-        b.iter(|| test::black_box(v).iter().cloned().sum::<u32>());
-    }
-
-    #[bench]
-    fn bench_bitset_initialize(b: &mut test::Bencher) {
-        b.iter(|| BitSet::with_max_value(1_000_000));
-    }
-}

common/src/file_slice.rs

@@ -1,333 +0,0 @@
-use std::ops::{Deref, Range, RangeBounds};
-use std::sync::Arc;
-use std::{fmt, io};
-
-use async_trait::async_trait;
-use ownedbytes::{OwnedBytes, StableDeref};
-
-use crate::HasLen;
-
-/// Objects that represents files sections in tantivy.
-///
-/// By contract, whatever happens to the directory file, as long as a FileHandle
-/// is alive, the data associated with it cannot be altered or destroyed.
-///
-/// The underlying behavior is therefore specific to the `Directory` that
-/// created it. Despite its name, a [`FileSlice`] may or may not directly map to an actual file
-/// on the filesystem.
-
-#[async_trait]
-pub trait FileHandle: 'static + Send + Sync + HasLen + fmt::Debug {
-    /// Reads a slice of bytes.
-    ///
-    /// This method may panic if the range requested is invalid.
-    fn read_bytes(&self, range: Range<usize>) -> io::Result<OwnedBytes>;
-
-    #[doc(hidden)]
-    async fn read_bytes_async(&self, byte_range: Range<usize>) -> io::Result<OwnedBytes> {
-        self.read_bytes(byte_range)
-    }
-}
-
-#[async_trait]
-impl FileHandle for &'static [u8] {
-    fn read_bytes(&self, range: Range<usize>) -> io::Result<OwnedBytes> {
-        let bytes = &self[range];
-        Ok(OwnedBytes::new(bytes))
-    }
-
-    #[cfg(feature = "quickwit")]
-    async fn read_bytes_async(&self, byte_range: Range<usize>) -> io::Result<OwnedBytes> {
-        Ok(self.read_bytes(byte_range)?)
-    }
-}
-
-impl<B> From<B> for FileSlice
-where B: StableDeref + Deref<Target = [u8]> + 'static + Send + Sync
-{
-    fn from(bytes: B) -> FileSlice {
-        FileSlice::new(Arc::new(OwnedBytes::new(bytes)))
-    }
-}
-
-/// Logical slice of read only file in tantivy.
-///
-/// It can be cloned and sliced cheaply.
-#[derive(Clone)]
-pub struct FileSlice {
-    data: Arc<dyn FileHandle>,
-    range: Range<usize>,
-}
-
-impl fmt::Debug for FileSlice {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        write!(f, "FileSlice({:?}, {:?})", &self.data, self.range)
-    }
-}
-
-#[inline]
-fn combine_ranges<R: RangeBounds<usize>>(orig_range: Range<usize>, rel_range: R) -> Range<usize> {
-    let start: usize = orig_range.start
-        + match rel_range.start_bound().cloned() {
-            std::ops::Bound::Included(rel_start) => rel_start,
-            std::ops::Bound::Excluded(rel_start) => rel_start + 1,
-            std::ops::Bound::Unbounded => 0,
-        };
-    assert!(start <= orig_range.end);
-    let end: usize = match rel_range.end_bound().cloned() {
-        std::ops::Bound::Included(rel_end) => orig_range.start + rel_end + 1,
-        std::ops::Bound::Excluded(rel_end) => orig_range.start + rel_end,
-        std::ops::Bound::Unbounded => orig_range.end,
-    };
-    assert!(end >= start);
-    assert!(end <= orig_range.end);
-    start..end
-}
-
-impl FileSlice {
-    /// Wraps a FileHandle.
-    pub fn new(file_handle: Arc<dyn FileHandle>) -> Self {
-        let num_bytes = file_handle.len();
-        FileSlice::new_with_num_bytes(file_handle, num_bytes)
-    }
-
-    /// Wraps a FileHandle.
-    #[doc(hidden)]
-    #[must_use]
-    pub fn new_with_num_bytes(file_handle: Arc<dyn FileHandle>, num_bytes: usize) -> Self {
-        FileSlice {
-            data: file_handle,
-            range: 0..num_bytes,
-        }
-    }
-
-    /// Creates a fileslice that is just a view over a slice of the data.
-    ///
-    /// # Panics
-    ///
-    /// Panics if `byte_range.end` exceeds the filesize.
-    #[must_use]
-    #[inline]
-    pub fn slice<R: RangeBounds<usize>>(&self, byte_range: R) -> FileSlice {
-        FileSlice {
-            data: self.data.clone(),
-            range: combine_ranges(self.range.clone(), byte_range),
-        }
-    }
-
-    /// Creates an empty FileSlice
-    pub fn empty() -> FileSlice {
-        const EMPTY_SLICE: &[u8] = &[];
-        FileSlice::from(EMPTY_SLICE)
-    }
-
-    /// Returns a `OwnedBytes` with all of the data in the `FileSlice`.
-    ///
-    /// The behavior is strongly dependent on the implementation of the underlying
-    /// `Directory` and the `FileSliceTrait` it creates.
-    /// In particular, it is  up to the `Directory` implementation
-    /// to handle caching if needed.
-    pub fn read_bytes(&self) -> io::Result<OwnedBytes> {
-        self.data.read_bytes(self.range.clone())
-    }
-
-    #[doc(hidden)]
-    pub async fn read_bytes_async(&self) -> io::Result<OwnedBytes> {
-        self.data.read_bytes_async(self.range.clone()).await
-    }
-
-    /// Reads a specific slice of data.
-    ///
-    /// This is equivalent to running `file_slice.slice(from, to).read_bytes()`.
-    pub fn read_bytes_slice(&self, range: Range<usize>) -> io::Result<OwnedBytes> {
-        assert!(
-            range.end <= self.len(),
-            "end of requested range exceeds the fileslice length ({} > {})",
-            range.end,
-            self.len()
-        );
-        self.data
-            .read_bytes(self.range.start + range.start..self.range.start + range.end)
-    }
-
-    #[doc(hidden)]
-    pub async fn read_bytes_slice_async(&self, byte_range: Range<usize>) -> io::Result<OwnedBytes> {
-        assert!(
-            self.range.start + byte_range.end <= self.range.end,
-            "`to` exceeds the fileslice length"
-        );
-        self.data
-            .read_bytes_async(
-                self.range.start + byte_range.start..self.range.start + byte_range.end,
-            )
-            .await
-    }
-
-    /// Splits the FileSlice at the given offset and return two file slices.
-    /// `file_slice[..split_offset]` and `file_slice[split_offset..]`.
-    ///
-    /// This operation is cheap and must not copy any underlying data.
-    pub fn split(self, left_len: usize) -> (FileSlice, FileSlice) {
-        let left = self.slice_to(left_len);
-        let right = self.slice_from(left_len);
-        (left, right)
-    }
-
-    /// Splits the file slice at the given offset and return two file slices.
-    /// `file_slice[..split_offset]` and `file_slice[split_offset..]`.
-    pub fn split_from_end(self, right_len: usize) -> (FileSlice, FileSlice) {
-        let left_len = self.len() - right_len;
-        self.split(left_len)
-    }
-
-    /// Like `.slice(...)` but enforcing only the `from`
-    /// boundary.
-    ///
-    /// Equivalent to `.slice(from_offset, self.len())`
-    #[must_use]
-    pub fn slice_from(&self, from_offset: usize) -> FileSlice {
-        self.slice(from_offset..self.len())
-    }
-
-    /// Returns a slice from the end.
-    ///
-    /// Equivalent to `.slice(self.len() - from_offset, self.len())`
-    #[must_use]
-    pub fn slice_from_end(&self, from_offset: usize) -> FileSlice {
-        self.slice(self.len() - from_offset..self.len())
-    }
-
-    /// Like `.slice(...)` but enforcing only the `to`
-    /// boundary.
-    ///
-    /// Equivalent to `.slice(0, to_offset)`
-    #[must_use]
-    pub fn slice_to(&self, to_offset: usize) -> FileSlice {
-        self.slice(0..to_offset)
-    }
-}
-
-#[async_trait]
-impl FileHandle for FileSlice {
-    fn read_bytes(&self, range: Range<usize>) -> io::Result<OwnedBytes> {
-        self.read_bytes_slice(range)
-    }
-
-    #[cfg(feature = "quickwit")]
-    async fn read_bytes_async(&self, byte_range: Range<usize>) -> io::Result<OwnedBytes> {
-        self.read_bytes_slice_async(byte_range).await
-    }
-}
-
-impl HasLen for FileSlice {
-    fn len(&self) -> usize {
-        self.range.len()
-    }
-}
-
-#[async_trait]
-impl FileHandle for OwnedBytes {
-    fn read_bytes(&self, range: Range<usize>) -> io::Result<OwnedBytes> {
-        Ok(self.slice(range))
-    }
-
-    #[cfg(feature = "quickwit")]
-    async fn read_bytes_async(&self, range: Range<usize>) -> io::Result<OwnedBytes> {
-        let bytes = self.read_bytes(range)?;
-        Ok(bytes)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use std::io;
-    use std::sync::Arc;
-
-    use super::{FileHandle, FileSlice};
-    use crate::file_slice::combine_ranges;
-    use crate::HasLen;
-
-    #[test]
-    fn test_file_slice() -> io::Result<()> {
-        let file_slice = FileSlice::new(Arc::new(b"abcdef".as_ref()));
-        assert_eq!(file_slice.len(), 6);
-        assert_eq!(file_slice.slice_from(2).read_bytes()?.as_slice(), b"cdef");
-        assert_eq!(file_slice.slice_to(2).read_bytes()?.as_slice(), b"ab");
-        assert_eq!(
-            file_slice
-                .slice_from(1)
-                .slice_to(2)
-                .read_bytes()?
-                .as_slice(),
-            b"bc"
-        );
-        {
-            let (left, right) = file_slice.clone().split(0);
-            assert_eq!(left.read_bytes()?.as_slice(), b"");
-            assert_eq!(right.read_bytes()?.as_slice(), b"abcdef");
-        }
-        {
-            let (left, right) = file_slice.clone().split(2);
-            assert_eq!(left.read_bytes()?.as_slice(), b"ab");
-            assert_eq!(right.read_bytes()?.as_slice(), b"cdef");
-        }
-        {
-            let (left, right) = file_slice.clone().split_from_end(0);
-            assert_eq!(left.read_bytes()?.as_slice(), b"abcdef");
-            assert_eq!(right.read_bytes()?.as_slice(), b"");
-        }
-        {
-            let (left, right) = file_slice.split_from_end(2);
-            assert_eq!(left.read_bytes()?.as_slice(), b"abcd");
-            assert_eq!(right.read_bytes()?.as_slice(), b"ef");
-        }
-        Ok(())
-    }
-
-    #[test]
-    fn test_file_slice_trait_slice_len() {
-        let blop: &'static [u8] = b"abc";
-        let owned_bytes: Box<dyn FileHandle> = Box::new(blop);
-        assert_eq!(owned_bytes.len(), 3);
-    }
-
-    #[test]
-    fn test_slice_simple_read() -> io::Result<()> {
-        let slice = FileSlice::new(Arc::new(&b"abcdef"[..]));
-        assert_eq!(slice.len(), 6);
-        assert_eq!(slice.read_bytes()?.as_ref(), b"abcdef");
-        assert_eq!(slice.slice(1..4).read_bytes()?.as_ref(), b"bcd");
-        Ok(())
-    }
-
-    #[test]
-    fn test_slice_read_slice() -> io::Result<()> {
-        let slice_deref = FileSlice::new(Arc::new(&b"abcdef"[..]));
-        assert_eq!(slice_deref.read_bytes_slice(1..4)?.as_ref(), b"bcd");
-        Ok(())
-    }
-
-    #[test]
-    #[should_panic(expected = "end of requested range exceeds the fileslice length (10 > 6)")]
-    fn test_slice_read_slice_invalid_range_exceeds() {
-        let slice_deref = FileSlice::new(Arc::new(&b"abcdef"[..]));
-        assert_eq!(
-            slice_deref.read_bytes_slice(0..10).unwrap().as_ref(),
-            b"bcd"
-        );
-    }
-
-    #[test]
-    fn test_combine_range() {
-        assert_eq!(combine_ranges(1..3, 0..1), 1..2);
-        assert_eq!(combine_ranges(1..3, 1..), 2..3);
-        assert_eq!(combine_ranges(1..4, ..2), 1..3);
-        assert_eq!(combine_ranges(3..10, 2..5), 5..8);
-    }
-
-    #[test]
-    #[should_panic]
-    fn test_combine_range_panics() {
-        let _ = combine_ranges(3..5, 1..4);
-    }
-}

common/src/lib.rs

@@ -1,172 +0,0 @@
-#![allow(clippy::len_without_is_empty)]
-
-use std::ops::Deref;
-
-pub use byteorder::LittleEndian as Endianness;
-
-mod bitset;
-pub mod file_slice;
-mod serialize;
-mod vint;
-mod writer;
-pub use bitset::*;
-pub use ownedbytes::OwnedBytes;
-pub use serialize::{BinarySerializable, DeserializeFrom, FixedSize};
-pub use vint::{
-    deserialize_vint_u128, read_u32_vint, read_u32_vint_no_advance, serialize_vint_u128,
-    serialize_vint_u32, write_u32_vint, VInt, VIntU128,
-};
-pub use writer::{AntiCallToken, CountingWriter, TerminatingWrite};
-
-/// Has length trait
-pub trait HasLen {
-    /// Return length
-    fn len(&self) -> usize;
-
-    /// Returns true iff empty.
-    fn is_empty(&self) -> bool {
-        self.len() == 0
-    }
-}
-
-impl<T: Deref<Target = [u8]>> HasLen for T {
-    fn len(&self) -> usize {
-        self.deref().len()
-    }
-}
-
-const HIGHEST_BIT: u64 = 1 << 63;
-
-/// Maps a `i64` to `u64`
-///
-/// For simplicity, tantivy internally handles `i64` as `u64`.
-/// The mapping is defined by this function.
-///
-/// Maps `i64` to `u64` so that
-/// `-2^63 .. 2^63-1` is mapped
-///     to
-/// `0 .. 2^64-1`
-/// in that order.
-///
-/// This is more suited than simply casting (`val as u64`)
-/// because of bitpacking.
-///
-/// Imagine a list of `i64` ranging from -10 to 10.
-/// When casting negative values, the negative values are projected
-/// to values over 2^63, and all values end up requiring 64 bits.
-///
-/// # See also
-/// The reverse mapping is [`u64_to_i64()`].
-#[inline]
-pub fn i64_to_u64(val: i64) -> u64 {
-    (val as u64) ^ HIGHEST_BIT
-}
-
-/// Reverse the mapping given by [`i64_to_u64()`].
-#[inline]
-pub fn u64_to_i64(val: u64) -> i64 {
-    (val ^ HIGHEST_BIT) as i64
-}
-
-/// Maps a `f64` to `u64`
-///
-/// For simplicity, tantivy internally handles `f64` as `u64`.
-/// The mapping is defined by this function.
-///
-/// Maps `f64` to `u64` in a monotonic manner, so that bytes lexical order is preserved.
-///
-/// This is more suited than simply casting (`val as u64`)
-/// which would truncate the result
-///
-/// # Reference
-///
-/// Daniel Lemire's [blog post](https://lemire.me/blog/2020/12/14/converting-floating-point-numbers-to-integers-while-preserving-order/)
-/// explains the mapping in a clear manner.
-///
-/// # See also
-/// The reverse mapping is [`u64_to_f64()`].
-#[inline]
-pub fn f64_to_u64(val: f64) -> u64 {
-    let bits = val.to_bits();
-    if val.is_sign_positive() {
-        bits ^ HIGHEST_BIT
-    } else {
-        !bits
-    }
-}
-
-/// Reverse the mapping given by [`f64_to_u64()`].
-#[inline]
-pub fn u64_to_f64(val: u64) -> f64 {
-    f64::from_bits(if val & HIGHEST_BIT != 0 {
-        val ^ HIGHEST_BIT
-    } else {
-        !val
-    })
-}
-
-#[cfg(test)]
-pub mod test {
-
-    use proptest::prelude::*;
-
-    use super::{f64_to_u64, i64_to_u64, u64_to_f64, u64_to_i64, BinarySerializable, FixedSize};
-
-    fn test_i64_converter_helper(val: i64) {
-        assert_eq!(u64_to_i64(i64_to_u64(val)), val);
-    }
-
-    fn test_f64_converter_helper(val: f64) {
-        assert_eq!(u64_to_f64(f64_to_u64(val)), val);
-    }
-
-    pub fn fixed_size_test<O: BinarySerializable + FixedSize + Default>() {
-        let mut buffer = Vec::new();
-        O::default().serialize(&mut buffer).unwrap();
-        assert_eq!(buffer.len(), O::SIZE_IN_BYTES);
-    }
-
-    proptest! {
-        #[test]
-        fn test_f64_converter_monotonicity_proptest((left, right) in (proptest::num::f64::NORMAL, proptest::num::f64::NORMAL)) {
-            let left_u64 = f64_to_u64(left);
-            let right_u64 = f64_to_u64(right);
-            assert_eq!(left_u64 < right_u64,  left < right);
-        }
-    }
-
-    #[test]
-    fn test_i64_converter() {
-        assert_eq!(i64_to_u64(i64::MIN), u64::MIN);
-        assert_eq!(i64_to_u64(i64::MAX), u64::MAX);
-        test_i64_converter_helper(0i64);
-        test_i64_converter_helper(i64::MIN);
-        test_i64_converter_helper(i64::MAX);
-        for i in -1000i64..1000i64 {
-            test_i64_converter_helper(i);
-        }
-    }
-
-    #[test]
-    fn test_f64_converter() {
-        test_f64_converter_helper(f64::INFINITY);
-        test_f64_converter_helper(f64::NEG_INFINITY);
-        test_f64_converter_helper(0.0);
-        test_f64_converter_helper(-0.0);
-        test_f64_converter_helper(1.0);
-        test_f64_converter_helper(-1.0);
-    }
-
-    #[test]
-    fn test_f64_order() {
-        assert!(!(f64_to_u64(f64::NEG_INFINITY)..f64_to_u64(f64::INFINITY))
-            .contains(&f64_to_u64(f64::NAN))); // nan is not a number
-        assert!(f64_to_u64(1.5) > f64_to_u64(1.0)); // same exponent, different mantissa
-        assert!(f64_to_u64(2.0) > f64_to_u64(1.0)); // same mantissa, different exponent
-        assert!(f64_to_u64(2.0) > f64_to_u64(1.5)); // different exponent and mantissa
-        assert!(f64_to_u64(1.0) > f64_to_u64(-1.0)); // pos > neg
-        assert!(f64_to_u64(-1.5) < f64_to_u64(-1.0));
-        assert!(f64_to_u64(-2.0) < f64_to_u64(1.0));
-        assert!(f64_to_u64(-2.0) < f64_to_u64(-1.5));
-    }
-}

common/src/writer.rs

@@ -1,114 +0,0 @@
-use std::io::{self, BufWriter, Write};
-
-pub struct CountingWriter<W> {
-    underlying: W,
-    written_bytes: u64,
-}
-
-impl<W: Write> CountingWriter<W> {
-    pub fn wrap(underlying: W) -> CountingWriter<W> {
-        CountingWriter {
-            underlying,
-            written_bytes: 0,
-        }
-    }
-
-    #[inline]
-    pub fn written_bytes(&self) -> u64 {
-        self.written_bytes
-    }
-
-    /// Returns the underlying write object.
-    /// Note that this method does not trigger any flushing.
-    #[inline]
-    pub fn finish(self) -> W {
-        self.underlying
-    }
-}
-
-impl<W: Write> Write for CountingWriter<W> {
-    #[inline]
-    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
-        let written_size = self.underlying.write(buf)?;
-        self.written_bytes += written_size as u64;
-        Ok(written_size)
-    }
-
-    #[inline]
-    fn write_all(&mut self, buf: &[u8]) -> io::Result<()> {
-        self.underlying.write_all(buf)?;
-        self.written_bytes += buf.len() as u64;
-        Ok(())
-    }
-
-    #[inline]
-    fn flush(&mut self) -> io::Result<()> {
-        self.underlying.flush()
-    }
-}
-
-impl<W: TerminatingWrite> TerminatingWrite for CountingWriter<W> {
-    #[inline]
-    fn terminate_ref(&mut self, token: AntiCallToken) -> io::Result<()> {
-        self.underlying.terminate_ref(token)
-    }
-}
-
-/// Struct used to prevent from calling
-/// [`terminate_ref`](TerminatingWrite::terminate_ref) directly
-///
-/// The point is that while the type is public, it cannot be built by anyone
-/// outside of this module.
-pub struct AntiCallToken(());
-
-/// Trait used to indicate when no more write need to be done on a writer
-pub trait TerminatingWrite: Write + Send + Sync {
-    /// Indicate that the writer will no longer be used. Internally call terminate_ref.
-    fn terminate(mut self) -> io::Result<()>
-    where Self: Sized {
-        self.terminate_ref(AntiCallToken(()))
-    }
-
-    /// You should implement this function to define custom behavior.
-    /// This function should flush any buffer it may hold.
-    fn terminate_ref(&mut self, _: AntiCallToken) -> io::Result<()>;
-}
-
-impl<W: TerminatingWrite + ?Sized> TerminatingWrite for Box<W> {
-    fn terminate_ref(&mut self, token: AntiCallToken) -> io::Result<()> {
-        self.as_mut().terminate_ref(token)
-    }
-}
-
-impl<W: TerminatingWrite> TerminatingWrite for BufWriter<W> {
-    fn terminate_ref(&mut self, a: AntiCallToken) -> io::Result<()> {
-        self.flush()?;
-        self.get_mut().terminate_ref(a)
-    }
-}
-
-impl<'a> TerminatingWrite for &'a mut Vec<u8> {
-    fn terminate_ref(&mut self, _a: AntiCallToken) -> io::Result<()> {
-        self.flush()
-    }
-}
-
-#[cfg(test)]
-mod test {
-
-    use std::io::Write;
-
-    use super::CountingWriter;
-
-    #[test]
-    fn test_counting_writer() {
-        let buffer: Vec<u8> = vec![];
-        let mut counting_writer = CountingWriter::wrap(buffer);
-        let bytes = (0u8..10u8).collect::<Vec<u8>>();
-        counting_writer.write_all(&bytes).unwrap();
-        let len = counting_writer.written_bytes();
-        let buffer_restituted: Vec<u8> = counting_writer.finish();
-        assert_eq!(len, 10u64);
-        assert_eq!(buffer_restituted.len(), 10);
-    }
-}

doc/assets/images/element.io.svg

@@ -1,8 +0,0 @@
-<svg width="518" height="112" viewBox="0 0 518 112" fill="none" xmlns="http://www.w3.org/2000/svg">
-<path fill-rule="evenodd" clip-rule="evenodd" d="M56 112C86.9279 112 112 86.9279 112 56C112 25.0721 86.9279 0 56 0C25.0721 0 0 25.0721 0 56C0 86.9279 25.0721 112 56 112Z" fill="#0DBD8B"/>
-<path fill-rule="evenodd" clip-rule="evenodd" d="M45.7615 26.093C45.7615 23.8325 47.5977 22.0001 49.8629 22.0001C65.2154 22.0001 77.6611 34.4199 77.6611 49.7406C77.6611 52.001 75.8248 53.8335 73.5597 53.8335C71.2945 53.8335 69.4583 52.001 69.4583 49.7406C69.4583 38.9408 60.6851 30.1859 49.8629 30.1859C47.5977 30.1859 45.7615 28.3534 45.7615 26.093Z" fill="white"/>
-<path fill-rule="evenodd" clip-rule="evenodd" d="M85.8986 45.6477C88.1637 45.6477 89.9999 47.4801 89.9999 49.7406C89.9999 65.0612 77.5543 77.4811 62.2017 77.4811C59.9366 77.4811 58.1003 75.6486 58.1003 73.3882C58.1003 71.1277 59.9366 69.2953 62.2017 69.2953C73.024 69.2953 81.7972 60.5403 81.7972 49.7406C81.7972 47.4801 83.6334 45.6477 85.8986 45.6477Z" fill="white"/>
-<path fill-rule="evenodd" clip-rule="evenodd" d="M66.3031 85.907C66.3031 88.1675 64.4668 89.9999 62.2017 89.9999C46.8492 89.9999 34.4035 77.58 34.4035 62.2594C34.4035 59.9989 36.2398 58.1665 38.5049 58.1665C40.77 58.1665 42.6063 59.9989 42.6063 62.2594C42.6063 73.0592 51.3795 81.8141 62.2017 81.8141C64.4668 81.8141 66.3031 83.6466 66.3031 85.907Z" fill="white"/>
-<path fill-rule="evenodd" clip-rule="evenodd" d="M26.1014 66.3523C23.8363 66.3523 22.0001 64.5199 22.0001 62.2594C22 46.9388 34.4457 34.5189 49.7983 34.5189C52.0634 34.5189 53.8997 36.3514 53.8997 38.6118C53.8997 40.8723 52.0634 42.7047 49.7983 42.7047C38.976 42.7047 30.2028 51.4597 30.2028 62.2594C30.2028 64.5199 28.3666 66.3523 26.1014 66.3523Z" fill="white"/>
-<path d="M197 63.5H157.5C157.967 67.6333 159.467 70.9333 162 73.4C164.533 75.8 167.867 77 172 77C174.733 77 177.2 76.3333 179.4 75C181.6 73.6667 183.167 71.8667 184.1 69.6H196.1C194.5 74.8667 191.5 79.1333 187.1 82.4C182.767 85.6 177.633 87.2 171.7 87.2C163.967 87.2 157.7 84.6333 152.9 79.5C148.167 74.3667 145.8 67.8667 145.8 60C145.8 52.3333 148.2 45.9 153 40.7C157.8 35.5 164 32.9 171.6 32.9C179.2 32.9 185.333 35.4667 190 40.6C194.733 45.6667 197.1 52.0667 197.1 59.8L197 63.5ZM171.6 42.6C167.867 42.6 164.767 43.7 162.3 45.9C159.833 48.1 158.3 51.0333 157.7 54.7H185.3C184.767 51.0333 183.3 48.1 180.9 45.9C178.5 43.7 175.4 42.6 171.6 42.6ZM205.289 70.5V11H217.189V70.7C217.189 73.3667 218.656 74.7 221.589 74.7L223.689 74.6V85.9C222.556 86.1 221.356 86.2 220.089 86.2C214.956 86.2 211.189 84.9 208.789 82.3C206.456 79.7 205.289 75.7667 205.289 70.5ZM279.109 63.5H239.609C240.076 67.6333 241.576 70.9333 244.109 73.4C246.643 75.8 249.976 77 254.109 77C256.843 77 259.309 76.3333 261.509 75C263.709 73.6667 265.276 71.8667 266.209 69.6H278.209C276.609 74.8667 273.609 79.1333 269.209 82.4C264.876 85.6 259.743 87.2 253.809 87.2C246.076 87.2 239.809 84.6333 235.009 79.5C230.276 74.3667 227.909 67.8667 227.909 60C227.909 52.3333 230.309 45.9 235.109 40.7C239.909 35.5 246.109 32.9 253.709 32.9C261.309 32.9 267.443 35.4667 272.109 40.6C276.843 45.6667 279.209 52.0667 279.209 59.8L279.109 63.5ZM253.709 42.6C249.976 42.6 246.876 43.7 244.409 45.9C241.943 48.1 240.409 51.0333 239.809 54.7H267.409C266.876 51.0333 265.409 48.1 263.009 45.9C260.609 43.7 257.509 42.6 253.709 42.6ZM332.798 56.2V86H320.898V54.9C320.898 47.0333 317.632 43.1 311.098 43.1C307.565 43.1 304.732 44.2333 302.598 46.5C300.532 48.7667 299.498 51.8667 299.498 55.8V86H287.598V34.1H298.598V41C299.865 38.6667 301.798 36.7333 304.398 35.2C306.998 33.6667 310.232 32.9 314.098 32.9C321.298 32.9 326.498 35.6333 329.698 41.1C334.098 35.6333 339.965 32.9 347.298 32.9C353.365 32.9 358.032 34.8 361.298 38.6C364.565 42.3333 366.198 47.2667 366.198 53.4V86H354.298V54.9C354.298 47.0333 351.032 43.1 344.498 43.1C340.898 43.1 338.032 44.2667 335.898 46.6C333.832 48.8667 332.798 52.0667 332.798 56.2ZM425.379 63.5H385.879C386.346 67.6333 387.846 70.9333 390.379 73.4C392.912 75.8 396.246 77 400.379 77C403.112 77 405.579 76.3333 407.779 75C409.979 73.6667 411.546 71.8667 412.479 69.6H424.479C422.879 74.8667 419.879 79.1333 415.479 82.4C411.146 85.6 406.012 87.2 400.079 87.2C392.346 87.2 386.079 84.6333 381.279 79.5C376.546 74.3667 374.179 67.8667 374.179 60C374.179 52.3333 376.579 45.9 381.379 40.7C386.179 35.5 392.379 32.9 399.979 32.9C407.579 32.9 413.712 35.4667 418.379 40.6C423.112 45.6667 425.479 52.0667 425.479 59.8L425.379 63.5ZM399.979 42.6C396.246 42.6 393.146 43.7 390.679 45.9C388.212 48.1 386.679 51.0333 386.079 54.7H413.679C413.146 51.0333 411.679 48.1 409.279 45.9C406.879 43.7 403.779 42.6 399.979 42.6ZM444.868 34.1V41C446.068 38.7333 448.035 36.8333 450.768 35.3C453.568 33.7 456.935 32.9 460.868 32.9C467.001 32.9 471.735 34.7667 475.068 38.5C478.468 42.2333 480.168 47.2 480.168 53.4V86H468.268V54.9C468.268 51.2333 467.401 48.3667 465.668 46.3C464.001 44.1667 461.435 43.1 457.968 43.1C454.168 43.1 451.168 44.2333 448.968 46.5C446.835 48.7667 445.768 51.9 445.768 55.9V86H433.868V34.1H444.868ZM514.922 75.4V85.7C513.455 86.1 511.389 86.3 508.722 86.3C498.589 86.3 493.522 81.2 493.522 71V43.6H485.622V34.1H493.522V20.6H505.422V34.1H515.122V43.6H505.422V69.8C505.422 73.8667 507.355 75.9 511.222 75.9L514.922 75.4Z" fill="black"/>
-</svg>

doc/src/SUMMARY.md

@@ -1,11 +1,12 @@
 # Summary
 
+
+
 [Avant Propos](./avant-propos.md)
 
 - [Segments](./basis.md)
 - [Defining your schema](./schema.md)
 - [Facetting](./facetting.md)
-- [Index Sorting](./index_sorting.md)
 - [Innerworkings](./innerworkings.md)
   - [Inverted index](./inverted_index.md)
 - [Best practise](./inverted_index.md)

doc/src/avant-propos.md

@@ -3,7 +3,7 @@
 > Tantivy is a **search** engine **library** for Rust.
 
 If you are familiar with Lucene, it's an excellent approximation to consider tantivy as Lucene for rust. tantivy is heavily inspired by Lucene's design and
-they both have the same scope and targeted use cases.
+they both have the same scope and targetted use cases.
 
 If you are not familiar with Lucene, let's break down our little tagline.
 
@@ -31,4 +31,4 @@ relevancy, collapsing, highlighting, spatial search.
   index from a different format.
 
   Tantivy exposes a lot of low level API to do all of these things.
-  
+  

doc/src/basis.md

@@ -11,7 +11,7 @@ directory shipped with tantivy is the `MmapDirectory`.
 While this design has some downsides, this greatly simplifies the source code of
 tantivy. Caching is also entirely delegated to the OS.
 
-`tantivy` works entirely (or almost) by directly reading the datastructures as they are laid on disk. As a result, the act of opening an indexing does not involve loading different datastructures from the disk into random access memory : starting a process, opening an index, and performing your first query can typically be done in a matter of milliseconds.
+`tantivy` works entirely (or almost) by directly reading the datastructures as they are layed on disk. As a result, the act of opening an indexing does not involve loading different datastructures from the disk into random access memory : starting a process, opening an index, and performing your first query can typically be done in a matter of milliseconds.
 
 This is an interesting property for a command line search engine, or for some multi-tenant log search engine : spawning a new process for each new query can be a perfectly sensible solution in some use case.
 
@@ -22,6 +22,7 @@ Of course this is crucial to reduce IO, and ensure that as much of our index can
 Also, whenever possible its data is accessed sequentially. Of course, this is an amazing property when tantivy needs to access the data from your spinning hard disk, but this is also
 critical for performance, if your data is read from and an `SSD` or even already in your pagecache.
 
+
 ## Segments, and the log method
 
 That kind of compact layout comes at one cost: it prevents our datastructures from being dynamic.
@@ -50,9 +51,13 @@ to get tantivy to fit your use case:
 
 *Example 1* You could for instance use hadoop to build a very large search index in a timely manner, copy all of the resulting segment files in the same directory and edit the `meta.json` to get a functional index.[^2]
 
-*Example 2* You could also disable your merge policy and enforce daily segments. Removing data after one week can then be done very efficiently by just editing the `meta.json` and deleting the files associated with segment `D-7`.
+*Example 2* You could also disable your merge policy and enforce daily segments. Removing data after one week can then be done very efficiently by just editing the `meta.json` and deleting the files associated to segment `D-7`.
 
-## Merging
+
+
+
+
+# Merging
 
 As you index more and more data, your index will accumulate more and more segments.
 Having a lot of small segments is not really optimal. There is a bit of redundancy in having
@@ -61,7 +66,11 @@ all these term dictionary. Also when searching, we will need to do term lookups
 That's where merging or compacting comes into place. Tantivy will continuously consider merge
 opportunities and start merging segments in the background.
 
-## Indexing throughput, number of indexing threads
+
+# Indexing throughput, number of indexing threads
+
+
+
 
 [^1]: This may eventually change.
 

doc/src/examples.md

@@ -1,3 +1,3 @@
 # Examples
 
-- [Basic search](/examples/basic_search.html)
+- [Basic search](/examples/basic_search.html)

doc/src/index_sorting.md

@@ -1,62 +0,0 @@
-
-- [Index Sorting](#index-sorting)
-  - [Why Sorting](#why-sorting)
-    - [Compression](#compression)
-    - [Top-N Optimization](#top-n-optimization)
-    - [Pruning](#pruning)
-    - [Other](#other)
-  - [Usage](#usage)
-
-# Index Sorting
-
-Tantivy allows you to sort the index according to a property.
-
-## Why Sorting
-
-Presorting an index has several advantages:
-
-### Compression
-
-When data is sorted it is easier to compress the data. E.g. the numbers sequence [5, 2, 3, 1, 4] would be sorted to [1, 2, 3, 4, 5].
-If we apply delta encoding this list would be unsorted [5, -3, 1, -2, 3] vs. [1, 1, 1, 1, 1].
-Compression ratio is mainly affected on the fast field of the sorted property, every thing else is likely unaffected.
-
-### Top-N Optimization
-
-When data is presorted by a field and search queries request sorting by the same field, we can leverage the natural order of the documents.
-E.g. if the data is sorted by timestamp and want the top n newest docs containing a term, we can simply leveraging the order of the docids.
-
-Note: Tantivy 0.16 does not do this optimization yet.
-
-### Pruning
-
-Let's say we want all documents and want to apply the filter `>= 2010-08-11`. When the data is sorted, we could make a lookup in the fast field to find the docid range and use this as the filter.
-
-Note: Tantivy 0.16 does not do this optimization yet.
-
-### Other?
-
-In principle there are many algorithms possible that exploit the monotonically increasing nature. (aggregations maybe?)
-
-## Usage
-
-The index sorting can be configured setting [`sort_by_field`](https://github.com/quickwit-oss/tantivy/blob/000d76b11a139a84b16b9b95060a1c93e8b9851c/src/core/index_meta.rs#L238) on `IndexSettings` and passing it to a `IndexBuilder`. As of Tantivy 0.16 only fast fields are allowed to be used.
-
-```rust
-let settings = IndexSettings {
-    sort_by_field: Some(IndexSortByField {
-        field: "intval".to_string(),
-        order: Order::Desc,
-    }),
-    ..Default::default()
-};
-let mut index_builder = Index::builder().schema(schema);
-index_builder = index_builder.settings(settings);
-let index = index_builder.create_in_ram().unwrap();
-```
-
-## Implementation details
-
-Sorting an index is applied in the serialization step. In general there are two serialization steps: [Finishing a single segment](https://github.com/quickwit-oss/tantivy/blob/000d76b11a139a84b16b9b95060a1c93e8b9851c/src/indexer/segment_writer.rs#L338) and [merging multiple segments](https://github.com/quickwit-oss/tantivy/blob/000d76b11a139a84b16b9b95060a1c93e8b9851c/src/indexer/merger.rs#L1073).
-
-In both cases we generate a docid mapping reflecting the sort. This mapping is used when serializing the different components (doc store, fastfields, posting list, normfield, facets).

doc/src/json.md

@@ -1,130 +0,0 @@
-# Json
-
-As of tantivy 0.17, tantivy supports a json object type.
-This type can be used to allow for a schema-less search index.
-
-When indexing a json object, we "flatten" the JSON. This operation emits terms that represent a triplet `(json_path, value_type, value)`
-
-For instance,  if user is a json field, the following document:
-
-```json
-{
-    "user": {
-        "name": "Paul Masurel",
-        "address": {
-            "city": "Tokyo",
-            "country": "Japan"
-        },
-        "created_at": "2018-11-12T23:20:50.52Z"
-    }
-}
-```
-
-emits the following tokens:
-
-- ("name", Text, "Paul")
-- ("name", Text, "Masurel")
-- ("address.city", Text, "Tokyo")
-- ("address.country", Text, "Japan")
-- ("created_at", Date, 15420648505)
-
-## Bytes-encoding and lexicographical sort
-
-Like any other terms, these triplets are encoded into a binary format as follows.
-
-- `json_path`: the json path is a sequence of "segments". In the example above, `address.city`
-is just a debug representation of the json path `["address", "city"]`.
-Its representation is done by separating segments by a unicode char `\x01`, and ending the path by `\x00`.
-- `value type`: One byte represents the `Value` type.
-- `value`: The value representation is just the regular Value representation.
-
-This representation is designed to align the natural sort of Terms with the lexicographical sort
-of their binary representation (Tantivy's dictionary (whether fst or sstable) is sorted and does prefix encoding).
-
-In the example above, the terms will be sorted as
-
-- ("address.city", Text, "Tokyo")
-- ("address.country", Text, "Japan")
-- ("name", Text, "Masurel")
-- ("name", Text, "Paul")
-- ("created_at", Date, 15420648505)
-
-As seen in "pitfalls", we may end up having to search for a value for a same path in several different fields. Putting the field code after the path makes it maximizes compression opportunities but also increases the chances for the two terms to end up in the actual same term dictionary block.
-
-## Pitfalls, limitation and corner cases
-
-Json gives very little information about the type of the literals it stores.
-All numeric types end up mapped as a "Number" and there are no types for dates.
-
-At indexing, tantivy will try to interpret number and strings as different type with a
-priority order.
-
-Numbers will be interpreted as u64, i64 and f64 in that order.
-Strings will be interpreted as rfc3999 dates or simple strings.
-
-The first working type is picked and is the only term that is emitted for indexing.
-Note this interpretation happens on a per-document basis, and there is no effort to try to sniff
-a consistent field type at the scale of a segment.
-
-On the query parser side on the other hand, we may end up emitting more than one type.
-For instance, we do not even know if the type is a number or string based.
-
-So the query
-
-```rust
-my_path.my_segment:233
-```
-
-Will be interpreted as
-
-```rust
-(my_path.my_segment, String, 233) or (my_path.my_segment, u64, 233)
-```
-
-Likewise, we need to emit two tokens if the query contains an rfc3999 date.
-Indeed the date could have been actually a single token inside the text of a document at ingestion time. Generally speaking, we will always at least emit a string token in query parsing, and sometimes more.
-
-If one more json field is defined, things get even more complicated.
-
-## Default json field
-
-If the schema contains a text field called "text" and a json field that is set as a default field:
-`text:hello` could be reasonably interpreted as targeting the text field or as targeting the json field called `json_dynamic` with the json_path "text".
-
-If there is such an ambiguity, we decide to only search in the "text" field: `text:hello`.
-
-In other words, the parser will not search in default json fields if there is a schema hit.
-This is a product decision.
-
-The user can still target the JSON field by specifying its name explicitly:
-`json_dynamic.text:hello`.
-
-## Range queries are not supported
-
-Json field do not support range queries.
-
-## Arrays do not work like nested object
-
-If json object contains an array, a search query might return more documents
-than what might be expected.
-
-Let's take an example.
-
-```json
-{
-    "cart_id": 3234234 ,
-    "cart": [
-        {"product_type": "sneakers", "attributes": {"color": "white"} },
-        {"product_type": "t-shirt", "attributes": {"color": "red"}},
-    ]
-}
-```
-
-Despite the array structure, a document in tantivy is a bag of terms.
-The query:
-
-```rust
-cart.product_type:sneakers AND cart.attributes.color:red
-```
-
-Actually match the document above.

examples/aggregation.rs

@@ -1,130 +0,0 @@
-// # Aggregation example
-//
-// This example shows how you can use built-in aggregations.
-// We will use range buckets and compute the average in each bucket.
-//
-
-use serde_json::Value;
-use tantivy::aggregation::agg_req::{
-    Aggregation, Aggregations, BucketAggregation, BucketAggregationType, MetricAggregation,
-    RangeAggregation,
-};
-use tantivy::aggregation::agg_result::AggregationResults;
-use tantivy::aggregation::metric::AverageAggregation;
-use tantivy::aggregation::AggregationCollector;
-use tantivy::query::TermQuery;
-use tantivy::schema::{self, Cardinality, IndexRecordOption, Schema, TextFieldIndexing};
-use tantivy::{doc, Index, Term};
-
-fn main() -> tantivy::Result<()> {
-    let mut schema_builder = Schema::builder();
-    let text_fieldtype = schema::TextOptions::default()
-        .set_indexing_options(
-            TextFieldIndexing::default().set_index_option(IndexRecordOption::WithFreqs),
-        )
-        .set_stored();
-    let text_field = schema_builder.add_text_field("text", text_fieldtype);
-    let score_fieldtype =
-        crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue);
-    let highscore_field = schema_builder.add_f64_field("highscore", score_fieldtype.clone());
-    let price_field = schema_builder.add_f64_field("price", score_fieldtype.clone());
-
-    let schema = schema_builder.build();
-
-    // # Indexing documents
-    //
-    // Lets index a bunch of documents for this example.
-    let index = Index::create_in_ram(schema);
-
-    let mut index_writer = index.writer(50_000_000)?;
-    // writing the segment
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 1f64,
-        price_field => 0f64,
-    ))?;
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 3f64,
-        price_field => 1f64,
-    ))?;
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 5f64,
-        price_field => 1f64,
-    ))?;
-    index_writer.add_document(doc!(
-        text_field => "nohit",
-        highscore_field => 6f64,
-        price_field => 2f64,
-    ))?;
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 7f64,
-        price_field => 2f64,
-    ))?;
-    index_writer.commit()?;
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 11f64,
-        price_field => 10f64,
-    ))?;
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 14f64,
-        price_field => 15f64,
-    ))?;
-
-    index_writer.add_document(doc!(
-        text_field => "cool",
-        highscore_field => 15f64,
-        price_field => 20f64,
-    ))?;
-
-    index_writer.commit()?;
-
-    let reader = index.reader()?;
-    let text_field = reader.searcher().schema().get_field("text").unwrap();
-
-    let term_query = TermQuery::new(
-        Term::from_field_text(text_field, "cool"),
-        IndexRecordOption::Basic,
-    );
-
-    let sub_agg_req_1: Aggregations = vec![(
-        "average_price".to_string(),
-        Aggregation::Metric(MetricAggregation::Average(
-            AverageAggregation::from_field_name("price".to_string()),
-        )),
-    )]
-    .into_iter()
-    .collect();
-
-    let agg_req_1: Aggregations = vec![(
-        "score_ranges".to_string(),
-        Aggregation::Bucket(BucketAggregation {
-            bucket_agg: BucketAggregationType::Range(RangeAggregation {
-                field: "highscore".to_string(),
-                ranges: vec![
-                    (-1f64..9f64).into(),
-                    (9f64..14f64).into(),
-                    (14f64..20f64).into(),
-                ],
-                ..Default::default()
-            }),
-            sub_aggregation: sub_agg_req_1.clone(),
-        }),
-    )]
-    .into_iter()
-    .collect();
-
-    let collector = AggregationCollector::from_aggs(agg_req_1, None, index.schema());
-
-    let searcher = reader.searcher();
-    let agg_res: AggregationResults = searcher.search(&term_query, &collector).unwrap();
-
-    let res: Value = serde_json::to_value(&agg_res)?;
-    println!("{}", serde_json::to_string_pretty(&res)?);
-
-    Ok(())
-}

examples/basic_search.rs

@@ -5,23 +5,28 @@
 //
 // We will :
 // - define our schema
-// - create an index in a directory
-// - index a few documents into our index
-// - search for the best document matching a basic query
-// - retrieve the best document's original content.
+// = create an index in a directory
+// - index few documents in our index
+// - search for the best document matchings "sea whale"
+// - retrieve the best document original content.
+
+extern crate tempdir;
 
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
-use tantivy::{doc, Index, ReloadPolicy};
-use tempfile::TempDir;
+use tantivy::Index;
+use tantivy::ReloadPolicy;
+use tempdir::TempDir;
 
 fn main() -> tantivy::Result<()> {
     // Let's create a temporary directory for the
     // sake of this example
-    let index_path = TempDir::new()?;
+    let index_path = TempDir::new("tantivy_example_dir")?;
 
     // # Defining the schema
     //
@@ -30,7 +35,7 @@ fn main() -> tantivy::Result<()> {
     // and for each field, its type and "the way it should
     // be indexed".
 
-    // First we need to define a schema ...
+    // first we need to define a schema ...
     let mut schema_builder = Schema::builder();
 
     // Our first field is title.
@@ -45,7 +50,7 @@ fn main() -> tantivy::Result<()> {
     //
     // `STORED` means that the field will also be saved
     // in a compressed, row-oriented key-value store.
-    // This store is useful for reconstructing the
+    // This store is useful to reconstruct the
     // documents that were selected during the search phase.
     schema_builder.add_text_field("title", TEXT | STORED);
 
@@ -54,7 +59,8 @@ fn main() -> tantivy::Result<()> {
     // need to be able to be able to retrieve it
     // for our application.
     //
-    // We can make our index lighter by omitting the `STORED` flag.
+    // We can make our index lighter and
+    // by omitting `STORED` flag.
     schema_builder.add_text_field("body", TEXT);
 
     let schema = schema_builder.build();
@@ -67,13 +73,13 @@ fn main() -> tantivy::Result<()> {
     // with our schema in the directory.
     let index = Index::create_in_dir(&index_path, schema.clone())?;
 
-    // To insert a document we will need an index writer.
+    // To insert document we need an index writer.
     // There must be only one writer at a time.
     // This single `IndexWriter` is already
     // multithreaded.
     //
     // Here we give tantivy a budget of `50MB`.
-    // Using a bigger memory_arena for the indexer may increase
+    // Using a bigger heap for the indexer may increase
     // throughput, but 50 MB is already plenty.
     let mut index_writer = index.writer(50_000_000)?;
 
@@ -91,12 +97,12 @@ fn main() -> tantivy::Result<()> {
     old_man_doc.add_text(title, "The Old Man and the Sea");
     old_man_doc.add_text(
         body,
-        "He was an old man who fished alone in a skiff in the Gulf Stream and he had gone \
-         eighty-four days now without taking a fish.",
+        "He was an old man who fished alone in a skiff in the Gulf Stream and \
+         he had gone eighty-four days now without taking a fish.",
     );
 
     // ... and add it to the `IndexWriter`.
-    index_writer.add_document(old_man_doc)?;
+    index_writer.add_document(old_man_doc);
 
     // For convenience, tantivy also comes with a macro to
     // reduce the boilerplate above.
@@ -110,7 +116,19 @@ fn main() -> tantivy::Result<()> {
             fresh and green with every spring, carrying in their lower leaf junctures the \
             debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
             limbs and branches that arch over the pool"
-    ))?;
+    ));
+
+    index_writer.add_document(doc!(
+    title => "Of Mice and Men",
+    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+            over the yellow sands in the sunlight before reaching the narrow pool. On one \
+            side of the river the golden foothill slopes curve up to the strong and rocky \
+            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+            fresh and green with every spring, carrying in their lower leaf junctures the \
+            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+            limbs and branches that arch over the pool"
+    ));
 
     // Multivalued field just need to be repeated.
     index_writer.add_document(doc!(
@@ -120,7 +138,7 @@ fn main() -> tantivy::Result<()> {
              enterprise which you have regarded with such evil forebodings.  I arrived here \
              yesterday, and my first task is to assure my dear sister of my welfare and \
              increasing confidence in the success of my undertaking."
-    ))?;
+    ));
 
     // This is an example, so we will only index 3 documents
     // here. You can check out tantivy's tutorial to index
@@ -133,8 +151,8 @@ fn main() -> tantivy::Result<()> {
     // At this point our documents are not searchable.
     //
     //
-    // We need to call `.commit()` explicitly to force the
-    // `index_writer` to finish processing the documents in the queue,
+    // We need to call .commit() explicitly to force the
+    // index_writer to finish processing the documents in the queue,
     // flush the current index to the disk, and advertise
     // the existence of new documents.
     //
@@ -146,14 +164,14 @@ fn main() -> tantivy::Result<()> {
     // persistently indexed.
     //
     // In the scenario of a crash or a power failure,
-    // tantivy behaves as if it has rolled back to its last
+    // tantivy behaves as if has rolled back to its last
     // commit.
 
     // # Searching
     //
     // ### Searcher
     //
-    // A reader is required first in order to search an index.
+    // A reader is required to get search the index.
     // It acts as a `Searcher` pool that reloads itself,
     // depending on a `ReloadPolicy`.
     //
@@ -169,7 +187,7 @@ fn main() -> tantivy::Result<()> {
 
     // We now need to acquire a searcher.
     //
-    // A searcher points to a snapshotted, immutable version of the index.
+    // A searcher points to snapshotted, immutable version of the index.
     //
     // Some search experience might require more than
     // one query. Using the same searcher ensures that all of these queries will run on the
@@ -189,7 +207,7 @@ fn main() -> tantivy::Result<()> {
     // in both title and body.
     let query_parser = QueryParser::for_index(&index, vec![title, body]);
 
-    // `QueryParser` may fail if the query is not in the right
+    // QueryParser may fail if the query is not in the right
     // format. For user facing applications, this can be a problem.
     // A ticket has been opened regarding this problem.
     let query = query_parser.parse_query("sea whale")?;
@@ -205,7 +223,7 @@ fn main() -> tantivy::Result<()> {
     //
     // We are not interested in all of the documents but
     // only in the top 10. Keeping track of our top 10 best documents
-    // is the role of the `TopDocs` collector.
+    // is the role of the TopDocs.
 
     // We can now perform our query.
     let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;

examples/custom_collector.rs

@@ -7,15 +7,19 @@
 // Of course, you can have a look at the tantivy's built-in collectors
 // such as the `CountCollector` for more examples.
 
-use std::sync::Arc;
+extern crate tempdir;
 
-use fastfield_codecs::Column;
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::{Collector, SegmentCollector};
+use tantivy::fastfield::FastFieldReader;
 use tantivy::query::QueryParser;
-use tantivy::schema::{Field, Schema, FAST, INDEXED, TEXT};
-use tantivy::{doc, Index, Score, SegmentReader};
+use tantivy::schema::Field;
+use tantivy::schema::{Schema, FAST, INDEXED, TEXT};
+use tantivy::SegmentReader;
+use tantivy::{Index, TantivyError};
 
 #[derive(Default)]
 struct Stats {
@@ -73,7 +77,16 @@ impl Collector for StatsCollector {
         _segment_local_id: u32,
         segment_reader: &SegmentReader,
     ) -> tantivy::Result<StatsSegmentCollector> {
-        let fast_field_reader = segment_reader.fast_fields().u64(self.field)?;
+        let fast_field_reader = segment_reader
+            .fast_fields()
+            .u64(self.field)
+            .ok_or_else(|| {
+                let field_name = segment_reader.schema().get_field_name(self.field);
+                TantivyError::SchemaError(format!(
+                    "Field {:?} is not a u64 fast field.",
+                    field_name
+                ))
+            })?;
         Ok(StatsSegmentCollector {
             fast_field_reader,
             stats: Stats::default(),
@@ -87,25 +100,27 @@ impl Collector for StatsCollector {
 
     fn merge_fruits(&self, segment_stats: Vec<Option<Stats>>) -> tantivy::Result<Option<Stats>> {
         let mut stats = Stats::default();
-        for segment_stats in segment_stats.into_iter().flatten() {
-            stats.count += segment_stats.count;
-            stats.sum += segment_stats.sum;
-            stats.squared_sum += segment_stats.squared_sum;
+        for segment_stats_opt in segment_stats {
+            if let Some(segment_stats) = segment_stats_opt {
+                stats.count += segment_stats.count;
+                stats.sum += segment_stats.sum;
+                stats.squared_sum += segment_stats.squared_sum;
+            }
         }
         Ok(stats.non_zero_count())
     }
 }
 
 struct StatsSegmentCollector {
-    fast_field_reader: Arc<dyn Column<u64>>,
+    fast_field_reader: FastFieldReader<u64>,
     stats: Stats,
 }
 
 impl SegmentCollector for StatsSegmentCollector {
     type Fruit = Option<Stats>;
 
-    fn collect(&mut self, doc: u32, _score: Score) {
-        let value = self.fast_field_reader.get_val(doc) as f64;
+    fn collect(&mut self, doc: u32, _score: f32) {
+        let value = self.fast_field_reader.get(doc) as f64;
         self.stats.count += 1;
         self.stats.sum += value;
         self.stats.squared_sum += value * value;
@@ -138,7 +153,7 @@ fn main() -> tantivy::Result<()> {
     //
     // Lets index a bunch of fake documents for the sake of
     // this example.
-    let index = Index::create_in_ram(schema);
+    let index = Index::create_in_ram(schema.clone());
 
     let mut index_writer = index.writer(50_000_000)?;
     index_writer.add_document(doc!(
@@ -146,23 +161,23 @@ fn main() -> tantivy::Result<()> {
         product_description => "While it is ok for short distance travel, this broom \
         was designed quiditch. It will up your game.",
         price => 30_200u64
-    ))?;
+    ));
     index_writer.add_document(doc!(
         product_name => "Turbulobroom",
         product_description => "You might have heard of this broom before : it is the sponsor of the Wales team.\
             You'll enjoy its sharp turns, and rapid acceleration",
         price => 29_240u64
-    ))?;
+    ));
     index_writer.add_document(doc!(
         product_name => "Broomio",
         product_description => "Great value for the price. This broom is a market favorite",
         price => 21_240u64
-    ))?;
+    ));
     index_writer.add_document(doc!(
         product_name => "Whack a Mole",
         product_description => "Prime quality bat.",
         price => 5_200u64
-    ))?;
+    ));
     index_writer.commit()?;
 
     let reader = index.reader()?;

examples/custom_tokenizer.rs

@@ -2,11 +2,14 @@
 //
 // In this example, we'll see how to define a tokenizer pipeline
 // by aligning a bunch of `TokenFilter`.
+
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::NgramTokenizer;
-use tantivy::{doc, Index};
+use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
     // # Defining the schema
@@ -36,7 +39,8 @@ fn main() -> tantivy::Result<()> {
     // need to be able to be able to retrieve it
     // for our application.
     //
-    // We can make our index lighter by omitting the `STORED` flag.
+    // We can make our index lighter and
+    // by omitting `STORED` flag.
     let body = schema_builder.add_text_field("body", TEXT);
 
     let schema = schema_builder.build();
@@ -49,7 +53,7 @@ fn main() -> tantivy::Result<()> {
     // for your unit tests... Or this example.
     let index = Index::create_in_ram(schema.clone());
 
-    // here we are registering our custom tokenizer
+    // here we are registering our custome tokenizer
     // this will store tokens of 3 characters each
     index
         .tokenizers()
@@ -61,13 +65,13 @@ fn main() -> tantivy::Result<()> {
     // multithreaded.
     //
     // Here we use a buffer of 50MB per thread. Using a bigger
-    // memory arena for the indexer can increase its throughput.
+    // heap for the indexer can increase its throughput.
     let mut index_writer = index.writer(50_000_000)?;
     index_writer.add_document(doc!(
     title => "The Old Man and the Sea",
     body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
      he had gone eighty-four days now without taking a fish."
-    ))?;
+    ));
     index_writer.add_document(doc!(
     title => "Of Mice and Men",
        body => r#"A few miles south of Soledad, the Salinas River drops in close to the hillside
@@ -78,14 +82,14 @@ fn main() -> tantivy::Result<()> {
                 fresh and green with every spring, carrying in their lower leaf junctures the
                 debris of the winter’s flooding; and sycamores with mottled, white, recumbent
                 limbs and branches that arch over the pool"#
-    ))?;
+    ));
     index_writer.add_document(doc!(
     title => "Frankenstein",
         body => r#"You will rejoice to hear that no disaster has accompanied the commencement of an
                 enterprise which you have regarded with such evil forebodings.  I arrived here
                 yesterday, and my first task is to assure my dear sister of my welfare and
                 increasing confidence in the success of my undertaking."#
-    ))?;
+    ));
     index_writer.commit()?;
 
     let reader = index.reader()?;

examples/date_time_field.rs

@@ -1,69 +0,0 @@
-// # DateTime field example
-//
-// This example shows how the DateTime field can be used
-
-use tantivy::collector::TopDocs;
-use tantivy::query::QueryParser;
-use tantivy::schema::{Cardinality, DateOptions, Schema, Value, INDEXED, STORED, STRING};
-use tantivy::Index;
-
-fn main() -> tantivy::Result<()> {
-    // # Defining the schema
-    let mut schema_builder = Schema::builder();
-    let opts = DateOptions::from(INDEXED)
-        .set_stored()
-        .set_fast(Cardinality::SingleValue)
-        .set_precision(tantivy::DatePrecision::Seconds);
-    let occurred_at = schema_builder.add_date_field("occurred_at", opts);
-    let event_type = schema_builder.add_text_field("event", STRING | STORED);
-    let schema = schema_builder.build();
-
-    // # Indexing documents
-    let index = Index::create_in_ram(schema.clone());
-
-    let mut index_writer = index.writer(50_000_000)?;
-    let doc = schema.parse_document(
-        r#"{
-        "occurred_at": "2022-06-22T12:53:50.53Z",
-        "event": "pull-request"
-    }"#,
-    )?;
-    index_writer.add_document(doc)?;
-    let doc = schema.parse_document(
-        r#"{
-        "occurred_at": "2022-06-22T13:00:00.22Z",
-        "event": "comment"
-    }"#,
-    )?;
-    index_writer.add_document(doc)?;
-    index_writer.commit()?;
-
-    let reader = index.reader()?;
-    let searcher = reader.searcher();
-
-    // # Default fields: event_type
-    let query_parser = QueryParser::for_index(&index, vec![event_type]);
-    {
-        let query = query_parser.parse_query("event:comment")?;
-        let count_docs = searcher.search(&*query, &TopDocs::with_limit(5))?;
-        assert_eq!(count_docs.len(), 1);
-    }
-    {
-        let query = query_parser
-            .parse_query(r#"occurred_at:[2022-06-22T12:58:00Z TO 2022-06-23T00:00:00Z}"#)?;
-        let count_docs = searcher.search(&*query, &TopDocs::with_limit(4))?;
-        assert_eq!(count_docs.len(), 1);
-        for (_score, doc_address) in count_docs {
-            let retrieved_doc = searcher.doc(doc_address)?;
-            assert!(matches!(
-                retrieved_doc.get_first(occurred_at),
-                Some(Value::Date(_))
-            ));
-            assert_eq!(
-                schema.to_json(&retrieved_doc),
-                r#"{"event":["comment"],"occurred_at":["2022-06-22T13:00:00.22Z"]}"#
-            );
-        }
-    }
-    Ok(())
-}

examples/deleting_updating_documents.rs

@@ -8,10 +8,13 @@
 //
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::TopDocs;
 use tantivy::query::TermQuery;
 use tantivy::schema::*;
-use tantivy::{doc, Index, IndexReader};
+use tantivy::Index;
+use tantivy::IndexReader;
 
 // A simple helper function to fetch a single document
 // given its id from our index.
@@ -56,9 +59,8 @@ fn main() -> tantivy::Result<()> {
     // If it is `text`, let's make sure to keep it `raw` and let's avoid
     // running any text processing on it.
     // This is done by associating this field to the tokenizer named `raw`.
-    // Rather than building our
-    // [`TextOptions`](//docs.rs/tantivy/~0/tantivy/schema/struct.TextOptions.html) manually, We
-    // use the `STRING` shortcut. `STRING` stands for indexed (without term frequency or positions)
+    // Rather than building our [`TextOptions`](//docs.rs/tantivy/~0/tantivy/schema/struct.TextOptions.html) manually,
+    // We use the `STRING` shortcut. `STRING` stands for indexed (without term frequency or positions)
     // and untokenized.
     //
     // Because we also want to be able to see this `id` in our returned documents,
@@ -77,21 +79,21 @@ fn main() -> tantivy::Result<()> {
     index_writer.add_document(doc!(
         isbn => "978-0099908401",
         title => "The old Man and the see"
-    ))?;
+    ));
     index_writer.add_document(doc!(
         isbn => "978-0140177398",
         title => "Of Mice and Men",
-    ))?;
+    ));
     index_writer.add_document(doc!(
        title => "Frankentein", //< Oops there is a typo here.
        isbn => "978-9176370711",
-    ))?;
+    ));
     index_writer.commit()?;
     let reader = index.reader()?;
 
     let frankenstein_isbn = Term::from_field_text(isbn, "978-9176370711");
 
-    // Oops our frankenstein doc seems misspelled
+    // Oops our frankenstein doc seems mispelled
     let frankenstein_doc_misspelled = extract_doc_given_isbn(&reader, &frankenstein_isbn)?.unwrap();
     assert_eq!(
         schema.to_json(&frankenstein_doc_misspelled),
@@ -113,7 +115,7 @@ fn main() -> tantivy::Result<()> {
     // on its id.
     //
     // Note that `tantivy` does nothing to enforce the idea that
-    // there is only one document associated with this id.
+    // there is only one document associated to this id.
     //
     // Also you might have noticed that we apply the delete before
     // having committed. This does not matter really...
@@ -123,7 +125,7 @@ fn main() -> tantivy::Result<()> {
     index_writer.add_document(doc!(
        title => "Frankenstein",
        isbn => "978-9176370711",
-    ))?;
+    ));
 
     // You are guaranteed that your clients will only observe your index in
     // the state it was in after a commit.

examples/faceted_search.rs

@@ -10,103 +10,71 @@
 // - search for the best document matchings "sea whale"
 // - retrieve the best document original content.
 
+extern crate tempdir;
+
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::FacetCollector;
-use tantivy::query::{AllQuery, TermQuery};
+use tantivy::query::AllQuery;
 use tantivy::schema::*;
-use tantivy::{doc, Index};
+use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
-    // Let's create a temporary directory for the sake of this example
+    // Let's create a temporary directory for the
+    // sake of this example
+    let index_path = TempDir::new("tantivy_facet_example_dir")?;
     let mut schema_builder = Schema::builder();
 
-    let name = schema_builder.add_text_field("felin_name", TEXT | STORED);
-    // this is our faceted field: its scientific classification
-    let classification = schema_builder.add_facet_field("classification", FacetOptions::default());
+    schema_builder.add_text_field("name", TEXT | STORED);
+
+    // this is our faceted field
+    schema_builder.add_facet_field("tags");
 
     let schema = schema_builder.build();
-    let index = Index::create_in_ram(schema);
 
-    let mut index_writer = index.writer(30_000_000)?;
+    let index = Index::create_in_dir(&index_path, schema.clone())?;
+
+    let mut index_writer = index.writer(50_000_000)?;
+
+    let name = schema.get_field("name").unwrap();
+    let tags = schema.get_field("tags").unwrap();
 
     // For convenience, tantivy also comes with a macro to
     // reduce the boilerplate above.
     index_writer.add_document(doc!(
-        name => "Cat",
-        classification => Facet::from("/Felidae/Felinae/Felis")
-    ))?;
+        name => "the ditch",
+        tags => Facet::from("/pools/north")
+    ));
+
     index_writer.add_document(doc!(
-        name => "Canada lynx",
-        classification => Facet::from("/Felidae/Felinae/Lynx")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Cheetah",
-        classification => Facet::from("/Felidae/Felinae/Acinonyx")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Tiger",
-        classification => Facet::from("/Felidae/Pantherinae/Panthera")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Lion",
-        classification => Facet::from("/Felidae/Pantherinae/Panthera")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Jaguar",
-        classification => Facet::from("/Felidae/Pantherinae/Panthera")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Sunda clouded leopard",
-        classification => Facet::from("/Felidae/Pantherinae/Neofelis")
-    ))?;
-    index_writer.add_document(doc!(
-        name => "Fossa",
-        classification => Facet::from("/Eupleridae/Cryptoprocta")
-    ))?;
+        name => "little stacey",
+        tags => Facet::from("/pools/south")
+    ));
+
     index_writer.commit()?;
 
     let reader = index.reader()?;
+
     let searcher = reader.searcher();
-    {
-        let mut facet_collector = FacetCollector::for_field(classification);
-        facet_collector.add_facet("/Felidae");
-        let facet_counts = searcher.search(&AllQuery, &facet_collector)?;
-        // This lists all of the facet counts, right below "/Felidae".
-        let facets: Vec<(&Facet, u64)> = facet_counts.get("/Felidae").collect();
-        assert_eq!(
-            facets,
-            vec![
-                (&Facet::from("/Felidae/Felinae"), 3),
-                (&Facet::from("/Felidae/Pantherinae"), 4),
-            ]
-        );
-    }
 
-    // Facets are also searchable.
-    //
-    // For instance a common UI pattern is to allow the user someone to click on a facet link
-    // (e.g: `Pantherinae`) to drill down and filter the current result set with this subfacet.
-    //
-    // The search would then look as follows.
+    let mut facet_collector = FacetCollector::for_field(tags);
+    facet_collector.add_facet("/pools");
 
-    // Check the reference doc for different ways to create a `Facet` object.
-    {
-        let facet = Facet::from("/Felidae/Pantherinae");
-        let facet_term = Term::from_facet(classification, &facet);
-        let facet_term_query = TermQuery::new(facet_term, IndexRecordOption::Basic);
-        let mut facet_collector = FacetCollector::for_field(classification);
-        facet_collector.add_facet("/Felidae/Pantherinae");
-        let facet_counts = searcher.search(&facet_term_query, &facet_collector)?;
-        let facets: Vec<(&Facet, u64)> = facet_counts.get("/Felidae/Pantherinae").collect();
-        assert_eq!(
-            facets,
-            vec![
-                (&Facet::from("/Felidae/Pantherinae/Neofelis"), 1),
-                (&Facet::from("/Felidae/Pantherinae/Panthera"), 3),
-            ]
-        );
-    }
+    let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+
+    // This lists all of the facet counts
+    let facets: Vec<(&Facet, u64)> = facet_counts.get("/pools").collect();
+    assert_eq!(
+        facets,
+        vec![
+            (&Facet::from("/pools/north"), 1),
+            (&Facet::from("/pools/south"), 1),
+        ]
+    );
 
     Ok(())
 }
+
+use tempdir::TempDir;

examples/faceted_search_with_tweaked_score.rs

@@ -1,98 +0,0 @@
-use std::collections::HashSet;
-
-use tantivy::collector::TopDocs;
-use tantivy::query::BooleanQuery;
-use tantivy::schema::*;
-use tantivy::{doc, DocId, Index, Score, SegmentReader};
-
-fn main() -> tantivy::Result<()> {
-    let mut schema_builder = Schema::builder();
-
-    let title = schema_builder.add_text_field("title", STORED);
-    let ingredient = schema_builder.add_facet_field("ingredient", FacetOptions::default());
-
-    let schema = schema_builder.build();
-    let index = Index::create_in_ram(schema);
-
-    let mut index_writer = index.writer(30_000_000)?;
-
-    index_writer.add_document(doc!(
-        title => "Fried egg",
-        ingredient => Facet::from("/ingredient/egg"),
-        ingredient => Facet::from("/ingredient/oil"),
-    ))?;
-    index_writer.add_document(doc!(
-        title => "Scrambled egg",
-        ingredient => Facet::from("/ingredient/egg"),
-        ingredient => Facet::from("/ingredient/butter"),
-        ingredient => Facet::from("/ingredient/milk"),
-        ingredient => Facet::from("/ingredient/salt"),
-    ))?;
-    index_writer.add_document(doc!(
-        title => "Egg rolls",
-        ingredient => Facet::from("/ingredient/egg"),
-        ingredient => Facet::from("/ingredient/garlic"),
-        ingredient => Facet::from("/ingredient/salt"),
-        ingredient => Facet::from("/ingredient/oil"),
-        ingredient => Facet::from("/ingredient/tortilla-wrap"),
-        ingredient => Facet::from("/ingredient/mushroom"),
-    ))?;
-    index_writer.commit()?;
-
-    let reader = index.reader()?;
-    let searcher = reader.searcher();
-    {
-        let facets = vec![
-            Facet::from("/ingredient/egg"),
-            Facet::from("/ingredient/oil"),
-            Facet::from("/ingredient/garlic"),
-            Facet::from("/ingredient/mushroom"),
-        ];
-        let query = BooleanQuery::new_multiterms_query(
-            facets
-                .iter()
-                .map(|key| Term::from_facet(ingredient, key))
-                .collect(),
-        );
-        let top_docs_by_custom_score =
-            TopDocs::with_limit(2).tweak_score(move |segment_reader: &SegmentReader| {
-                let ingredient_reader = segment_reader.facet_reader(ingredient).unwrap();
-                let facet_dict = ingredient_reader.facet_dict();
-
-                let query_ords: HashSet<u64> = facets
-                    .iter()
-                    .filter_map(|key| facet_dict.term_ord(key.encoded_str()).unwrap())
-                    .collect();
-
-                let mut facet_ords_buffer: Vec<u64> = Vec::with_capacity(20);
-
-                move |doc: DocId, original_score: Score| {
-                    ingredient_reader.facet_ords(doc, &mut facet_ords_buffer);
-                    let missing_ingredients = facet_ords_buffer
-                        .iter()
-                        .filter(|ord| !query_ords.contains(ord))
-                        .count();
-                    let tweak = 1.0 / 4_f32.powi(missing_ingredients as i32);
-
-                    original_score * tweak
-                }
-            });
-        let top_docs = searcher.search(&query, &top_docs_by_custom_score)?;
-
-        let titles: Vec<String> = top_docs
-            .iter()
-            .map(|(_, doc_id)| {
-                searcher
-                    .doc(*doc_id)
-                    .unwrap()
-                    .get_first(title)
-                    .unwrap()
-                    .as_text()
-                    .unwrap()
-                    .to_owned()
-            })
-            .collect();
-        assert_eq!(titles, vec!["Fried egg", "Egg rolls"]);
-    }
-    Ok(())
-}

examples/integer_range_search.rs

@@ -2,12 +2,16 @@
 //
 // Below is an example of creating an indexed integer field in your schema
 // You can use RangeQuery to get a Count of all occurrences in a given range.
+
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::Count;
 use tantivy::query::RangeQuery;
 use tantivy::schema::{Schema, INDEXED};
-use tantivy::{doc, Index, Result};
+use tantivy::Index;
+use tantivy::Result;
 
-fn main() -> Result<()> {
+fn run() -> Result<()> {
     // For the sake of simplicity, this schema will only have 1 field
     let mut schema_builder = Schema::builder();
 
@@ -19,7 +23,7 @@ fn main() -> Result<()> {
     {
         let mut index_writer = index.writer_with_num_threads(1, 6_000_000)?;
         for year in 1950u64..2019u64 {
-            index_writer.add_document(doc!(year_field => year))?;
+            index_writer.add_document(doc!(year_field => year));
         }
         index_writer.commit()?;
         // The index will be a range of years
@@ -33,3 +37,7 @@ fn main() -> Result<()> {
     assert_eq!(num_60s_books, 10);
     Ok(())
 }
+
+fn main() {
+    run().unwrap()
+}

examples/iterating_docs_and_positions.rs

@@ -1,4 +1,4 @@
-// # Iterating docs and positions.
+// # Iterating docs and positioms.
 //
 // At its core of tantivy, relies on a data structure
 // called an inverted index.
@@ -9,8 +9,11 @@
 
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::schema::*;
-use tantivy::{doc, DocSet, Index, Postings, TERMINATED};
+use tantivy::Index;
+use tantivy::{DocId, DocSet, Postings};
 
 fn main() -> tantivy::Result<()> {
     // We first create a schema for the sake of the
@@ -22,12 +25,12 @@ fn main() -> tantivy::Result<()> {
     let title = schema_builder.add_text_field("title", TEXT | STORED);
     let schema = schema_builder.build();
 
-    let index = Index::create_in_ram(schema);
+    let index = Index::create_in_ram(schema.clone());
 
     let mut index_writer = index.writer_with_num_threads(1, 50_000_000)?;
-    index_writer.add_document(doc!(title => "The Old Man and the Sea"))?;
-    index_writer.add_document(doc!(title => "Of Mice and Men"))?;
-    index_writer.add_document(doc!(title => "The modern Promotheus"))?;
+    index_writer.add_document(doc!(title => "The Old Man and the Sea"));
+    index_writer.add_document(doc!(title => "Of Mice and Men"));
+    index_writer.add_document(doc!(title => "The modern Promotheus"));
     index_writer.commit()?;
 
     let reader = index.reader()?;
@@ -44,29 +47,30 @@ fn main() -> tantivy::Result<()> {
         // A segment contains different data structure.
         // Inverted index stands for the combination of
         // - the term dictionary
-        // - the inverted lists associated with each terms and their positions
-        let inverted_index = segment_reader.inverted_index(title)?;
+        // - the inverted lists associated to each terms and their positions
+        let inverted_index = segment_reader.inverted_index(title);
 
         // A `Term` is a text token associated with a field.
         // Let's go through all docs containing the term `title:the` and access their position
         let term_the = Term::from_field_text(title, "the");
 
         // This segment posting object is like a cursor over the documents matching the term.
-        // The `IndexRecordOption` arguments tells tantivy we will be interested in both term
-        // frequencies and positions.
+        // The `IndexRecordOption` arguments tells tantivy we will be interested in both term frequencies
+        // and positions.
         //
-        // If you don't need all this information, you may get better performance by decompressing
-        // less information.
+        // If you don't need all this information, you may get better performance by decompressing less
+        // information.
         if let Some(mut segment_postings) =
-            inverted_index.read_postings(&term_the, IndexRecordOption::WithFreqsAndPositions)?
+            inverted_index.read_postings(&term_the, IndexRecordOption::WithFreqsAndPositions)
         {
             // this buffer will be used to request for positions
             let mut positions: Vec<u32> = Vec::with_capacity(100);
-            let mut doc_id = segment_postings.doc();
-            while doc_id != TERMINATED {
+            while segment_postings.advance() {
+                // the number of time the term appears in the document.
+                let doc_id: DocId = segment_postings.doc(); //< do not try to access this before calling advance once.
+
                 // This MAY contains deleted documents as well.
                 if segment_reader.is_deleted(doc_id) {
-                    doc_id = segment_postings.advance();
                     continue;
                 }
 
@@ -85,7 +89,6 @@ fn main() -> tantivy::Result<()> {
                 // Doc 2: TermFreq 1: [0]
                 // ```
                 println!("Doc {}: TermFreq {}: {:?}", doc_id, term_freq, positions);
-                doc_id = segment_postings.advance();
             }
         }
     }
@@ -105,28 +108,23 @@ fn main() -> tantivy::Result<()> {
         // A segment contains different data structure.
         // Inverted index stands for the combination of
         // - the term dictionary
-        // - the inverted lists associated with each terms and their positions
-        let inverted_index = segment_reader.inverted_index(title)?;
+        // - the inverted lists associated to each terms and their positions
+        let inverted_index = segment_reader.inverted_index(title);
 
         // This segment posting object is like a cursor over the documents matching the term.
-        // The `IndexRecordOption` arguments tells tantivy we will be interested in both term
-        // frequencies and positions.
+        // The `IndexRecordOption` arguments tells tantivy we will be interested in both term frequencies
+        // and positions.
         //
-        // If you don't need all this information, you may get better performance by decompressing
-        // less information.
+        // If you don't need all this information, you may get better performance by decompressing less
+        // information.
         if let Some(mut block_segment_postings) =
-            inverted_index.read_block_postings(&term_the, IndexRecordOption::Basic)?
+            inverted_index.read_block_postings(&term_the, IndexRecordOption::Basic)
         {
-            loop {
-                let docs = block_segment_postings.docs();
-                if docs.is_empty() {
-                    break;
-                }
+            while block_segment_postings.advance() {
                 // Once again these docs MAY contains deleted documents as well.
                 let docs = block_segment_postings.docs();
                 // Prints `Docs [0, 2].`
                 println!("Docs {:?}", docs);
-                block_segment_postings.advance();
             }
         }
     }

examples/json_field.rs

@@ -1,105 +0,0 @@
-// # Json field example
-//
-// This example shows how the json field can be used
-// to make tantivy partially schemaless by setting it as
-// default query parser field.
-
-use tantivy::collector::{Count, TopDocs};
-use tantivy::query::QueryParser;
-use tantivy::schema::{Schema, FAST, STORED, STRING, TEXT};
-use tantivy::Index;
-
-fn main() -> tantivy::Result<()> {
-    // # Defining the schema
-    let mut schema_builder = Schema::builder();
-    schema_builder.add_date_field("timestamp", FAST | STORED);
-    let event_type = schema_builder.add_text_field("event_type", STRING | STORED);
-    let attributes = schema_builder.add_json_field("attributes", STORED | TEXT);
-    let schema = schema_builder.build();
-
-    // # Indexing documents
-    let index = Index::create_in_ram(schema.clone());
-
-    let mut index_writer = index.writer(50_000_000)?;
-    let doc = schema.parse_document(
-        r#"{
-        "timestamp": "2022-02-22T23:20:50.53Z",
-        "event_type": "click",
-        "attributes": {
-            "target": "submit-button",
-            "cart": {"product_id": 103},
-            "description": "the best vacuum cleaner ever"
-        }
-    }"#,
-    )?;
-    index_writer.add_document(doc)?;
-    let doc = schema.parse_document(
-        r#"{
-        "timestamp": "2022-02-22T23:20:51.53Z",
-        "event_type": "click",
-        "attributes": {
-            "target": "submit-button",
-            "cart": {"product_id": 133},
-            "description": "das keyboard",
-            "event_type": "holiday-sale"
-        }
-    }"#,
-    )?;
-    index_writer.add_document(doc)?;
-    index_writer.commit()?;
-
-    let reader = index.reader()?;
-    let searcher = reader.searcher();
-
-    // # Default fields: event_type and attributes
-    // By setting attributes as a default field it allows omitting attributes itself, e.g. "target",
-    // instead of "attributes.target"
-    let query_parser = QueryParser::for_index(&index, vec![event_type, attributes]);
-    {
-        let query = query_parser.parse_query("target:submit-button")?;
-        let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(count_docs.len(), 2);
-    }
-    {
-        let query = query_parser.parse_query("target:submit")?;
-        let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(count_docs.len(), 2);
-    }
-    {
-        let query = query_parser.parse_query("cart.product_id:103")?;
-        let count_docs = searcher.search(&*query, &Count)?;
-        assert_eq!(count_docs, 1);
-    }
-    {
-        let query = query_parser.parse_query("click AND cart.product_id:133")?;
-        let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(hits.len(), 1);
-    }
-    {
-        // The sub-fields in the json field marked as default field still need to be explicitly
-        // addressed
-        let query = query_parser.parse_query("click AND 133")?;
-        let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(hits.len(), 0);
-    }
-    {
-        // Default json fields are ignored if they collide with the schema
-        let query = query_parser.parse_query("event_type:holiday-sale")?;
-        let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(hits.len(), 0);
-    }
-    // # Query via full attribute path
-    {
-        // This only searches in our schema's `event_type` field
-        let query = query_parser.parse_query("event_type:click")?;
-        let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(hits.len(), 2);
-    }
-    {
-        // Default json fields can still be accessed by full path
-        let query = query_parser.parse_query("attributes.event_type:holiday-sale")?;
-        let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
-        assert_eq!(hits.len(), 1);
-    }
-    Ok(())
-}

examples/multiple_producer.rs

@@ -1,104 +0,0 @@
-// # Indexing from different threads.
-//
-// It is fairly common to have to index from different threads.
-// Tantivy forbids to create more than one `IndexWriter` at a time.
-//
-// This `IndexWriter` itself has its own multithreaded layer, so managing your own
-// indexing threads will not help. However, it can still be useful for some applications.
-//
-// For instance, if preparing documents to send to tantivy before indexing is the bottleneck of
-// your application, it is reasonable to have multiple threads.
-//
-// Another very common reason to want to index from multiple threads, is implementing a webserver
-// with CRUD capabilities. The server framework will most likely handle request from
-// different threads.
-//
-// The recommended way to address both of these use case is to wrap your `IndexWriter` into a
-// `Arc<RwLock<IndexWriter>>`.
-//
-// While this is counterintuitive, adding and deleting documents do not require mutability
-// over the `IndexWriter`, so several threads will be able to do this operation concurrently.
-//
-// The example below does not represent an actual real-life use case (who would spawn thread to
-// index a single document?), but aims at demonstrating the mechanism that makes indexing
-// from several threads possible.
-
-// ---
-// Importing tantivy...
-use std::sync::{Arc, RwLock};
-use std::thread;
-use std::time::Duration;
-
-use tantivy::schema::{Schema, STORED, TEXT};
-use tantivy::{doc, Index, IndexWriter, Opstamp, TantivyError};
-
-fn main() -> tantivy::Result<()> {
-    // # Defining the schema
-    let mut schema_builder = Schema::builder();
-    let title = schema_builder.add_text_field("title", TEXT | STORED);
-    let body = schema_builder.add_text_field("body", TEXT);
-    let schema = schema_builder.build();
-
-    let index = Index::create_in_ram(schema);
-    let index_writer: Arc<RwLock<IndexWriter>> = Arc::new(RwLock::new(index.writer(50_000_000)?));
-
-    // # First indexing thread.
-    let index_writer_clone_1 = index_writer.clone();
-    thread::spawn(move || {
-        // we index 100 times the document... for the sake of the example.
-        for i in 0..100 {
-            let opstamp = index_writer_clone_1
-                .read().unwrap() //< A read lock is sufficient here.
-                .add_document(
-                    doc!(
-                        title => "Of Mice and Men",
-                        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-                        bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-                        over the yellow sands in the sunlight before reaching the narrow pool. On one \
-                        side of the river the golden foothill slopes curve up to the strong and rocky \
-                        Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-                        fresh and green with every spring, carrying in their lower leaf junctures the \
-                        debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-                        limbs and branches that arch over the pool"
-                    ))?;
-            println!("add doc {} from thread 1 - opstamp {}", i, opstamp);
-            thread::sleep(Duration::from_millis(20));
-        }
-        Result::<(), TantivyError>::Ok(())
-    });
-
-    // # Second indexing thread.
-    let index_writer_clone_2 = index_writer.clone();
-    // For convenience, tantivy also comes with a macro to
-    // reduce the boilerplate above.
-    thread::spawn(move || {
-        // we index 100 times the document... for the sake of the example.
-        for i in 0..100 {
-            // A read lock is sufficient here.
-            let opstamp = {
-                let index_writer_rlock = index_writer_clone_2.read().unwrap();
-                index_writer_rlock.add_document(doc!(
-                    title => "Manufacturing consent",
-                    body => "Some great book description..."
-                ))?
-            };
-            println!("add doc {} from thread 2 - opstamp {}", i, opstamp);
-            thread::sleep(Duration::from_millis(10));
-        }
-        Result::<(), TantivyError>::Ok(())
-    });
-
-    // # In the main thread, we commit 10 times, once every 500ms.
-    for _ in 0..10 {
-        let opstamp: Opstamp = {
-            // Committing or rollbacking on the other hand requires write lock. This will block
-            // other threads.
-            let mut index_writer_wlock = index_writer.write().unwrap();
-            index_writer_wlock.commit()?
-        };
-        println!("committed with opstamp {}", opstamp);
-        thread::sleep(Duration::from_millis(500));
-    }
-
-    Ok(())
-}

examples/pre_tokenized_text.rs

@@ -1,135 +0,0 @@
-// # Pre-tokenized text example
-//
-// This example shows how to use pre-tokenized text. Sometimes you might
-// want to index and search through text which is already split into
-// tokens by some external tool.
-//
-// In this example we will:
-// - use tantivy tokenizer to create tokens and load them directly into tantivy,
-// - import tokenized text straight from json,
-// - perform a search on documents with pre-tokenized text
-
-use tantivy::collector::{Count, TopDocs};
-use tantivy::query::TermQuery;
-use tantivy::schema::*;
-use tantivy::tokenizer::{PreTokenizedString, SimpleTokenizer, Token, Tokenizer};
-use tantivy::{doc, Index, ReloadPolicy};
-use tempfile::TempDir;
-
-fn pre_tokenize_text(text: &str) -> Vec<Token> {
-    let mut token_stream = SimpleTokenizer.token_stream(text);
-    let mut tokens = vec![];
-    while token_stream.advance() {
-        tokens.push(token_stream.token().clone());
-    }
-    tokens
-}
-
-fn main() -> tantivy::Result<()> {
-    let index_path = TempDir::new()?;
-
-    let mut schema_builder = Schema::builder();
-
-    schema_builder.add_text_field("title", TEXT | STORED);
-    schema_builder.add_text_field("body", TEXT);
-
-    let schema = schema_builder.build();
-
-    let index = Index::create_in_dir(&index_path, schema.clone())?;
-
-    let mut index_writer = index.writer(50_000_000)?;
-
-    // We can create a document manually, by setting the fields
-    // one by one in a Document object.
-    let title = schema.get_field("title").unwrap();
-    let body = schema.get_field("body").unwrap();
-
-    let title_text = "The Old Man and the Sea";
-    let body_text = "He was an old man who fished alone in a skiff in the Gulf Stream";
-
-    // Content of our first document
-    // We create `PreTokenizedString` which contains original text and vector of tokens
-    let title_tok = PreTokenizedString {
-        text: String::from(title_text),
-        tokens: pre_tokenize_text(title_text),
-    };
-
-    println!(
-        "Original text: \"{}\" and tokens: {:?}",
-        title_tok.text, title_tok.tokens
-    );
-
-    let body_tok = PreTokenizedString {
-        text: String::from(body_text),
-        tokens: pre_tokenize_text(body_text),
-    };
-
-    // Now lets create a document and add our `PreTokenizedString`
-    let old_man_doc = doc!(title => title_tok, body => body_tok);
-
-    // ... now let's just add it to the IndexWriter
-    index_writer.add_document(old_man_doc)?;
-
-    // Pretokenized text can also be fed as JSON
-    let short_man_json = r#"{
-        "title":[{
-            "text":"The Old Man",
-            "tokens":[
-                {"offset_from":0,"offset_to":3,"position":0,"text":"The","position_length":1},
-                {"offset_from":4,"offset_to":7,"position":1,"text":"Old","position_length":1},
-                {"offset_from":8,"offset_to":11,"position":2,"text":"Man","position_length":1}
-            ]
-        }]
-    }"#;
-
-    let short_man_doc = schema.parse_document(short_man_json)?;
-
-    index_writer.add_document(short_man_doc)?;
-
-    // Let's commit changes
-    index_writer.commit()?;
-
-    // ... and now is the time to query our index
-
-    let reader = index
-        .reader_builder()
-        .reload_policy(ReloadPolicy::OnCommit)
-        .try_into()?;
-
-    let searcher = reader.searcher();
-
-    // We want to get documents with token "Man", we will use TermQuery to do it
-    // Using PreTokenizedString means the tokens are stored as is avoiding stemming
-    // and lowercasing, which preserves full words in their original form
-    let query = TermQuery::new(
-        Term::from_field_text(title, "Man"),
-        IndexRecordOption::Basic,
-    );
-
-    let (top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count))?;
-
-    assert_eq!(count, 2);
-
-    // Now let's print out the results.
-    // Note that the tokens are not stored along with the original text
-    // in the document store
-    for (_score, doc_address) in top_docs {
-        let retrieved_doc = searcher.doc(doc_address)?;
-        println!("Document: {}", schema.to_json(&retrieved_doc));
-    }
-
-    // In contrary to the previous query, when we search for the "man" term we
-    // should get no results, as it's not one of the indexed tokens. SimpleTokenizer
-    // only splits text on whitespace / punctuation.
-
-    let query = TermQuery::new(
-        Term::from_field_text(title, "man"),
-        IndexRecordOption::Basic,
-    );
-
-    let (_top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count))?;
-
-    assert_eq!(count, 0);
-
-    Ok(())
-}

examples/snippet.rs

@@ -4,19 +4,23 @@
 // your hit result.
 // Snippet are an extracted of a target document, and returned in HTML format.
 // The keyword searched by the user are highlighted with a `<b>` tag.
+extern crate tempdir;
 
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
-use tantivy::{doc, Index, Snippet, SnippetGenerator};
-use tempfile::TempDir;
+use tantivy::Index;
+use tantivy::{Snippet, SnippetGenerator};
+use tempdir::TempDir;
 
 fn main() -> tantivy::Result<()> {
     // Let's create a temporary directory for the
     // sake of this example
-    let index_path = TempDir::new()?;
+    let index_path = TempDir::new("tantivy_example_dir")?;
 
     // # Defining the schema
     let mut schema_builder = Schema::builder();
@@ -25,7 +29,7 @@ fn main() -> tantivy::Result<()> {
     let schema = schema_builder.build();
 
     // # Indexing documents
-    let index = Index::create_in_dir(&index_path, schema)?;
+    let index = Index::create_in_dir(&index_path, schema.clone())?;
 
     let mut index_writer = index.writer(50_000_000)?;
 
@@ -40,7 +44,7 @@ fn main() -> tantivy::Result<()> {
             fresh and green with every spring, carrying in their lower leaf junctures the \
             debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
             limbs and branches that arch over the pool"
-    ))?;
+    ));
     // ...
     index_writer.commit()?;
 
@@ -57,10 +61,7 @@ fn main() -> tantivy::Result<()> {
         let doc = searcher.doc(doc_address)?;
         let snippet = snippet_generator.snippet_from_doc(&doc);
         println!("Document score {}:", score);
-        println!(
-            "title: {}",
-            doc.get_first(title).unwrap().as_text().unwrap()
-        );
+        println!("title: {}", doc.get_first(title).unwrap().text().unwrap());
         println!("snippet: {}", snippet.to_html());
         println!("custom highlighting: {}", highlight(snippet));
     }
@@ -72,14 +73,14 @@ fn highlight(snippet: Snippet) -> String {
     let mut result = String::new();
     let mut start_from = 0;
 
-    for fragment_range in snippet.highlighted() {
-        result.push_str(&snippet.fragment()[start_from..fragment_range.start]);
+    for (start, end) in snippet.highlighted().iter().map(|h| h.bounds()) {
+        result.push_str(&snippet.fragments()[start_from..start]);
         result.push_str(" --> ");
-        result.push_str(&snippet.fragment()[fragment_range.clone()]);
+        result.push_str(&snippet.fragments()[start..end]);
         result.push_str(" <-- ");
-        start_from = fragment_range.end;
+        start_from = end;
     }
 
-    result.push_str(&snippet.fragment()[start_from..]);
+    result.push_str(&snippet.fragments()[start_from..]);
     result
 }

examples/stop_words.rs

@@ -9,13 +9,17 @@
 // - add a few stop words
 // - index few documents in our index
 
+extern crate tempdir;
+
 // ---
 // Importing tantivy...
+#[macro_use]
+extern crate tantivy;
 use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::*;
-use tantivy::{doc, Index};
+use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
     // this example assumes you understand the content in `basic_search`
@@ -50,7 +54,7 @@ fn main() -> tantivy::Result<()> {
 
     // This tokenizer lowers all of the text (to help with stop word matching)
     // then removes all instances of `the` and `and` from the corpus
-    let tokenizer = TextAnalyzer::from(SimpleTokenizer)
+    let tokenizer = SimpleTokenizer
         .filter(LowerCaser)
         .filter(StopWordFilter::remove(vec![
             "the".to_string(),
@@ -68,7 +72,7 @@ fn main() -> tantivy::Result<()> {
     title => "The Old Man and the Sea",
     body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
      he had gone eighty-four days now without taking a fish."
-    ))?;
+    ));
 
     index_writer.add_document(doc!(
     title => "Of Mice and Men",
@@ -80,7 +84,7 @@ fn main() -> tantivy::Result<()> {
             fresh and green with every spring, carrying in their lower leaf junctures the \
             debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
             limbs and branches that arch over the pool"
-    ))?;
+    ));
 
     index_writer.add_document(doc!(
     title => "Frankenstein",
@@ -88,7 +92,7 @@ fn main() -> tantivy::Result<()> {
              enterprise which you have regarded with such evil forebodings.  I arrived here \
              yesterday, and my first task is to assure my dear sister of my welfare and \
              increasing confidence in the success of my undertaking."
-    ))?;
+    ));
 
     index_writer.commit()?;
 

examples/warmer.rs

@@ -1,219 +0,0 @@
-use std::cmp::Reverse;
-use std::collections::{HashMap, HashSet};
-use std::sync::{Arc, RwLock, Weak};
-
-use tantivy::collector::TopDocs;
-use tantivy::query::QueryParser;
-use tantivy::schema::{Field, Schema, FAST, TEXT};
-use tantivy::{
-    doc, DocAddress, DocId, Index, IndexReader, Opstamp, Searcher, SearcherGeneration, SegmentId,
-    SegmentReader, Warmer,
-};
-
-// This example shows how warmers can be used to
-// load a values from an external sources using the Warmer API.
-//
-// In this example, we assume an e-commerce search engine.
-
-type ProductId = u64;
-
-/// Price
-type Price = u32;
-
-pub trait PriceFetcher: Send + Sync + 'static {
-    fn fetch_prices(&self, product_ids: &[ProductId]) -> Vec<Price>;
-}
-
-struct DynamicPriceColumn {
-    field: Field,
-    price_cache: RwLock<HashMap<(SegmentId, Option<Opstamp>), Arc<Vec<Price>>>>,
-    price_fetcher: Box<dyn PriceFetcher>,
-}
-
-impl DynamicPriceColumn {
-    pub fn with_product_id_field<T: PriceFetcher>(field: Field, price_fetcher: T) -> Self {
-        DynamicPriceColumn {
-            field,
-            price_cache: Default::default(),
-            price_fetcher: Box::new(price_fetcher),
-        }
-    }
-
-    pub fn price_for_segment(&self, segment_reader: &SegmentReader) -> Option<Arc<Vec<Price>>> {
-        let segment_key = (segment_reader.segment_id(), segment_reader.delete_opstamp());
-        self.price_cache.read().unwrap().get(&segment_key).cloned()
-    }
-}
-impl Warmer for DynamicPriceColumn {
-    fn warm(&self, searcher: &Searcher) -> tantivy::Result<()> {
-        for segment in searcher.segment_readers() {
-            let key = (segment.segment_id(), segment.delete_opstamp());
-            let product_id_reader = segment.fast_fields().u64(self.field)?;
-            let product_ids: Vec<ProductId> = segment
-                .doc_ids_alive()
-                .map(|doc| product_id_reader.get_val(doc))
-                .collect();
-            let mut prices_it = self.price_fetcher.fetch_prices(&product_ids).into_iter();
-            let mut price_vals: Vec<Price> = Vec::new();
-            for doc in 0..segment.max_doc() {
-                if segment.is_deleted(doc) {
-                    price_vals.push(0);
-                } else {
-                    price_vals.push(prices_it.next().unwrap())
-                }
-            }
-            self.price_cache
-                .write()
-                .unwrap()
-                .insert(key, Arc::new(price_vals));
-        }
-        Ok(())
-    }
-
-    fn garbage_collect(&self, live_generations: &[&SearcherGeneration]) {
-        let live_segment_id_and_delete_ops: HashSet<(SegmentId, Option<Opstamp>)> =
-            live_generations
-                .iter()
-                .flat_map(|gen| gen.segments())
-                .map(|(&segment_id, &opstamp)| (segment_id, opstamp))
-                .collect();
-        let mut price_cache_wrt = self.price_cache.write().unwrap();
-        // let price_cache = std::mem::take(&mut *price_cache_wrt);
-        // Drain would be nicer here.
-        *price_cache_wrt = std::mem::take(&mut *price_cache_wrt)
-            .into_iter()
-            .filter(|(seg_id_and_op, _)| !live_segment_id_and_delete_ops.contains(seg_id_and_op))
-            .collect();
-    }
-}
-
-/// For the sake of this example, the table is just an editable HashMap behind a RwLock.
-/// This map represents a map (ProductId -> Price)
-///
-/// In practise, it could be fetching things from an external service, like a SQL table.
-#[derive(Default, Clone)]
-pub struct ExternalPriceTable {
-    prices: Arc<RwLock<HashMap<ProductId, Price>>>,
-}
-
-impl ExternalPriceTable {
-    pub fn update_price(&self, product_id: ProductId, price: Price) {
-        let mut prices_wrt = self.prices.write().unwrap();
-        prices_wrt.insert(product_id, price);
-    }
-}
-
-impl PriceFetcher for ExternalPriceTable {
-    fn fetch_prices(&self, product_ids: &[ProductId]) -> Vec<Price> {
-        let prices_read = self.prices.read().unwrap();
-        product_ids
-            .iter()
-            .map(|product_id| prices_read.get(product_id).cloned().unwrap_or(0))
-            .collect()
-    }
-}
-
-fn main() -> tantivy::Result<()> {
-    // Declaring our schema.
-    let mut schema_builder = Schema::builder();
-    // The product id is assumed to be a primary id for our external price source.
-    let product_id = schema_builder.add_u64_field("product_id", FAST);
-    let text = schema_builder.add_text_field("text", TEXT);
-    let schema: Schema = schema_builder.build();
-
-    let price_table = ExternalPriceTable::default();
-    let price_dynamic_column = Arc::new(DynamicPriceColumn::with_product_id_field(
-        product_id,
-        price_table.clone(),
-    ));
-    price_table.update_price(OLIVE_OIL, 12);
-    price_table.update_price(GLOVES, 13);
-    price_table.update_price(SNEAKERS, 80);
-
-    const OLIVE_OIL: ProductId = 323423;
-    const GLOVES: ProductId = 3966623;
-    const SNEAKERS: ProductId = 23222;
-
-    let index = Index::create_in_ram(schema);
-    let mut writer = index.writer_with_num_threads(1, 10_000_000)?;
-    writer.add_document(doc!(product_id=>OLIVE_OIL, text=>"cooking olive oil from greece"))?;
-    writer.add_document(doc!(product_id=>GLOVES, text=>"kitchen gloves, perfect for cooking"))?;
-    writer.add_document(doc!(product_id=>SNEAKERS, text=>"uber sweet sneakers"))?;
-    writer.commit()?;
-
-    let warmers: Vec<Weak<dyn Warmer>> = vec![Arc::downgrade(
-        &(price_dynamic_column.clone() as Arc<dyn Warmer>),
-    )];
-    let reader: IndexReader = index.reader_builder().warmers(warmers).try_into()?;
-    reader.reload()?;
-
-    let query_parser = QueryParser::for_index(&index, vec![text]);
-    let query = query_parser.parse_query("cooking")?;
-
-    let searcher = reader.searcher();
-    let score_by_price = move |segment_reader: &SegmentReader| {
-        let price = price_dynamic_column
-            .price_for_segment(segment_reader)
-            .unwrap();
-        move |doc_id: DocId| Reverse(price[doc_id as usize])
-    };
-
-    let most_expensive_first = TopDocs::with_limit(10).custom_score(score_by_price);
-
-    let hits = searcher.search(&query, &most_expensive_first)?;
-    assert_eq!(
-        &hits,
-        &[
-            (
-                Reverse(12u32),
-                DocAddress {
-                    segment_ord: 0,
-                    doc_id: 0u32
-                }
-            ),
-            (
-                Reverse(13u32),
-                DocAddress {
-                    segment_ord: 0,
-                    doc_id: 1u32
-                }
-            ),
-        ]
-    );
-
-    // Olive oil just got more expensive!
-    price_table.update_price(OLIVE_OIL, 15);
-
-    // The price update are directly reflected on `reload`.
-    //
-    // Be careful here though!...
-    // You may have spotted that we are still using the same `Searcher`.
-    //
-    // It is up to the `Warmer` implementer to decide how
-    // to control this behavior.
-
-    reader.reload()?;
-
-    let hits_with_new_prices = searcher.search(&query, &most_expensive_first)?;
-    assert_eq!(
-        &hits_with_new_prices,
-        &[
-            (
-                Reverse(13u32),
-                DocAddress {
-                    segment_ord: 0,
-                    doc_id: 1u32
-                }
-            ),
-            (
-                Reverse(15u32),
-                DocAddress {
-                    segment_ord: 0,
-                    doc_id: 0u32
-                }
-            ),
-        ]
-    );
-
-    Ok(())
-}

examples/working_with_json.rs

@@ -1,3 +1,4 @@
+extern crate tantivy;
 use tantivy::schema::*;
 
 // # Document from json
@@ -21,7 +22,7 @@ fn main() -> tantivy::Result<()> {
     }"#;
 
     // We can parse our document
-    let _mice_and_men_doc = schema.parse_document(mice_and_men_doc_json)?;
+    let _mice_and_men_doc = schema.parse_document(&mice_and_men_doc_json)?;
 
     // Multi-valued field are allowed, they are
     // expressed in JSON by an array.
@@ -30,7 +31,7 @@ fn main() -> tantivy::Result<()> {
        "title": ["Frankenstein", "The Modern Prometheus"],
        "year": 1818
     }"#;
-    let _frankenstein_doc = schema.parse_document(frankenstein_json)?;
+    let _frankenstein_doc = schema.parse_document(&frankenstein_json)?;
 
     // Note that the schema is saved in your index directory.
     //

fastfield_codecs/Cargo.toml

@@ -1,35 +0,0 @@
-[package]
-name = "fastfield_codecs"
-version = "0.3.0"
-authors = ["Pascal Seitz <pascal@quickwit.io>"]
-license = "MIT"
-edition = "2021"
-description = "Fast field codecs used by tantivy"
-documentation = "https://docs.rs/fastfield_codecs/"
-homepage = "https://github.com/quickwit-oss/tantivy"
-repository = "https://github.com/quickwit-oss/tantivy"
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-common = { version = "0.5", path = "../common/", package = "tantivy-common" }
-tantivy-bitpacker = { version= "0.3", path = "../bitpacker/" }
-ownedbytes = { version = "0.5", path = "../ownedbytes" }
-prettytable-rs = {version="0.9.0", optional= true}
-rand = {version="0.8.3", optional= true}
-fastdivide = "0.4"
-log = "0.4"
-itertools = { version = "0.10.3" }
-measure_time = { version="0.8.2", optional=true}
-ordered-float = "3.4"
-
-[dev-dependencies]
-more-asserts = "0.3.0"
-proptest = "1.0.0"
-rand = "0.8.3"
-
-[features]
-bin = ["prettytable-rs", "rand", "measure_time"]
-default = ["bin"]
-unstable = []
-

fastfield_codecs/README.md

@@ -1,68 +0,0 @@
-
-
-# Fast Field Codecs
-
-This crate contains various fast field codecs, used to compress/decompress fast field data in tantivy.
-
-## Contributing
-
-Contributing is pretty straightforward. Since the bitpacking is the simplest compressor, you can check it for reference.
-
-A codec needs to implement 2 traits:
-
-- A reader implementing `FastFieldCodecReader` to read the codec.
-- A serializer implementing `FastFieldCodecSerializer` for compression estimation and codec name + id.
-
-### Tests
-
-Once the traits are implemented test and benchmark integration is pretty easy (see `test_with_codec_data_sets` and `bench.rs`).
-
-Make sure to add the codec to the main.rs, which tests the compression ratio and estimation against different data sets. You can run it with:
-```
-cargo run --features bin
-```
-
-### TODO
-- Add real world data sets in comparison
-- Add codec to cover sparse data sets
-
-
-### Codec Comparison
-```
-+----------------------------------+-------------------+------------------------+
-|                                  | Compression Ratio | Compression Estimation |
-+----------------------------------+-------------------+------------------------+
-| Autoincrement                    |                   |                        |
-+----------------------------------+-------------------+------------------------+
-| LinearInterpol                   | 0.000039572664    | 0.000004396963         |
-+----------------------------------+-------------------+------------------------+
-| MultiLinearInterpol              | 0.1477348         | 0.17275847             |
-+----------------------------------+-------------------+------------------------+
-| Bitpacked                        | 0.28126493        | 0.28125                |
-+----------------------------------+-------------------+------------------------+
-| Monotonically increasing concave |                   |                        |
-+----------------------------------+-------------------+------------------------+
-| LinearInterpol                   | 0.25003937        | 0.26562938             |
-+----------------------------------+-------------------+------------------------+
-| MultiLinearInterpol              | 0.190665          | 0.1883836              |
-+----------------------------------+-------------------+------------------------+
-| Bitpacked                        | 0.31251436        | 0.3125                 |
-+----------------------------------+-------------------+------------------------+
-| Monotonically increasing convex  |                   |                        |
-+----------------------------------+-------------------+------------------------+
-| LinearInterpol                   | 0.25003937        | 0.28125438             |
-+----------------------------------+-------------------+------------------------+
-| MultiLinearInterpol              | 0.18676           | 0.2040086              |
-+----------------------------------+-------------------+------------------------+
-| Bitpacked                        | 0.31251436        | 0.3125                 |
-+----------------------------------+-------------------+------------------------+
-| Almost monotonically increasing  |                   |                        |
-+----------------------------------+-------------------+------------------------+
-| LinearInterpol                   | 0.14066513        | 0.1562544              |
-+----------------------------------+-------------------+------------------------+
-| MultiLinearInterpol              | 0.16335973        | 0.17275847             |
-+----------------------------------+-------------------+------------------------+
-| Bitpacked                        | 0.28126493        | 0.28125                |
-+----------------------------------+-------------------+------------------------+
-
-```

fastfield_codecs/benches/bench.rs

@@ -1,246 +0,0 @@
-#![feature(test)]
-
-extern crate test;
-
-#[cfg(test)]
-mod tests {
-    use std::iter;
-    use std::sync::Arc;
-
-    use fastfield_codecs::*;
-    use ownedbytes::OwnedBytes;
-    use rand::prelude::*;
-    use test::Bencher;
-
-    use super::*;
-
-    // Warning: this generates the same permutation at each call
-    fn generate_permutation() -> Vec<u64> {
-        let mut permutation: Vec<u64> = (0u64..100_000u64).collect();
-        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
-        permutation
-    }
-
-    fn generate_random() -> Vec<u64> {
-        let mut permutation: Vec<u64> = (0u64..100_000u64)
-            .map(|el| el + random::<u16>() as u64)
-            .collect();
-        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
-        permutation
-    }
-
-    // Warning: this generates the same permutation at each call
-    fn generate_permutation_gcd() -> Vec<u64> {
-        let mut permutation: Vec<u64> = (1u64..100_000u64).map(|el| el * 1000).collect();
-        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
-        permutation
-    }
-
-    pub fn serialize_and_load<T: MonotonicallyMappableToU64 + Ord + Default>(
-        column: &[T],
-    ) -> Arc<dyn Column<T>> {
-        let mut buffer = Vec::new();
-        serialize(VecColumn::from(&column), &mut buffer, &ALL_CODEC_TYPES).unwrap();
-        open(OwnedBytes::new(buffer)).unwrap()
-    }
-
-    #[bench]
-    fn bench_intfastfield_jumpy_veclookup(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        let n = permutation.len();
-        b.iter(|| {
-            let mut a = 0u64;
-            for _ in 0..n {
-                a = permutation[a as usize];
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_jumpy_fflookup(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        let n = permutation.len();
-        let column: Arc<dyn Column<u64>> = serialize_and_load(&permutation);
-        b.iter(|| {
-            let mut a = 0u64;
-            for _ in 0..n {
-                a = column.get_val(a as u32);
-            }
-            a
-        });
-    }
-
-    fn get_exp_data() -> Vec<u64> {
-        let mut data = vec![];
-        for i in 0..100 {
-            let num = i * i;
-            data.extend(iter::repeat(i as u64).take(num));
-        }
-        data.shuffle(&mut StdRng::from_seed([1u8; 32]));
-
-        // lengt = 328350
-        data
-    }
-
-    fn get_data_50percent_item() -> (u128, u128, Vec<u128>) {
-        let mut permutation = get_exp_data();
-        let major_item = 20;
-        let minor_item = 10;
-        permutation.extend(iter::repeat(major_item).take(permutation.len()));
-        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
-        let permutation = permutation.iter().map(|el| *el as u128).collect::<Vec<_>>();
-        (major_item as u128, minor_item as u128, permutation)
-    }
-    fn get_u128_column_random() -> Arc<dyn Column<u128>> {
-        let permutation = generate_random();
-        let permutation = permutation.iter().map(|el| *el as u128).collect::<Vec<_>>();
-        get_u128_column_from_data(&permutation)
-    }
-
-    fn get_u128_column_from_data(data: &[u128]) -> Arc<dyn Column<u128>> {
-        let mut out = vec![];
-        let iter_gen = || data.iter().cloned();
-        serialize_u128(iter_gen, data.len() as u32, &mut out).unwrap();
-        let out = OwnedBytes::new(out);
-        open_u128::<u128>(out).unwrap()
-    }
-
-    #[bench]
-    fn bench_intfastfield_getrange_u128_50percent_hit(b: &mut Bencher) {
-        let (major_item, _minor_item, data) = get_data_50percent_item();
-        let column = get_u128_column_from_data(&data);
-
-        b.iter(|| {
-            let mut positions = Vec::new();
-            column.get_docids_for_value_range(
-                major_item..=major_item,
-                0..data.len() as u32,
-                &mut positions,
-            );
-            positions
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_getrange_u128_single_hit(b: &mut Bencher) {
-        let (_major_item, minor_item, data) = get_data_50percent_item();
-        let column = get_u128_column_from_data(&data);
-
-        b.iter(|| {
-            let mut positions = Vec::new();
-            column.get_docids_for_value_range(
-                minor_item..=minor_item,
-                0..data.len() as u32,
-                &mut positions,
-            );
-            positions
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_getrange_u128_hit_all(b: &mut Bencher) {
-        let (_major_item, _minor_item, data) = get_data_50percent_item();
-        let column = get_u128_column_from_data(&data);
-
-        b.iter(|| {
-            let mut positions = Vec::new();
-            column.get_docids_for_value_range(0..=u128::MAX, 0..data.len() as u32, &mut positions);
-            positions
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_scan_all_fflookup_u128(b: &mut Bencher) {
-        let column = get_u128_column_random();
-
-        b.iter(|| {
-            let mut a = 0u128;
-            for i in 0u64..column.num_vals() as u64 {
-                a += column.get_val(i as u32);
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_jumpy_stride5_u128(b: &mut Bencher) {
-        let column = get_u128_column_random();
-
-        b.iter(|| {
-            let n = column.num_vals();
-            let mut a = 0u128;
-            for i in (0..n / 5).map(|val| val * 5) {
-                a += column.get_val(i);
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_stride7_vec(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        let n = permutation.len();
-        b.iter(|| {
-            let mut a = 0u64;
-            for i in (0..n / 7).map(|val| val * 7) {
-                a += permutation[i as usize];
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_stride7_fflookup(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        let n = permutation.len();
-        let column: Arc<dyn Column<u64>> = serialize_and_load(&permutation);
-        b.iter(|| {
-            let mut a = 0;
-            for i in (0..n / 7).map(|val| val * 7) {
-                a += column.get_val(i as u32);
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_scan_all_fflookup(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        let n = permutation.len();
-        let column: Arc<dyn Column<u64>> = serialize_and_load(&permutation);
-        b.iter(|| {
-            let mut a = 0u64;
-            for i in 0u32..n as u32 {
-                a += column.get_val(i);
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_scan_all_fflookup_gcd(b: &mut Bencher) {
-        let permutation = generate_permutation_gcd();
-        let n = permutation.len();
-        let column: Arc<dyn Column<u64>> = serialize_and_load(&permutation);
-        b.iter(|| {
-            let mut a = 0u64;
-            for i in 0..n {
-                a += column.get_val(i as u32);
-            }
-            a
-        });
-    }
-
-    #[bench]
-    fn bench_intfastfield_scan_all_vec(b: &mut Bencher) {
-        let permutation = generate_permutation();
-        b.iter(|| {
-            let mut a = 0u64;
-            for i in 0..permutation.len() {
-                a += permutation[i as usize] as u64;
-            }
-            a
-        });
-    }
-}

fastfield_codecs/src/bitpacked.rs

@@ -1,116 +0,0 @@
-use std::io::{self, Write};
-
-use ownedbytes::OwnedBytes;
-use tantivy_bitpacker::{compute_num_bits, BitPacker, BitUnpacker};
-
-use crate::serialize::NormalizedHeader;
-use crate::{Column, FastFieldCodec, FastFieldCodecType};
-
-/// Depending on the field type, a different
-/// fast field is required.
-#[derive(Clone)]
-pub struct BitpackedReader {
-    data: OwnedBytes,
-    bit_unpacker: BitUnpacker,
-    normalized_header: NormalizedHeader,
-}
-
-impl Column for BitpackedReader {
-    #[inline]
-    fn get_val(&self, doc: u32) -> u64 {
-        self.bit_unpacker.get(doc, &self.data)
-    }
-    #[inline]
-    fn min_value(&self) -> u64 {
-        // The BitpackedReader assumes a normalized vector.
-        0
-    }
-    #[inline]
-    fn max_value(&self) -> u64 {
-        self.normalized_header.max_value
-    }
-    #[inline]
-    fn num_vals(&self) -> u32 {
-        self.normalized_header.num_vals
-    }
-}
-
-pub struct BitpackedCodec;
-
-impl FastFieldCodec for BitpackedCodec {
-    /// The CODEC_TYPE is an enum value used for serialization.
-    const CODEC_TYPE: FastFieldCodecType = FastFieldCodecType::Bitpacked;
-
-    type Reader = BitpackedReader;
-
-    /// Opens a fast field given a file.
-    fn open_from_bytes(
-        data: OwnedBytes,
-        normalized_header: NormalizedHeader,
-    ) -> io::Result<Self::Reader> {
-        let num_bits = compute_num_bits(normalized_header.max_value);
-        let bit_unpacker = BitUnpacker::new(num_bits);
-        Ok(BitpackedReader {
-            data,
-            bit_unpacker,
-            normalized_header,
-        })
-    }
-
-    /// Serializes data with the BitpackedFastFieldSerializer.
-    ///
-    /// The bitpacker assumes that the column has been normalized.
-    /// i.e. It has already been shifted by its minimum value, so that its
-    /// current minimum value is 0.
-    ///
-    /// Ideally, we made a shift upstream on the column so that `col.min_value() == 0`.
-    fn serialize(column: &dyn Column, write: &mut impl Write) -> io::Result<()> {
-        assert_eq!(column.min_value(), 0u64);
-        let num_bits = compute_num_bits(column.max_value());
-        let mut bit_packer = BitPacker::new();
-        for val in column.iter() {
-            bit_packer.write(val, num_bits, write)?;
-        }
-        bit_packer.close(write)?;
-        Ok(())
-    }
-
-    fn estimate(column: &dyn Column) -> Option<f32> {
-        let num_bits = compute_num_bits(column.max_value());
-        let num_bits_uncompressed = 64;
-        Some(num_bits as f32 / num_bits_uncompressed as f32)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use crate::tests::get_codec_test_datasets;
-
-    fn create_and_validate(data: &[u64], name: &str) {
-        crate::tests::create_and_validate::<BitpackedCodec>(data, name);
-    }
-
-    #[test]
-    fn test_with_codec_data_sets() {
-        let data_sets = get_codec_test_datasets();
-        for (mut data, name) in data_sets {
-            create_and_validate(&data, name);
-            data.reverse();
-            create_and_validate(&data, name);
-        }
-    }
-
-    #[test]
-    fn bitpacked_fast_field_rand() {
-        for _ in 0..500 {
-            let mut data = (0..1 + rand::random::<u8>() as usize)
-                .map(|_| rand::random::<i64>() as u64 / 2)
-                .collect::<Vec<_>>();
-            create_and_validate(&data, "rand");
-
-            data.reverse();
-            create_and_validate(&data, "rand");
-        }
-    }
-}

fastfield_codecs/src/blockwise_linear.rs

@@ -1,186 +0,0 @@
-use std::sync::Arc;
-use std::{io, iter};
-
-use common::{BinarySerializable, CountingWriter, DeserializeFrom};
-use ownedbytes::OwnedBytes;
-use tantivy_bitpacker::{compute_num_bits, BitPacker, BitUnpacker};
-
-use crate::line::Line;
-use crate::serialize::NormalizedHeader;
-use crate::{Column, FastFieldCodec, FastFieldCodecType, VecColumn};
-
-const CHUNK_SIZE: usize = 512;
-
-#[derive(Debug, Default)]
-struct Block {
-    line: Line,
-    bit_unpacker: BitUnpacker,
-    data_start_offset: usize,
-}
-
-impl BinarySerializable for Block {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        self.line.serialize(writer)?;
-        self.bit_unpacker.bit_width().serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let line = Line::deserialize(reader)?;
-        let bit_width = u8::deserialize(reader)?;
-        Ok(Block {
-            line,
-            bit_unpacker: BitUnpacker::new(bit_width),
-            data_start_offset: 0,
-        })
-    }
-}
-
-fn compute_num_blocks(num_vals: u32) -> usize {
-    (num_vals as usize + CHUNK_SIZE - 1) / CHUNK_SIZE
-}
-
-pub struct BlockwiseLinearCodec;
-
-impl FastFieldCodec for BlockwiseLinearCodec {
-    const CODEC_TYPE: crate::FastFieldCodecType = FastFieldCodecType::BlockwiseLinear;
-    type Reader = BlockwiseLinearReader;
-
-    fn open_from_bytes(
-        bytes: ownedbytes::OwnedBytes,
-        normalized_header: NormalizedHeader,
-    ) -> io::Result<Self::Reader> {
-        let footer_len: u32 = (&bytes[bytes.len() - 4..]).deserialize()?;
-        let footer_offset = bytes.len() - 4 - footer_len as usize;
-        let (data, mut footer) = bytes.split(footer_offset);
-        let num_blocks = compute_num_blocks(normalized_header.num_vals);
-        let mut blocks: Vec<Block> = iter::repeat_with(|| Block::deserialize(&mut footer))
-            .take(num_blocks)
-            .collect::<io::Result<_>>()?;
-
-        let mut start_offset = 0;
-        for block in &mut blocks {
-            block.data_start_offset = start_offset;
-            start_offset += (block.bit_unpacker.bit_width() as usize) * CHUNK_SIZE / 8;
-        }
-        Ok(BlockwiseLinearReader {
-            blocks: Arc::new(blocks),
-            data,
-            normalized_header,
-        })
-    }
-
-    // Estimate first_chunk and extrapolate
-    fn estimate(column: &dyn crate::Column) -> Option<f32> {
-        if column.num_vals() < 10 * CHUNK_SIZE as u32 {
-            return None;
-        }
-        let mut first_chunk: Vec<u64> = column.iter().take(CHUNK_SIZE).collect();
-        let line = Line::train(&VecColumn::from(&first_chunk));
-        for (i, buffer_val) in first_chunk.iter_mut().enumerate() {
-            let interpolated_val = line.eval(i as u32);
-            *buffer_val = buffer_val.wrapping_sub(interpolated_val);
-        }
-        let estimated_bit_width = first_chunk
-            .iter()
-            .map(|el| ((el + 1) as f32 * 3.0) as u64)
-            .map(compute_num_bits)
-            .max()
-            .unwrap();
-
-        let metadata_per_block = {
-            let mut out = vec![];
-            Block::default().serialize(&mut out).unwrap();
-            out.len()
-        };
-        let num_bits = estimated_bit_width as u64 * column.num_vals() as u64
-            // function metadata per block
-            + metadata_per_block as u64 * (column.num_vals() as u64 / CHUNK_SIZE as u64);
-        let num_bits_uncompressed = 64 * column.num_vals();
-        Some(num_bits as f32 / num_bits_uncompressed as f32)
-    }
-
-    fn serialize(column: &dyn Column, wrt: &mut impl io::Write) -> io::Result<()> {
-        // The BitpackedReader assumes a normalized vector.
-        assert_eq!(column.min_value(), 0);
-        let mut buffer = Vec::with_capacity(CHUNK_SIZE);
-        let num_vals = column.num_vals();
-
-        let num_blocks = compute_num_blocks(num_vals);
-        let mut blocks = Vec::with_capacity(num_blocks);
-
-        let mut vals = column.iter();
-
-        let mut bit_packer = BitPacker::new();
-
-        for _ in 0..num_blocks {
-            buffer.clear();
-            buffer.extend((&mut vals).take(CHUNK_SIZE));
-            let line = Line::train(&VecColumn::from(&buffer));
-
-            assert!(!buffer.is_empty());
-
-            for (i, buffer_val) in buffer.iter_mut().enumerate() {
-                let interpolated_val = line.eval(i as u32);
-                *buffer_val = buffer_val.wrapping_sub(interpolated_val);
-            }
-            let bit_width = buffer.iter().copied().map(compute_num_bits).max().unwrap();
-
-            for &buffer_val in &buffer {
-                bit_packer.write(buffer_val, bit_width, wrt)?;
-            }
-
-            blocks.push(Block {
-                line,
-                bit_unpacker: BitUnpacker::new(bit_width),
-                data_start_offset: 0,
-            });
-        }
-
-        bit_packer.close(wrt)?;
-
-        assert_eq!(blocks.len(), compute_num_blocks(num_vals));
-
-        let mut counting_wrt = CountingWriter::wrap(wrt);
-        for block in &blocks {
-            block.serialize(&mut counting_wrt)?;
-        }
-        let footer_len = counting_wrt.written_bytes();
-        (footer_len as u32).serialize(&mut counting_wrt)?;
-
-        Ok(())
-    }
-}
-
-#[derive(Clone)]
-pub struct BlockwiseLinearReader {
-    blocks: Arc<Vec<Block>>,
-    normalized_header: NormalizedHeader,
-    data: OwnedBytes,
-}
-
-impl Column for BlockwiseLinearReader {
-    #[inline(always)]
-    fn get_val(&self, idx: u32) -> u64 {
-        let block_id = (idx / CHUNK_SIZE as u32) as usize;
-        let idx_within_block = idx % (CHUNK_SIZE as u32);
-        let block = &self.blocks[block_id];
-        let interpoled_val: u64 = block.line.eval(idx_within_block);
-        let block_bytes = &self.data[block.data_start_offset..];
-        let bitpacked_diff = block.bit_unpacker.get(idx_within_block, block_bytes);
-        interpoled_val.wrapping_add(bitpacked_diff)
-    }
-
-    fn min_value(&self) -> u64 {
-        // The BlockwiseLinearReader assumes a normalized vector.
-        0u64
-    }
-
-    fn max_value(&self) -> u64 {
-        self.normalized_header.max_value
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.normalized_header.num_vals
-    }
-}

fastfield_codecs/src/column.rs

@@ -1,348 +0,0 @@
-use std::marker::PhantomData;
-use std::ops::{Range, RangeInclusive};
-
-use tantivy_bitpacker::minmax;
-
-use crate::monotonic_mapping::StrictlyMonotonicFn;
-
-/// `Column` provides columnar access on a field.
-pub trait Column<T: PartialOrd = u64>: Send + Sync {
-    /// Return the value associated with the given idx.
-    ///
-    /// This accessor should return as fast as possible.
-    ///
-    /// # Panics
-    ///
-    /// May panic if `idx` is greater than the column length.
-    fn get_val(&self, idx: u32) -> T;
-
-    /// Fills an output buffer with the fast field values
-    /// associated with the `DocId` going from
-    /// `start` to `start + output.len()`.
-    ///
-    /// # Panics
-    ///
-    /// Must panic if `start + output.len()` is greater than
-    /// the segment's `maxdoc`.
-    #[inline]
-    fn get_range(&self, start: u64, output: &mut [T]) {
-        for (out, idx) in output.iter_mut().zip(start..) {
-            *out = self.get_val(idx as u32);
-        }
-    }
-
-    /// Get the positions of values which are in the provided value range.
-    ///
-    /// Note that position == docid for single value fast fields
-    #[inline]
-    fn get_docids_for_value_range(
-        &self,
-        value_range: RangeInclusive<T>,
-        doc_id_range: Range<u32>,
-        positions: &mut Vec<u32>,
-    ) {
-        let doc_id_range = doc_id_range.start..doc_id_range.end.min(self.num_vals());
-
-        for idx in doc_id_range.start..doc_id_range.end {
-            let val = self.get_val(idx);
-            if value_range.contains(&val) {
-                positions.push(idx);
-            }
-        }
-    }
-
-    /// Returns the minimum value for this fast field.
-    ///
-    /// This min_value may not be exact.
-    /// For instance, the min value does not take in account of possible
-    /// deleted document. All values are however guaranteed to be higher than
-    /// `.min_value()`.
-    fn min_value(&self) -> T;
-
-    /// Returns the maximum value for this fast field.
-    ///
-    /// This max_value may not be exact.
-    /// For instance, the max value does not take in account of possible
-    /// deleted document. All values are however guaranteed to be higher than
-    /// `.max_value()`.
-    fn max_value(&self) -> T;
-
-    /// The number of values in the column.
-    fn num_vals(&self) -> u32;
-
-    /// Returns a iterator over the data
-    fn iter<'a>(&'a self) -> Box<dyn Iterator<Item = T> + 'a> {
-        Box::new((0..self.num_vals()).map(|idx| self.get_val(idx)))
-    }
-}
-
-/// VecColumn provides `Column` over a slice.
-pub struct VecColumn<'a, T = u64> {
-    values: &'a [T],
-    min_value: T,
-    max_value: T,
-}
-
-impl<'a, C: Column<T>, T: Copy + PartialOrd> Column<T> for &'a C {
-    fn get_val(&self, idx: u32) -> T {
-        (*self).get_val(idx)
-    }
-
-    fn min_value(&self) -> T {
-        (*self).min_value()
-    }
-
-    fn max_value(&self) -> T {
-        (*self).max_value()
-    }
-
-    fn num_vals(&self) -> u32 {
-        (*self).num_vals()
-    }
-
-    fn iter<'b>(&'b self) -> Box<dyn Iterator<Item = T> + 'b> {
-        (*self).iter()
-    }
-
-    fn get_range(&self, start: u64, output: &mut [T]) {
-        (*self).get_range(start, output)
-    }
-}
-
-impl<'a, T: Copy + PartialOrd + Send + Sync> Column<T> for VecColumn<'a, T> {
-    fn get_val(&self, position: u32) -> T {
-        self.values[position as usize]
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = T> + '_> {
-        Box::new(self.values.iter().copied())
-    }
-
-    fn min_value(&self) -> T {
-        self.min_value
-    }
-
-    fn max_value(&self) -> T {
-        self.max_value
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.values.len() as u32
-    }
-
-    fn get_range(&self, start: u64, output: &mut [T]) {
-        output.copy_from_slice(&self.values[start as usize..][..output.len()])
-    }
-}
-
-impl<'a, T: Copy + Ord + Default, V> From<&'a V> for VecColumn<'a, T>
-where V: AsRef<[T]> + ?Sized
-{
-    fn from(values: &'a V) -> Self {
-        let values = values.as_ref();
-        let (min_value, max_value) = minmax(values.iter().copied()).unwrap_or_default();
-        Self {
-            values,
-            min_value,
-            max_value,
-        }
-    }
-}
-
-struct MonotonicMappingColumn<C, T, Input> {
-    from_column: C,
-    monotonic_mapping: T,
-    _phantom: PhantomData<Input>,
-}
-
-/// Creates a view of a column transformed by a strictly monotonic mapping. See
-/// [`StrictlyMonotonicFn`].
-///
-/// E.g. apply a gcd monotonic_mapping([100, 200, 300]) == [1, 2, 3]
-/// monotonic_mapping.mapping() is expected to be injective, and we should always have
-/// monotonic_mapping.inverse(monotonic_mapping.mapping(el)) == el
-///
-/// The inverse of the mapping is required for:
-/// `fn get_positions_for_value_range(&self, range: RangeInclusive<T>) -> Vec<u64> `
-/// The user provides the original value range and we need to monotonic map them in the same way the
-/// serialization does before calling the underlying column.
-///
-/// Note that when opening a codec, the monotonic_mapping should be the inverse of the mapping
-/// during serialization. And therefore the monotonic_mapping_inv when opening is the same as
-/// monotonic_mapping during serialization.
-pub fn monotonic_map_column<C, T, Input, Output>(
-    from_column: C,
-    monotonic_mapping: T,
-) -> impl Column<Output>
-where
-    C: Column<Input>,
-    T: StrictlyMonotonicFn<Input, Output> + Send + Sync,
-    Input: PartialOrd + Send + Sync + Clone,
-    Output: PartialOrd + Send + Sync + Clone,
-{
-    MonotonicMappingColumn {
-        from_column,
-        monotonic_mapping,
-        _phantom: PhantomData,
-    }
-}
-
-impl<C, T, Input, Output> Column<Output> for MonotonicMappingColumn<C, T, Input>
-where
-    C: Column<Input>,
-    T: StrictlyMonotonicFn<Input, Output> + Send + Sync,
-    Input: PartialOrd + Send + Sync + Clone,
-    Output: PartialOrd + Send + Sync + Clone,
-{
-    #[inline]
-    fn get_val(&self, idx: u32) -> Output {
-        let from_val = self.from_column.get_val(idx);
-        self.monotonic_mapping.mapping(from_val)
-    }
-
-    fn min_value(&self) -> Output {
-        let from_min_value = self.from_column.min_value();
-        self.monotonic_mapping.mapping(from_min_value)
-    }
-
-    fn max_value(&self) -> Output {
-        let from_max_value = self.from_column.max_value();
-        self.monotonic_mapping.mapping(from_max_value)
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.from_column.num_vals()
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = Output> + '_> {
-        Box::new(
-            self.from_column
-                .iter()
-                .map(|el| self.monotonic_mapping.mapping(el)),
-        )
-    }
-
-    fn get_docids_for_value_range(
-        &self,
-        range: RangeInclusive<Output>,
-        doc_id_range: Range<u32>,
-        positions: &mut Vec<u32>,
-    ) {
-        self.from_column.get_docids_for_value_range(
-            self.monotonic_mapping.inverse(range.start().clone())
-                ..=self.monotonic_mapping.inverse(range.end().clone()),
-            doc_id_range,
-            positions,
-        )
-    }
-
-    // We voluntarily do not implement get_range as it yields a regression,
-    // and we do not have any specialized implementation anyway.
-}
-
-/// Wraps an iterator into a `Column`.
-pub struct IterColumn<T>(T);
-
-impl<T> From<T> for IterColumn<T>
-where T: Iterator + Clone + ExactSizeIterator
-{
-    fn from(iter: T) -> Self {
-        IterColumn(iter)
-    }
-}
-
-impl<T> Column<T::Item> for IterColumn<T>
-where
-    T: Iterator + Clone + ExactSizeIterator + Send + Sync,
-    T::Item: PartialOrd,
-{
-    fn get_val(&self, idx: u32) -> T::Item {
-        self.0.clone().nth(idx as usize).unwrap()
-    }
-
-    fn min_value(&self) -> T::Item {
-        self.0.clone().next().unwrap()
-    }
-
-    fn max_value(&self) -> T::Item {
-        self.0.clone().last().unwrap()
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.0.len() as u32
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = T::Item> + '_> {
-        Box::new(self.0.clone())
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use crate::monotonic_mapping::{
-        StrictlyMonotonicMappingInverter, StrictlyMonotonicMappingToInternalBaseval,
-        StrictlyMonotonicMappingToInternalGCDBaseval,
-    };
-
-    #[test]
-    fn test_monotonic_mapping() {
-        let vals = &[3u64, 5u64][..];
-        let col = VecColumn::from(vals);
-        let mapped = monotonic_map_column(col, StrictlyMonotonicMappingToInternalBaseval::new(2));
-        assert_eq!(mapped.min_value(), 1u64);
-        assert_eq!(mapped.max_value(), 3u64);
-        assert_eq!(mapped.num_vals(), 2);
-        assert_eq!(mapped.num_vals(), 2);
-        assert_eq!(mapped.get_val(0), 1);
-        assert_eq!(mapped.get_val(1), 3);
-    }
-
-    #[test]
-    fn test_range_as_col() {
-        let col = IterColumn::from(10..100);
-        assert_eq!(col.num_vals(), 90);
-        assert_eq!(col.max_value(), 99);
-    }
-
-    #[test]
-    fn test_monotonic_mapping_iter() {
-        let vals: Vec<u64> = (10..110u64).map(|el| el * 10).collect();
-        let col = VecColumn::from(&vals);
-        let mapped = monotonic_map_column(
-            col,
-            StrictlyMonotonicMappingInverter::from(
-                StrictlyMonotonicMappingToInternalGCDBaseval::new(10, 100),
-            ),
-        );
-        let val_i64s: Vec<u64> = mapped.iter().collect();
-        for i in 0..100 {
-            assert_eq!(val_i64s[i as usize], mapped.get_val(i));
-        }
-    }
-
-    #[test]
-    fn test_monotonic_mapping_get_range() {
-        let vals: Vec<u64> = (0..100u64).map(|el| el * 10).collect();
-        let col = VecColumn::from(&vals);
-        let mapped = monotonic_map_column(
-            col,
-            StrictlyMonotonicMappingInverter::from(
-                StrictlyMonotonicMappingToInternalGCDBaseval::new(10, 0),
-            ),
-        );
-
-        assert_eq!(mapped.min_value(), 0u64);
-        assert_eq!(mapped.max_value(), 9900u64);
-        assert_eq!(mapped.num_vals(), 100);
-        let val_u64s: Vec<u64> = mapped.iter().collect();
-        assert_eq!(val_u64s.len(), 100);
-        for i in 0..100 {
-            assert_eq!(val_u64s[i as usize], mapped.get_val(i));
-            assert_eq!(val_u64s[i as usize], vals[i as usize] * 10);
-        }
-        let mut buf = [0u64; 20];
-        mapped.get_range(7, &mut buf[..]);
-        assert_eq!(&val_u64s[7..][..20], &buf);
-    }
-}

fastfield_codecs/src/compact_space/blank_range.rs

@@ -1,43 +0,0 @@
-use std::ops::RangeInclusive;
-
-/// The range of a blank in value space.
-///
-/// A blank is an unoccupied space in the data.
-/// Use try_into() to construct.
-/// A range has to have at least length of 3. Invalid ranges will be rejected.
-///
-/// Ordered by range length.
-#[derive(Debug, Eq, PartialEq, Clone)]
-pub(crate) struct BlankRange {
-    blank_range: RangeInclusive<u128>,
-}
-impl TryFrom<RangeInclusive<u128>> for BlankRange {
-    type Error = &'static str;
-    fn try_from(range: RangeInclusive<u128>) -> Result<Self, Self::Error> {
-        let blank_size = range.end().saturating_sub(*range.start());
-        if blank_size < 2 {
-            Err("invalid range")
-        } else {
-            Ok(BlankRange { blank_range: range })
-        }
-    }
-}
-impl BlankRange {
-    pub(crate) fn blank_size(&self) -> u128 {
-        self.blank_range.end() - self.blank_range.start() + 1
-    }
-    pub(crate) fn blank_range(&self) -> RangeInclusive<u128> {
-        self.blank_range.clone()
-    }
-}
-
-impl Ord for BlankRange {
-    fn cmp(&self, other: &Self) -> std::cmp::Ordering {
-        self.blank_size().cmp(&other.blank_size())
-    }
-}
-impl PartialOrd for BlankRange {
-    fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
-        Some(self.blank_size().cmp(&other.blank_size()))
-    }
-}

fastfield_codecs/src/compact_space/build_compact_space.rs

@@ -1,231 +0,0 @@
-use std::collections::{BTreeSet, BinaryHeap};
-use std::iter;
-use std::ops::RangeInclusive;
-
-use itertools::Itertools;
-
-use super::blank_range::BlankRange;
-use super::{CompactSpace, RangeMapping};
-
-/// Put the blanks for the sorted values into a binary heap
-fn get_blanks(values_sorted: &BTreeSet<u128>) -> BinaryHeap<BlankRange> {
-    let mut blanks: BinaryHeap<BlankRange> = BinaryHeap::new();
-    for (first, second) in values_sorted.iter().tuple_windows() {
-        // Correctness Overflow: the values are deduped and sorted (BTreeSet property), that means
-        // there's always space between two values.
-        let blank_range = first + 1..=second - 1;
-        let blank_range: Result<BlankRange, _> = blank_range.try_into();
-        if let Ok(blank_range) = blank_range {
-            blanks.push(blank_range);
-        }
-    }
-
-    blanks
-}
-
-struct BlankCollector {
-    blanks: Vec<BlankRange>,
-    staged_blanks_sum: u128,
-}
-impl BlankCollector {
-    fn new() -> Self {
-        Self {
-            blanks: vec![],
-            staged_blanks_sum: 0,
-        }
-    }
-    fn stage_blank(&mut self, blank: BlankRange) {
-        self.staged_blanks_sum += blank.blank_size();
-        self.blanks.push(blank);
-    }
-    fn drain(&mut self) -> impl Iterator<Item = BlankRange> + '_ {
-        self.staged_blanks_sum = 0;
-        self.blanks.drain(..)
-    }
-    fn staged_blanks_sum(&self) -> u128 {
-        self.staged_blanks_sum
-    }
-    fn num_staged_blanks(&self) -> usize {
-        self.blanks.len()
-    }
-}
-fn num_bits(val: u128) -> u8 {
-    (128u32 - val.leading_zeros()) as u8
-}
-
-/// Will collect blanks and add them to compact space if more bits are saved than cost from
-/// metadata.
-pub fn get_compact_space(
-    values_deduped_sorted: &BTreeSet<u128>,
-    total_num_values: u32,
-    cost_per_blank: usize,
-) -> CompactSpace {
-    let mut compact_space_builder = CompactSpaceBuilder::new();
-    if values_deduped_sorted.is_empty() {
-        return compact_space_builder.finish();
-    }
-
-    let mut blanks: BinaryHeap<BlankRange> = get_blanks(values_deduped_sorted);
-    // Replace after stabilization of https://github.com/rust-lang/rust/issues/62924
-
-    // We start by space that's limited to min_value..=max_value
-    let min_value = *values_deduped_sorted.iter().next().unwrap_or(&0);
-    let max_value = *values_deduped_sorted.iter().last().unwrap_or(&0);
-
-    // +1 for null, in case min and max covers the whole space, we are off by one.
-    let mut amplitude_compact_space = (max_value - min_value).saturating_add(1);
-    if min_value != 0 {
-        compact_space_builder.add_blanks(iter::once(0..=min_value - 1));
-    }
-    if max_value != u128::MAX {
-        compact_space_builder.add_blanks(iter::once(max_value + 1..=u128::MAX));
-    }
-
-    let mut amplitude_bits: u8 = num_bits(amplitude_compact_space);
-
-    let mut blank_collector = BlankCollector::new();
-    // We will stage blanks until they reduce the compact space by at least 1 bit and then flush
-    // them if the metadata cost is lower than the total number of saved bits.
-    // Binary heap to process the gaps by their size
-    while let Some(blank_range) = blanks.pop() {
-        blank_collector.stage_blank(blank_range);
-
-        let staged_spaces_sum: u128 = blank_collector.staged_blanks_sum();
-        let amplitude_new_compact_space = amplitude_compact_space - staged_spaces_sum;
-        let amplitude_new_bits = num_bits(amplitude_new_compact_space);
-        if amplitude_bits == amplitude_new_bits {
-            continue;
-        }
-        let saved_bits = (amplitude_bits - amplitude_new_bits) as usize * total_num_values as usize;
-        // TODO: Maybe calculate exact cost of blanks and run this more expensive computation only,
-        // when amplitude_new_bits changes
-        let cost = blank_collector.num_staged_blanks() * cost_per_blank;
-        if cost >= saved_bits {
-            // Continue here, since although we walk over the blanks by size,
-            // we can potentially save a lot at the last bits, which are smaller blanks
-            //
-            // E.g. if the first range reduces the compact space by 1000 from 2000 to 1000, which
-            // saves 11-10=1 bit and the next range reduces the compact space by 950 to
-            // 50, which saves 10-6=4 bit
-            continue;
-        }
-
-        amplitude_compact_space = amplitude_new_compact_space;
-        amplitude_bits = amplitude_new_bits;
-        compact_space_builder.add_blanks(blank_collector.drain().map(|blank| blank.blank_range()));
-    }
-
-    // special case, when we don't collected any blanks because:
-    // * the data is empty (early exit)
-    // * the algorithm did decide it's not worth the cost, which can be the case for single values
-    //
-    // We drain one collected blank unconditionally, so the empty case is reserved for empty
-    // data, and therefore empty compact_space means the data is empty and no data is covered
-    // (conversely to all data) and we can assign null to it.
-    if compact_space_builder.is_empty() {
-        compact_space_builder.add_blanks(
-            blank_collector
-                .drain()
-                .map(|blank| blank.blank_range())
-                .take(1),
-        );
-    }
-
-    let compact_space = compact_space_builder.finish();
-    if max_value - min_value != u128::MAX {
-        debug_assert_eq!(
-            compact_space.amplitude_compact_space(),
-            amplitude_compact_space
-        );
-    }
-    compact_space
-}
-
-#[derive(Debug, Clone, Eq, PartialEq)]
-struct CompactSpaceBuilder {
-    blanks: Vec<RangeInclusive<u128>>,
-}
-
-impl CompactSpaceBuilder {
-    /// Creates a new compact space builder which will initially cover the whole space.
-    fn new() -> Self {
-        Self { blanks: Vec::new() }
-    }
-
-    /// Assumes that repeated add_blank calls don't overlap and are not adjacent,
-    /// e.g. [3..=5, 5..=10] is not allowed
-    ///
-    /// Both of those assumptions are true when blanks are produced from sorted values.
-    fn add_blanks(&mut self, blank: impl Iterator<Item = RangeInclusive<u128>>) {
-        self.blanks.extend(blank);
-    }
-
-    fn is_empty(&self) -> bool {
-        self.blanks.is_empty()
-    }
-
-    /// Convert blanks to covered space and assign null value
-    fn finish(mut self) -> CompactSpace {
-        // sort by start. ranges are not allowed to overlap
-        self.blanks.sort_unstable_by_key(|blank| *blank.start());
-
-        let mut covered_space = Vec::with_capacity(self.blanks.len());
-
-        // begining of the blanks
-        if let Some(first_blank_start) = self.blanks.first().map(RangeInclusive::start) {
-            if *first_blank_start != 0 {
-                covered_space.push(0..=first_blank_start - 1);
-            }
-        }
-
-        // Between the blanks
-        let between_blanks = self.blanks.iter().tuple_windows().map(|(left, right)| {
-            assert!(
-                left.end() < right.start(),
-                "overlapping or adjacent ranges detected"
-            );
-            *left.end() + 1..=*right.start() - 1
-        });
-        covered_space.extend(between_blanks);
-
-        // end of the blanks
-        if let Some(last_blank_end) = self.blanks.last().map(RangeInclusive::end) {
-            if *last_blank_end != u128::MAX {
-                covered_space.push(last_blank_end + 1..=u128::MAX);
-            }
-        }
-
-        if covered_space.is_empty() {
-            covered_space.push(0..=0); // empty data case
-        };
-
-        let mut compact_start: u64 = 1; // 0 is reserved for `null`
-        let mut ranges_mapping: Vec<RangeMapping> = Vec::with_capacity(covered_space.len());
-        for cov in covered_space {
-            let range_mapping = super::RangeMapping {
-                value_range: cov,
-                compact_start,
-            };
-            let covered_range_len = range_mapping.range_length();
-            ranges_mapping.push(range_mapping);
-            compact_start += covered_range_len;
-        }
-        // println!("num ranges {}", ranges_mapping.len());
-        CompactSpace { ranges_mapping }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_binary_heap_pop_order() {
-        let mut blanks: BinaryHeap<BlankRange> = BinaryHeap::new();
-        blanks.push((0..=10).try_into().unwrap());
-        blanks.push((100..=200).try_into().unwrap());
-        blanks.push((100..=110).try_into().unwrap());
-        assert_eq!(blanks.pop().unwrap().blank_size(), 101);
-        assert_eq!(blanks.pop().unwrap().blank_size(), 11);
-    }
-}

fastfield_codecs/src/compact_space/mod.rs

@@ -1,814 +0,0 @@
-/// This codec takes a large number space (u128) and reduces it to a compact number space.
-///
-/// It will find spaces in the number range. For example:
-///
-/// 100, 101, 102, 103, 104, 50000, 50001
-/// could be mapped to
-/// 100..104 -> 0..4
-/// 50000..50001 -> 5..6
-///
-/// Compact space 0..=6 requires much less bits than 100..=50001
-///
-/// The codec is created to compress ip addresses, but may be employed in other use cases.
-use std::{
-    cmp::Ordering,
-    collections::BTreeSet,
-    io::{self, Write},
-    ops::{Range, RangeInclusive},
-};
-
-use common::{BinarySerializable, CountingWriter, VInt, VIntU128};
-use ownedbytes::OwnedBytes;
-use tantivy_bitpacker::{self, BitPacker, BitUnpacker};
-
-use crate::compact_space::build_compact_space::get_compact_space;
-use crate::Column;
-
-mod blank_range;
-mod build_compact_space;
-
-/// The cost per blank is quite hard actually, since blanks are delta encoded, the actual cost of
-/// blanks depends on the number of blanks.
-///
-/// The number is taken by looking at a real dataset. It is optimized for larger datasets.
-const COST_PER_BLANK_IN_BITS: usize = 36;
-
-#[derive(Debug, Clone, Eq, PartialEq)]
-pub struct CompactSpace {
-    ranges_mapping: Vec<RangeMapping>,
-}
-
-/// Maps the range from the original space to compact_start + range.len()
-#[derive(Debug, Clone, Eq, PartialEq)]
-struct RangeMapping {
-    value_range: RangeInclusive<u128>,
-    compact_start: u64,
-}
-impl RangeMapping {
-    fn range_length(&self) -> u64 {
-        (self.value_range.end() - self.value_range.start()) as u64 + 1
-    }
-
-    // The last value of the compact space in this range
-    fn compact_end(&self) -> u64 {
-        self.compact_start + self.range_length() - 1
-    }
-}
-
-impl BinarySerializable for CompactSpace {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        VInt(self.ranges_mapping.len() as u64).serialize(writer)?;
-
-        let mut prev_value = 0;
-        for value_range in self
-            .ranges_mapping
-            .iter()
-            .map(|range_mapping| &range_mapping.value_range)
-        {
-            let blank_delta_start = value_range.start() - prev_value;
-            VIntU128(blank_delta_start).serialize(writer)?;
-            prev_value = *value_range.start();
-
-            let blank_delta_end = value_range.end() - prev_value;
-            VIntU128(blank_delta_end).serialize(writer)?;
-            prev_value = *value_range.end();
-        }
-
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let num_ranges = VInt::deserialize(reader)?.0;
-        let mut ranges_mapping: Vec<RangeMapping> = vec![];
-        let mut value = 0u128;
-        let mut compact_start = 1u64; // 0 is reserved for `null`
-        for _ in 0..num_ranges {
-            let blank_delta_start = VIntU128::deserialize(reader)?.0;
-            value += blank_delta_start;
-            let blank_start = value;
-
-            let blank_delta_end = VIntU128::deserialize(reader)?.0;
-            value += blank_delta_end;
-            let blank_end = value;
-
-            let range_mapping = RangeMapping {
-                value_range: blank_start..=blank_end,
-                compact_start,
-            };
-            let range_length = range_mapping.range_length();
-            ranges_mapping.push(range_mapping);
-            compact_start += range_length;
-        }
-
-        Ok(Self { ranges_mapping })
-    }
-}
-
-impl CompactSpace {
-    /// Amplitude is the value range of the compact space including the sentinel value used to
-    /// identify null values. The compact space is 0..=amplitude .
-    ///
-    /// It's only used to verify we don't exceed u64 number space, which would indicate a bug.
-    fn amplitude_compact_space(&self) -> u128 {
-        self.ranges_mapping
-            .last()
-            .map(|last_range| last_range.compact_end() as u128)
-            .unwrap_or(1) // compact space starts at 1, 0 == null
-    }
-
-    fn get_range_mapping(&self, pos: usize) -> &RangeMapping {
-        &self.ranges_mapping[pos]
-    }
-
-    /// Returns either Ok(the value in the compact space) or if it is outside the compact space the
-    /// Err(position where it would be inserted)
-    fn u128_to_compact(&self, value: u128) -> Result<u64, usize> {
-        self.ranges_mapping
-            .binary_search_by(|probe| {
-                let value_range = &probe.value_range;
-                if value < *value_range.start() {
-                    Ordering::Greater
-                } else if value > *value_range.end() {
-                    Ordering::Less
-                } else {
-                    Ordering::Equal
-                }
-            })
-            .map(|pos| {
-                let range_mapping = &self.ranges_mapping[pos];
-                let pos_in_range = (value - range_mapping.value_range.start()) as u64;
-                range_mapping.compact_start + pos_in_range
-            })
-    }
-
-    /// Unpacks a value from compact space u64 to u128 space
-    fn compact_to_u128(&self, compact: u64) -> u128 {
-        let pos = self
-            .ranges_mapping
-            .binary_search_by_key(&compact, |range_mapping| range_mapping.compact_start)
-            // Correctness: Overflow. The first range starts at compact space 0, the error from
-            // binary search can never be 0
-            .map_or_else(|e| e - 1, |v| v);
-
-        let range_mapping = &self.ranges_mapping[pos];
-        let diff = compact - range_mapping.compact_start;
-        range_mapping.value_range.start() + diff as u128
-    }
-}
-
-pub struct CompactSpaceCompressor {
-    params: IPCodecParams,
-}
-#[derive(Debug, Clone)]
-pub struct IPCodecParams {
-    compact_space: CompactSpace,
-    bit_unpacker: BitUnpacker,
-    min_value: u128,
-    max_value: u128,
-    num_vals: u32,
-    num_bits: u8,
-}
-
-impl CompactSpaceCompressor {
-    /// Taking the vals as Vec may cost a lot of memory. It is used to sort the vals.
-    pub fn train_from(iter: impl Iterator<Item = u128>, num_vals: u32) -> Self {
-        let mut values_sorted = BTreeSet::new();
-        values_sorted.extend(iter);
-        let total_num_values = num_vals;
-
-        let compact_space =
-            get_compact_space(&values_sorted, total_num_values, COST_PER_BLANK_IN_BITS);
-        let amplitude_compact_space = compact_space.amplitude_compact_space();
-
-        assert!(
-            amplitude_compact_space <= u64::MAX as u128,
-            "case unsupported."
-        );
-
-        let num_bits = tantivy_bitpacker::compute_num_bits(amplitude_compact_space as u64);
-        let min_value = *values_sorted.iter().next().unwrap_or(&0);
-        let max_value = *values_sorted.iter().last().unwrap_or(&0);
-        assert_eq!(
-            compact_space
-                .u128_to_compact(max_value)
-                .expect("could not convert max value to compact space"),
-            amplitude_compact_space as u64
-        );
-        CompactSpaceCompressor {
-            params: IPCodecParams {
-                compact_space,
-                bit_unpacker: BitUnpacker::new(num_bits),
-                min_value,
-                max_value,
-                num_vals: total_num_values,
-                num_bits,
-            },
-        }
-    }
-
-    fn write_footer(self, writer: &mut impl Write) -> io::Result<()> {
-        let writer = &mut CountingWriter::wrap(writer);
-        self.params.serialize(writer)?;
-
-        let footer_len = writer.written_bytes() as u32;
-        footer_len.serialize(writer)?;
-
-        Ok(())
-    }
-
-    pub fn compress_into(
-        self,
-        vals: impl Iterator<Item = u128>,
-        write: &mut impl Write,
-    ) -> io::Result<()> {
-        let mut bitpacker = BitPacker::default();
-        for val in vals {
-            let compact = self
-                .params
-                .compact_space
-                .u128_to_compact(val)
-                .map_err(|_| {
-                    io::Error::new(
-                        io::ErrorKind::InvalidData,
-                        "Could not convert value to compact_space. This is a bug.",
-                    )
-                })?;
-            bitpacker.write(compact, self.params.num_bits, write)?;
-        }
-        bitpacker.close(write)?;
-        self.write_footer(write)?;
-        Ok(())
-    }
-}
-
-#[derive(Debug, Clone)]
-pub struct CompactSpaceDecompressor {
-    data: OwnedBytes,
-    params: IPCodecParams,
-}
-
-impl BinarySerializable for IPCodecParams {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        // header flags for future optional dictionary encoding
-        let footer_flags = 0u64;
-        footer_flags.serialize(writer)?;
-
-        VIntU128(self.min_value).serialize(writer)?;
-        VIntU128(self.max_value).serialize(writer)?;
-        VIntU128(self.num_vals as u128).serialize(writer)?;
-        self.num_bits.serialize(writer)?;
-
-        self.compact_space.serialize(writer)?;
-
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let _header_flags = u64::deserialize(reader)?;
-        let min_value = VIntU128::deserialize(reader)?.0;
-        let max_value = VIntU128::deserialize(reader)?.0;
-        let num_vals = VIntU128::deserialize(reader)?.0 as u32;
-        let num_bits = u8::deserialize(reader)?;
-        let compact_space = CompactSpace::deserialize(reader)?;
-
-        Ok(Self {
-            compact_space,
-            bit_unpacker: BitUnpacker::new(num_bits),
-            min_value,
-            max_value,
-            num_vals,
-            num_bits,
-        })
-    }
-}
-
-impl Column<u128> for CompactSpaceDecompressor {
-    #[inline]
-    fn get_val(&self, doc: u32) -> u128 {
-        self.get(doc)
-    }
-
-    fn min_value(&self) -> u128 {
-        self.min_value()
-    }
-
-    fn max_value(&self) -> u128 {
-        self.max_value()
-    }
-
-    fn num_vals(&self) -> u32 {
-        self.params.num_vals
-    }
-
-    #[inline]
-    fn iter(&self) -> Box<dyn Iterator<Item = u128> + '_> {
-        Box::new(self.iter())
-    }
-
-    #[inline]
-    fn get_docids_for_value_range(
-        &self,
-        value_range: RangeInclusive<u128>,
-        positions_range: Range<u32>,
-        positions: &mut Vec<u32>,
-    ) {
-        self.get_positions_for_value_range(value_range, positions_range, positions)
-    }
-}
-
-impl CompactSpaceDecompressor {
-    pub fn open(data: OwnedBytes) -> io::Result<CompactSpaceDecompressor> {
-        let (data_slice, footer_len_bytes) = data.split_at(data.len() - 4);
-        let footer_len = u32::deserialize(&mut &footer_len_bytes[..])?;
-
-        let data_footer = &data_slice[data_slice.len() - footer_len as usize..];
-        let params = IPCodecParams::deserialize(&mut &data_footer[..])?;
-        let decompressor = CompactSpaceDecompressor { data, params };
-
-        Ok(decompressor)
-    }
-
-    /// Converting to compact space for the decompressor is more complex, since we may get values
-    /// which are outside the compact space. e.g. if we map
-    /// 1000 => 5
-    /// 2000 => 6
-    ///
-    /// and we want a mapping for 1005, there is no equivalent compact space. We instead return an
-    /// error with the index of the next range.
-    fn u128_to_compact(&self, value: u128) -> Result<u64, usize> {
-        self.params.compact_space.u128_to_compact(value)
-    }
-
-    fn compact_to_u128(&self, compact: u64) -> u128 {
-        self.params.compact_space.compact_to_u128(compact)
-    }
-
-    /// Comparing on compact space: Random dataset 0,24 (50% random hit) - 1.05 GElements/s
-    /// Comparing on compact space: Real dataset 1.08 GElements/s
-    ///
-    /// Comparing on original space: Real dataset .06 GElements/s (not completely optimized)
-    #[inline]
-    pub fn get_positions_for_value_range(
-        &self,
-        value_range: RangeInclusive<u128>,
-        position_range: Range<u32>,
-        positions: &mut Vec<u32>,
-    ) {
-        if value_range.start() > value_range.end() {
-            return;
-        }
-        let position_range = position_range.start..position_range.end.min(self.num_vals());
-        let from_value = *value_range.start();
-        let to_value = *value_range.end();
-        assert!(to_value >= from_value);
-        let compact_from = self.u128_to_compact(from_value);
-        let compact_to = self.u128_to_compact(to_value);
-
-        // Quick return, if both ranges fall into the same non-mapped space, the range can't cover
-        // any values, so we can early exit
-        match (compact_to, compact_from) {
-            (Err(pos1), Err(pos2)) if pos1 == pos2 => return,
-            _ => {}
-        }
-
-        let compact_from = compact_from.unwrap_or_else(|pos| {
-            // Correctness: Out of bounds, if this value is Err(last_index + 1), we early exit,
-            // since the to_value also mapps into the same non-mapped space
-            let range_mapping = self.params.compact_space.get_range_mapping(pos);
-            range_mapping.compact_start
-        });
-        // If there is no compact space, we go to the closest upperbound compact space
-        let compact_to = compact_to.unwrap_or_else(|pos| {
-            // Correctness: Overflow, if this value is Err(0), we early exit,
-            // since the from_value also mapps into the same non-mapped space
-
-            // Get end of previous range
-            let pos = pos - 1;
-            let range_mapping = self.params.compact_space.get_range_mapping(pos);
-            range_mapping.compact_end()
-        });
-
-        let range = compact_from..=compact_to;
-
-        let scan_num_docs = position_range.end - position_range.start;
-
-        let step_size = 4;
-        let cutoff = position_range.start + scan_num_docs - scan_num_docs % step_size;
-
-        let mut push_if_in_range = |idx, val| {
-            if range.contains(&val) {
-                positions.push(idx);
-            }
-        };
-        let get_val = |idx| self.params.bit_unpacker.get(idx, &self.data);
-        // unrolled loop
-        for idx in (position_range.start..cutoff).step_by(step_size as usize) {
-            let idx1 = idx;
-            let idx2 = idx + 1;
-            let idx3 = idx + 2;
-            let idx4 = idx + 3;
-            let val1 = get_val(idx1);
-            let val2 = get_val(idx2);
-            let val3 = get_val(idx3);
-            let val4 = get_val(idx4);
-            push_if_in_range(idx1, val1);
-            push_if_in_range(idx2, val2);
-            push_if_in_range(idx3, val3);
-            push_if_in_range(idx4, val4);
-        }
-
-        // handle rest
-        for idx in cutoff..position_range.end {
-            push_if_in_range(idx, get_val(idx));
-        }
-    }
-
-    #[inline]
-    fn iter_compact(&self) -> impl Iterator<Item = u64> + '_ {
-        (0..self.params.num_vals).map(move |idx| self.params.bit_unpacker.get(idx, &self.data))
-    }
-
-    #[inline]
-    fn iter(&self) -> impl Iterator<Item = u128> + '_ {
-        // TODO: Performance. It would be better to iterate on the ranges and check existence via
-        // the bit_unpacker.
-        self.iter_compact()
-            .map(|compact| self.compact_to_u128(compact))
-    }
-
-    #[inline]
-    pub fn get(&self, idx: u32) -> u128 {
-        let compact = self.params.bit_unpacker.get(idx, &self.data);
-        self.compact_to_u128(compact)
-    }
-
-    pub fn min_value(&self) -> u128 {
-        self.params.min_value
-    }
-
-    pub fn max_value(&self) -> u128 {
-        self.params.max_value
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use super::*;
-    use crate::format_version::read_format_version;
-    use crate::null_index_footer::read_null_index_footer;
-    use crate::serialize::U128Header;
-    use crate::{open_u128, serialize_u128};
-
-    #[test]
-    fn compact_space_test() {
-        let ips = &[
-            2u128, 4u128, 1000, 1001, 1002, 1003, 1004, 1005, 1008, 1010, 1012, 1260,
-        ]
-        .into_iter()
-        .collect();
-        let compact_space = get_compact_space(ips, ips.len() as u32, 11);
-        let amplitude = compact_space.amplitude_compact_space();
-        assert_eq!(amplitude, 17);
-        assert_eq!(1, compact_space.u128_to_compact(2).unwrap());
-        assert_eq!(2, compact_space.u128_to_compact(3).unwrap());
-        assert_eq!(compact_space.u128_to_compact(100).unwrap_err(), 1);
-
-        for (num1, num2) in (0..3).tuple_windows() {
-            assert_eq!(
-                compact_space.get_range_mapping(num1).compact_end() + 1,
-                compact_space.get_range_mapping(num2).compact_start
-            );
-        }
-
-        let mut output: Vec<u8> = Vec::new();
-        compact_space.serialize(&mut output).unwrap();
-
-        assert_eq!(
-            compact_space,
-            CompactSpace::deserialize(&mut &output[..]).unwrap()
-        );
-
-        for ip in ips {
-            let compact = compact_space.u128_to_compact(*ip).unwrap();
-            assert_eq!(compact_space.compact_to_u128(compact), *ip);
-        }
-    }
-
-    #[test]
-    fn compact_space_amplitude_test() {
-        let ips = &[100000u128, 1000000].into_iter().collect();
-        let compact_space = get_compact_space(ips, ips.len() as u32, 1);
-        let amplitude = compact_space.amplitude_compact_space();
-        assert_eq!(amplitude, 2);
-    }
-
-    fn test_all(mut data: OwnedBytes, expected: &[u128]) {
-        let _header = U128Header::deserialize(&mut data);
-        let decompressor = CompactSpaceDecompressor::open(data).unwrap();
-        for (idx, expected_val) in expected.iter().cloned().enumerate() {
-            let val = decompressor.get(idx as u32);
-            assert_eq!(val, expected_val);
-
-            let test_range = |range: RangeInclusive<u128>| {
-                let expected_positions = expected
-                    .iter()
-                    .positions(|val| range.contains(val))
-                    .map(|pos| pos as u32)
-                    .collect::<Vec<_>>();
-                let mut positions = Vec::new();
-                decompressor.get_positions_for_value_range(
-                    range,
-                    0..decompressor.num_vals(),
-                    &mut positions,
-                );
-                assert_eq!(positions, expected_positions);
-            };
-
-            test_range(expected_val.saturating_sub(1)..=expected_val);
-            test_range(expected_val..=expected_val);
-            test_range(expected_val..=expected_val.saturating_add(1));
-            test_range(expected_val.saturating_sub(1)..=expected_val.saturating_add(1));
-        }
-    }
-
-    fn test_aux_vals(u128_vals: &[u128]) -> OwnedBytes {
-        let mut out = Vec::new();
-        serialize_u128(
-            || u128_vals.iter().cloned(),
-            u128_vals.len() as u32,
-            &mut out,
-        )
-        .unwrap();
-
-        let data = OwnedBytes::new(out);
-        let (data, _format_version) = read_format_version(data).unwrap();
-        let (data, _null_index_footer) = read_null_index_footer(data).unwrap();
-        test_all(data.clone(), u128_vals);
-
-        data
-    }
-
-    #[test]
-    fn test_range_1() {
-        let vals = &[
-            1u128,
-            100u128,
-            3u128,
-            99999u128,
-            100000u128,
-            100001u128,
-            4_000_211_221u128,
-            4_000_211_222u128,
-            333u128,
-        ];
-        let mut data = test_aux_vals(vals);
-
-        let _header = U128Header::deserialize(&mut data);
-        let decomp = CompactSpaceDecompressor::open(data).unwrap();
-        let complete_range = 0..vals.len() as u32;
-        for (pos, val) in vals.iter().enumerate() {
-            let val = *val;
-            let pos = pos as u32;
-            let mut positions = Vec::new();
-            decomp.get_positions_for_value_range(val..=val, pos..pos + 1, &mut positions);
-            assert_eq!(positions, vec![pos]);
-        }
-
-        // handle docid range out of bounds
-        let positions = get_positions_for_value_range_helper(&decomp, 0..=1, 1..u32::MAX);
-        assert_eq!(positions, vec![]);
-
-        let positions =
-            get_positions_for_value_range_helper(&decomp, 0..=1, complete_range.clone());
-        assert_eq!(positions, vec![0]);
-        let positions =
-            get_positions_for_value_range_helper(&decomp, 0..=2, complete_range.clone());
-        assert_eq!(positions, vec![0]);
-        let positions =
-            get_positions_for_value_range_helper(&decomp, 0..=3, complete_range.clone());
-        assert_eq!(positions, vec![0, 2]);
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                99999u128..=99999u128,
-                complete_range.clone()
-            ),
-            vec![3]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                99999u128..=100000u128,
-                complete_range.clone()
-            ),
-            vec![3, 4]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                99998u128..=100000u128,
-                complete_range.clone()
-            ),
-            vec![3, 4]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                99998u128..=99999u128,
-                complete_range.clone()
-            ),
-            vec![3]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                99998u128..=99998u128,
-                complete_range.clone()
-            ),
-            vec![]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                333u128..=333u128,
-                complete_range.clone()
-            ),
-            vec![8]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                332u128..=333u128,
-                complete_range.clone()
-            ),
-            vec![8]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                332u128..=334u128,
-                complete_range.clone()
-            ),
-            vec![8]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                333u128..=334u128,
-                complete_range.clone()
-            ),
-            vec![8]
-        );
-
-        assert_eq!(
-            get_positions_for_value_range_helper(
-                &decomp,
-                4_000_211_221u128..=5_000_000_000u128,
-                complete_range
-            ),
-            vec![6, 7]
-        );
-    }
-
-    #[test]
-    fn test_empty() {
-        let vals = &[];
-        let data = test_aux_vals(vals);
-        let _decomp = CompactSpaceDecompressor::open(data).unwrap();
-    }
-
-    #[test]
-    fn test_range_2() {
-        let vals = &[
-            100u128,
-            99999u128,
-            100000u128,
-            100001u128,
-            4_000_211_221u128,
-            4_000_211_222u128,
-            333u128,
-        ];
-        let mut data = test_aux_vals(vals);
-        let _header = U128Header::deserialize(&mut data);
-        let decomp = CompactSpaceDecompressor::open(data).unwrap();
-        let complete_range = 0..vals.len() as u32;
-        assert_eq!(
-            get_positions_for_value_range_helper(&decomp, 0..=5, complete_range.clone()),
-            vec![]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(&decomp, 0..=100, complete_range.clone()),
-            vec![0]
-        );
-        assert_eq!(
-            get_positions_for_value_range_helper(&decomp, 0..=105, complete_range),
-            vec![0]
-        );
-    }
-
-    fn get_positions_for_value_range_helper<C: Column<T> + ?Sized, T: PartialOrd>(
-        column: &C,
-        value_range: RangeInclusive<T>,
-        doc_id_range: Range<u32>,
-    ) -> Vec<u32> {
-        let mut positions = Vec::new();
-        column.get_docids_for_value_range(value_range, doc_id_range, &mut positions);
-        positions
-    }
-
-    #[test]
-    fn test_range_3() {
-        let vals = &[
-            200u128,
-            201,
-            202,
-            203,
-            204,
-            204,
-            206,
-            207,
-            208,
-            209,
-            210,
-            1_000_000,
-            5_000_000_000,
-        ];
-        let mut out = Vec::new();
-        serialize_u128(|| vals.iter().cloned(), vals.len() as u32, &mut out).unwrap();
-        let decomp = open_u128::<u128>(OwnedBytes::new(out)).unwrap();
-        let complete_range = 0..vals.len() as u32;
-
-        assert_eq!(
-            get_positions_for_value_range_helper(&*decomp, 199..=200, complete_range.clone()),
-            vec![0]
-        );
-
-        assert_eq!(
-            get_positions_for_value_range_helper(&*decomp, 199..=201, complete_range.clone()),
-            vec![0, 1]
-        );
-
-        assert_eq!(
-            get_positions_for_value_range_helper(&*decomp, 200..=200, complete_range.clone()),
-            vec![0]
-        );
-
-        assert_eq!(
-            get_positions_for_value_range_helper(&*decomp, 1_000_000..=1_000_000, complete_range),
-            vec![11]
-        );
-    }
-
-    #[test]
-    fn test_bug1() {
-        let vals = &[9223372036854775806];
-        let _data = test_aux_vals(vals);
-    }
-
-    #[test]
-    fn test_bug2() {
-        let vals = &[340282366920938463463374607431768211455u128];
-        let _data = test_aux_vals(vals);
-    }
-
-    #[test]
-    fn test_bug3() {
-        let vals = &[340282366920938463463374607431768211454];
-        let _data = test_aux_vals(vals);
-    }
-
-    #[test]
-    fn test_bug4() {
-        let vals = &[340282366920938463463374607431768211455, 0];
-        let _data = test_aux_vals(vals);
-    }
-
-    #[test]
-    fn test_first_large_gaps() {
-        let vals = &[1_000_000_000u128; 100];
-        let _data = test_aux_vals(vals);
-    }
-    use itertools::Itertools;
-    use proptest::prelude::*;
-
-    fn num_strategy() -> impl Strategy<Value = u128> {
-        prop_oneof![
-            1 => prop::num::u128::ANY.prop_map(|num| u128::MAX - (num % 10) ),
-            1 => prop::num::u128::ANY.prop_map(|num| i64::MAX as u128 + 5 - (num % 10) ),
-            1 => prop::num::u128::ANY.prop_map(|num| i128::MAX as u128 + 5 - (num % 10) ),
-            1 => prop::num::u128::ANY.prop_map(|num| num % 10 ),
-            20 => prop::num::u128::ANY,
-        ]
-    }
-
-    proptest! {
-        #![proptest_config(ProptestConfig::with_cases(10))]
-
-            #[test]
-            fn compress_decompress_random(vals in proptest::collection::vec(num_strategy()
-    , 1..1000)) {
-                let _data = test_aux_vals(&vals);
-            }
-        }
-}

fastfield_codecs/src/format_version.rs

@@ -1,39 +0,0 @@
-use std::io;
-
-use common::BinarySerializable;
-use ownedbytes::OwnedBytes;
-
-const MAGIC_NUMBER: u16 = 4335u16;
-const FASTFIELD_FORMAT_VERSION: u8 = 1;
-
-pub(crate) fn append_format_version(output: &mut impl io::Write) -> io::Result<()> {
-    FASTFIELD_FORMAT_VERSION.serialize(output)?;
-    MAGIC_NUMBER.serialize(output)?;
-
-    Ok(())
-}
-
-pub(crate) fn read_format_version(data: OwnedBytes) -> io::Result<(OwnedBytes, u8)> {
-    let (data, magic_number_bytes) = data.rsplit(2);
-
-    let magic_number = u16::deserialize(&mut magic_number_bytes.as_slice())?;
-    if magic_number != MAGIC_NUMBER {
-        return Err(io::Error::new(
-            io::ErrorKind::InvalidData,
-            format!("magic number mismatch {} != {}", magic_number, MAGIC_NUMBER),
-        ));
-    }
-    let (data, format_version_bytes) = data.rsplit(1);
-    let format_version = u8::deserialize(&mut format_version_bytes.as_slice())?;
-    if format_version > FASTFIELD_FORMAT_VERSION {
-        return Err(io::Error::new(
-            io::ErrorKind::InvalidData,
-            format!(
-                "Unsupported fastfield format version: {}. Max supported version: {}",
-                format_version, FASTFIELD_FORMAT_VERSION
-            ),
-        ));
-    }
-
-    Ok((data, format_version))
-}

fastfield_codecs/src/gcd.rs

@@ -1,170 +0,0 @@
-use std::num::NonZeroU64;
-
-use fastdivide::DividerU64;
-
-/// Compute the gcd of two non null numbers.
-///
-/// It is recommended, but not required, to feed values such that `large >= small`.
-fn compute_gcd(mut large: NonZeroU64, mut small: NonZeroU64) -> NonZeroU64 {
-    loop {
-        let rem: u64 = large.get() % small;
-        if let Some(new_small) = NonZeroU64::new(rem) {
-            (large, small) = (small, new_small);
-        } else {
-            return small;
-        }
-    }
-}
-
-// Find GCD for iterator of numbers
-pub fn find_gcd(numbers: impl Iterator<Item = u64>) -> Option<NonZeroU64> {
-    let mut numbers = numbers.flat_map(NonZeroU64::new);
-    let mut gcd: NonZeroU64 = numbers.next()?;
-    if gcd.get() == 1 {
-        return Some(gcd);
-    }
-
-    let mut gcd_divider = DividerU64::divide_by(gcd.get());
-    for val in numbers {
-        let remainder = val.get() - (gcd_divider.divide(val.get())) * gcd.get();
-        if remainder == 0 {
-            continue;
-        }
-        gcd = compute_gcd(val, gcd);
-        if gcd.get() == 1 {
-            return Some(gcd);
-        }
-
-        gcd_divider = DividerU64::divide_by(gcd.get());
-    }
-    Some(gcd)
-}
-
-#[cfg(test)]
-mod tests {
-    use std::io;
-    use std::num::NonZeroU64;
-
-    use ownedbytes::OwnedBytes;
-
-    use crate::gcd::{compute_gcd, find_gcd};
-    use crate::{FastFieldCodecType, VecColumn};
-
-    fn test_fastfield_gcd_i64_with_codec(
-        codec_type: FastFieldCodecType,
-        num_vals: usize,
-    ) -> io::Result<()> {
-        let mut vals: Vec<i64> = (-4..=(num_vals as i64) - 5).map(|val| val * 1000).collect();
-        let mut buffer: Vec<u8> = Vec::new();
-        crate::serialize(VecColumn::from(&vals), &mut buffer, &[codec_type])?;
-        let buffer = OwnedBytes::new(buffer);
-        let column = crate::open::<i64>(buffer.clone())?;
-        assert_eq!(column.get_val(0), -4000i64);
-        assert_eq!(column.get_val(1), -3000i64);
-        assert_eq!(column.get_val(2), -2000i64);
-        assert_eq!(column.max_value(), (num_vals as i64 - 5) * 1000);
-        assert_eq!(column.min_value(), -4000i64);
-
-        // Can't apply gcd
-        let mut buffer_without_gcd = Vec::new();
-        vals.pop();
-        vals.push(1001i64);
-        crate::serialize(
-            VecColumn::from(&vals),
-            &mut buffer_without_gcd,
-            &[codec_type],
-        )?;
-        let buffer_without_gcd = OwnedBytes::new(buffer_without_gcd);
-        assert!(buffer_without_gcd.len() > buffer.len());
-
-        Ok(())
-    }
-
-    #[test]
-    fn test_fastfield_gcd_i64() -> io::Result<()> {
-        for &codec_type in &[
-            FastFieldCodecType::Bitpacked,
-            FastFieldCodecType::BlockwiseLinear,
-            FastFieldCodecType::Linear,
-        ] {
-            test_fastfield_gcd_i64_with_codec(codec_type, 5500)?;
-        }
-        Ok(())
-    }
-
-    fn test_fastfield_gcd_u64_with_codec(
-        codec_type: FastFieldCodecType,
-        num_vals: usize,
-    ) -> io::Result<()> {
-        let mut vals: Vec<u64> = (1..=num_vals).map(|i| i as u64 * 1000u64).collect();
-        let mut buffer: Vec<u8> = Vec::new();
-        crate::serialize(VecColumn::from(&vals), &mut buffer, &[codec_type])?;
-        let buffer = OwnedBytes::new(buffer);
-        let column = crate::open::<u64>(buffer.clone())?;
-        assert_eq!(column.get_val(0), 1000u64);
-        assert_eq!(column.get_val(1), 2000u64);
-        assert_eq!(column.get_val(2), 3000u64);
-        assert_eq!(column.max_value(), num_vals as u64 * 1000);
-        assert_eq!(column.min_value(), 1000u64);
-
-        // Can't apply gcd
-        let mut buffer_without_gcd = Vec::new();
-        vals.pop();
-        vals.push(1001u64);
-        crate::serialize(
-            VecColumn::from(&vals),
-            &mut buffer_without_gcd,
-            &[codec_type],
-        )?;
-        let buffer_without_gcd = OwnedBytes::new(buffer_without_gcd);
-        assert!(buffer_without_gcd.len() > buffer.len());
-        Ok(())
-    }
-
-    #[test]
-    fn test_fastfield_gcd_u64() -> io::Result<()> {
-        for &codec_type in &[
-            FastFieldCodecType::Bitpacked,
-            FastFieldCodecType::BlockwiseLinear,
-            FastFieldCodecType::Linear,
-        ] {
-            test_fastfield_gcd_u64_with_codec(codec_type, 5500)?;
-        }
-        Ok(())
-    }
-
-    #[test]
-    pub fn test_fastfield2() {
-        let test_fastfield = crate::serialize_and_load(&[100u64, 200u64, 300u64]);
-        assert_eq!(test_fastfield.get_val(0), 100);
-        assert_eq!(test_fastfield.get_val(1), 200);
-        assert_eq!(test_fastfield.get_val(2), 300);
-    }
-
-    #[test]
-    fn test_compute_gcd() {
-        let test_compute_gcd_aux = |large, small, expected| {
-            let large = NonZeroU64::new(large).unwrap();
-            let small = NonZeroU64::new(small).unwrap();
-            let expected = NonZeroU64::new(expected).unwrap();
-            assert_eq!(compute_gcd(small, large), expected);
-            assert_eq!(compute_gcd(large, small), expected);
-        };
-        test_compute_gcd_aux(1, 4, 1);
-        test_compute_gcd_aux(2, 4, 2);
-        test_compute_gcd_aux(10, 25, 5);
-        test_compute_gcd_aux(25, 25, 25);
-    }
-
-    #[test]
-    fn find_gcd_test() {
-        assert_eq!(find_gcd([0].into_iter()), None);
-        assert_eq!(find_gcd([0, 10].into_iter()), NonZeroU64::new(10));
-        assert_eq!(find_gcd([10, 0].into_iter()), NonZeroU64::new(10));
-        assert_eq!(find_gcd([].into_iter()), None);
-        assert_eq!(find_gcd([15, 30, 5, 10].into_iter()), NonZeroU64::new(5));
-        assert_eq!(find_gcd([15, 16, 10].into_iter()), NonZeroU64::new(1));
-        assert_eq!(find_gcd([0, 5, 5, 5].into_iter()), NonZeroU64::new(5));
-        assert_eq!(find_gcd([0, 0].into_iter()), None);
-    }
-}

fastfield_codecs/src/lib.rs

@@ -1,563 +0,0 @@
-#![warn(missing_docs)]
-#![cfg_attr(all(feature = "unstable", test), feature(test))]
-
-//! # `fastfield_codecs`
-//!
-//! - Columnar storage of data for tantivy [`Column`].
-//! - Encode data in different codecs.
-//! - Monotonically map values to u64/u128
-
-#[cfg(test)]
-#[macro_use]
-extern crate more_asserts;
-
-#[cfg(all(test, feature = "unstable"))]
-extern crate test;
-
-use std::io;
-use std::io::Write;
-use std::sync::Arc;
-
-use common::BinarySerializable;
-use compact_space::CompactSpaceDecompressor;
-use format_version::read_format_version;
-use monotonic_mapping::{
-    StrictlyMonotonicMappingInverter, StrictlyMonotonicMappingToInternal,
-    StrictlyMonotonicMappingToInternalBaseval, StrictlyMonotonicMappingToInternalGCDBaseval,
-};
-use null_index_footer::read_null_index_footer;
-use ownedbytes::OwnedBytes;
-use serialize::{Header, U128Header};
-
-mod bitpacked;
-mod blockwise_linear;
-mod compact_space;
-mod format_version;
-mod line;
-mod linear;
-mod monotonic_mapping;
-mod monotonic_mapping_u128;
-#[allow(dead_code)]
-mod null_index;
-mod null_index_footer;
-
-mod column;
-mod gcd;
-pub mod serialize;
-
-pub use ordered_float;
-
-use self::bitpacked::BitpackedCodec;
-use self::blockwise_linear::BlockwiseLinearCodec;
-pub use self::column::{monotonic_map_column, Column, IterColumn, VecColumn};
-use self::linear::LinearCodec;
-pub use self::monotonic_mapping::{MonotonicallyMappableToU64, StrictlyMonotonicFn};
-pub use self::monotonic_mapping_u128::MonotonicallyMappableToU128;
-pub use self::serialize::{
-    estimate, serialize, serialize_and_load, serialize_u128, NormalizedHeader,
-};
-
-#[derive(PartialEq, Eq, PartialOrd, Ord, Debug, Clone, Copy)]
-#[repr(u8)]
-/// Available codecs to use to encode the u64 (via [`MonotonicallyMappableToU64`]) converted data.
-pub enum FastFieldCodecType {
-    /// Bitpack all values in the value range. The number of bits is defined by the amplitude
-    /// `column.max_value() - column.min_value()`
-    Bitpacked = 1,
-    /// Linear interpolation puts a line between the first and last value and then bitpacks the
-    /// values by the offset from the line. The number of bits is defined by the max deviation from
-    /// the line.
-    Linear = 2,
-    /// Same as [`FastFieldCodecType::Linear`], but encodes in blocks of 512 elements.
-    BlockwiseLinear = 3,
-}
-
-impl BinarySerializable for FastFieldCodecType {
-    fn serialize<W: Write>(&self, wrt: &mut W) -> io::Result<()> {
-        self.to_code().serialize(wrt)
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let code = u8::deserialize(reader)?;
-        let codec_type: Self = Self::from_code(code)
-            .ok_or_else(|| io::Error::new(io::ErrorKind::InvalidData, "Unknown code `{code}.`"))?;
-        Ok(codec_type)
-    }
-}
-
-impl FastFieldCodecType {
-    pub(crate) fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub(crate) fn from_code(code: u8) -> Option<Self> {
-        match code {
-            1 => Some(Self::Bitpacked),
-            2 => Some(Self::Linear),
-            3 => Some(Self::BlockwiseLinear),
-            _ => None,
-        }
-    }
-}
-
-#[derive(PartialEq, Eq, PartialOrd, Ord, Debug, Clone, Copy)]
-#[repr(u8)]
-/// Available codecs to use to encode the u128 (via [`MonotonicallyMappableToU128`]) converted data.
-pub enum U128FastFieldCodecType {
-    /// This codec takes a large number space (u128) and reduces it to a compact number space, by
-    /// removing the holes.
-    CompactSpace = 1,
-}
-
-impl BinarySerializable for U128FastFieldCodecType {
-    fn serialize<W: Write>(&self, wrt: &mut W) -> io::Result<()> {
-        self.to_code().serialize(wrt)
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let code = u8::deserialize(reader)?;
-        let codec_type: Self = Self::from_code(code)
-            .ok_or_else(|| io::Error::new(io::ErrorKind::InvalidData, "Unknown code `{code}.`"))?;
-        Ok(codec_type)
-    }
-}
-
-impl U128FastFieldCodecType {
-    pub(crate) fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub(crate) fn from_code(code: u8) -> Option<Self> {
-        match code {
-            1 => Some(Self::CompactSpace),
-            _ => None,
-        }
-    }
-}
-
-/// Returns the correct codec reader wrapped in the `Arc` for the data.
-pub fn open_u128<Item: MonotonicallyMappableToU128>(
-    bytes: OwnedBytes,
-) -> io::Result<Arc<dyn Column<Item>>> {
-    let (bytes, _format_version) = read_format_version(bytes)?;
-    let (mut bytes, _null_index_footer) = read_null_index_footer(bytes)?;
-    let header = U128Header::deserialize(&mut bytes)?;
-    assert_eq!(header.codec_type, U128FastFieldCodecType::CompactSpace);
-    let reader = CompactSpaceDecompressor::open(bytes)?;
-    let inverted: StrictlyMonotonicMappingInverter<StrictlyMonotonicMappingToInternal<Item>> =
-        StrictlyMonotonicMappingToInternal::<Item>::new().into();
-    Ok(Arc::new(monotonic_map_column(reader, inverted)))
-}
-
-/// Returns the correct codec reader wrapped in the `Arc` for the data.
-pub fn open<T: MonotonicallyMappableToU64>(bytes: OwnedBytes) -> io::Result<Arc<dyn Column<T>>> {
-    let (bytes, _format_version) = read_format_version(bytes)?;
-    let (mut bytes, _null_index_footer) = read_null_index_footer(bytes)?;
-    let header = Header::deserialize(&mut bytes)?;
-    match header.codec_type {
-        FastFieldCodecType::Bitpacked => open_specific_codec::<BitpackedCodec, _>(bytes, &header),
-        FastFieldCodecType::Linear => open_specific_codec::<LinearCodec, _>(bytes, &header),
-        FastFieldCodecType::BlockwiseLinear => {
-            open_specific_codec::<BlockwiseLinearCodec, _>(bytes, &header)
-        }
-    }
-}
-
-fn open_specific_codec<C: FastFieldCodec, Item: MonotonicallyMappableToU64>(
-    bytes: OwnedBytes,
-    header: &Header,
-) -> io::Result<Arc<dyn Column<Item>>> {
-    let normalized_header = header.normalized();
-    let reader = C::open_from_bytes(bytes, normalized_header)?;
-    let min_value = header.min_value;
-    if let Some(gcd) = header.gcd {
-        let mapping = StrictlyMonotonicMappingInverter::from(
-            StrictlyMonotonicMappingToInternalGCDBaseval::new(gcd.get(), min_value),
-        );
-        Ok(Arc::new(monotonic_map_column(reader, mapping)))
-    } else {
-        let mapping = StrictlyMonotonicMappingInverter::from(
-            StrictlyMonotonicMappingToInternalBaseval::new(min_value),
-        );
-        Ok(Arc::new(monotonic_map_column(reader, mapping)))
-    }
-}
-
-/// The FastFieldSerializerEstimate trait is required on all variants
-/// of fast field compressions, to decide which one to choose.
-trait FastFieldCodec: 'static {
-    /// A codex needs to provide a unique name and id, which is
-    /// used for debugging and de/serialization.
-    const CODEC_TYPE: FastFieldCodecType;
-
-    type Reader: Column<u64> + 'static;
-
-    /// Reads the metadata and returns the CodecReader
-    fn open_from_bytes(bytes: OwnedBytes, header: NormalizedHeader) -> io::Result<Self::Reader>;
-
-    /// Serializes the data using the serializer into write.
-    ///
-    /// The column iterator should be preferred over using column `get_val` method for
-    /// performance reasons.
-    fn serialize(column: &dyn Column, write: &mut impl Write) -> io::Result<()>;
-
-    /// Returns an estimate of the compression ratio.
-    /// If the codec is not applicable, returns `None`.
-    ///
-    /// The baseline is uncompressed 64bit data.
-    ///
-    /// It could make sense to also return a value representing
-    /// computational complexity.
-    fn estimate(column: &dyn Column) -> Option<f32>;
-}
-
-/// The list of all available codecs for u64 convertible data.
-pub const ALL_CODEC_TYPES: [FastFieldCodecType; 3] = [
-    FastFieldCodecType::Bitpacked,
-    FastFieldCodecType::BlockwiseLinear,
-    FastFieldCodecType::Linear,
-];
-
-#[cfg(test)]
-mod tests {
-
-    use proptest::prelude::*;
-    use proptest::strategy::Strategy;
-    use proptest::{prop_oneof, proptest};
-
-    use crate::bitpacked::BitpackedCodec;
-    use crate::blockwise_linear::BlockwiseLinearCodec;
-    use crate::linear::LinearCodec;
-    use crate::serialize::Header;
-
-    pub(crate) fn create_and_validate<Codec: FastFieldCodec>(
-        data: &[u64],
-        name: &str,
-    ) -> Option<(f32, f32)> {
-        let col = &VecColumn::from(data);
-        let header = Header::compute_header(col, &[Codec::CODEC_TYPE])?;
-        let normalized_col = header.normalize_column(col);
-        let estimation = Codec::estimate(&normalized_col)?;
-
-        let mut out = Vec::new();
-        let col = VecColumn::from(data);
-        serialize(col, &mut out, &[Codec::CODEC_TYPE]).unwrap();
-
-        let actual_compression = out.len() as f32 / (data.len() as f32 * 8.0);
-
-        let reader = crate::open::<u64>(OwnedBytes::new(out)).unwrap();
-        assert_eq!(reader.num_vals(), data.len() as u32);
-        for (doc, orig_val) in data.iter().copied().enumerate() {
-            let val = reader.get_val(doc as u32);
-            assert_eq!(
-                val, orig_val,
-                "val `{val}` does not match orig_val {orig_val:?}, in data set {name}, data \
-                 `{data:?}`",
-            );
-        }
-
-        if !data.is_empty() {
-            let test_rand_idx = rand::thread_rng().gen_range(0..=data.len() - 1);
-            let expected_positions: Vec<u32> = data
-                .iter()
-                .enumerate()
-                .filter(|(_, el)| **el == data[test_rand_idx])
-                .map(|(pos, _)| pos as u32)
-                .collect();
-            let mut positions = Vec::new();
-            reader.get_docids_for_value_range(
-                data[test_rand_idx]..=data[test_rand_idx],
-                0..data.len() as u32,
-                &mut positions,
-            );
-            assert_eq!(expected_positions, positions);
-        }
-        Some((estimation, actual_compression))
-    }
-
-    proptest! {
-        #![proptest_config(ProptestConfig::with_cases(100))]
-
-        #[test]
-        fn test_proptest_small_bitpacked(data in proptest::collection::vec(num_strategy(), 1..10)) {
-            create_and_validate::<BitpackedCodec>(&data, "proptest bitpacked");
-        }
-
-        #[test]
-        fn test_proptest_small_linear(data in proptest::collection::vec(num_strategy(), 1..10)) {
-            create_and_validate::<LinearCodec>(&data, "proptest linearinterpol");
-        }
-
-        #[test]
-        fn test_proptest_small_blockwise_linear(data in proptest::collection::vec(num_strategy(), 1..10)) {
-            create_and_validate::<BlockwiseLinearCodec>(&data, "proptest multilinearinterpol");
-        }
-    }
-
-    proptest! {
-        #![proptest_config(ProptestConfig::with_cases(10))]
-
-        #[test]
-        fn test_proptest_large_bitpacked(data in proptest::collection::vec(num_strategy(), 1..6000)) {
-            create_and_validate::<BitpackedCodec>(&data, "proptest bitpacked");
-        }
-
-        #[test]
-        fn test_proptest_large_linear(data in proptest::collection::vec(num_strategy(), 1..6000)) {
-            create_and_validate::<LinearCodec>(&data, "proptest linearinterpol");
-        }
-
-        #[test]
-        fn test_proptest_large_blockwise_linear(data in proptest::collection::vec(num_strategy(), 1..6000)) {
-            create_and_validate::<BlockwiseLinearCodec>(&data, "proptest multilinearinterpol");
-        }
-    }
-
-    fn num_strategy() -> impl Strategy<Value = u64> {
-        prop_oneof![
-            1 => prop::num::u64::ANY.prop_map(|num| u64::MAX - (num % 10) ),
-            1 => prop::num::u64::ANY.prop_map(|num| num % 10 ),
-            20 => prop::num::u64::ANY,
-        ]
-    }
-
-    pub fn get_codec_test_datasets() -> Vec<(Vec<u64>, &'static str)> {
-        let mut data_and_names = vec![];
-
-        let data = (10..=10_000_u64).collect::<Vec<_>>();
-        data_and_names.push((data, "simple monotonically increasing"));
-
-        data_and_names.push((
-            vec![5, 6, 7, 8, 9, 10, 99, 100],
-            "offset in linear interpol",
-        ));
-        data_and_names.push((vec![5, 50, 3, 13, 1, 1000, 35], "rand small"));
-        data_and_names.push((vec![10], "single value"));
-
-        data_and_names.push((
-            vec![1572656989877777, 1170935903116329, 720575940379279, 0],
-            "overflow error",
-        ));
-
-        data_and_names
-    }
-
-    fn test_codec<C: FastFieldCodec>() {
-        let codec_name = format!("{:?}", C::CODEC_TYPE);
-        for (data, dataset_name) in get_codec_test_datasets() {
-            let estimate_actual_opt: Option<(f32, f32)> =
-                crate::tests::create_and_validate::<C>(&data, dataset_name);
-            let result = if let Some((estimate, actual)) = estimate_actual_opt {
-                format!("Estimate `{estimate}` Actual `{actual}`")
-            } else {
-                "Disabled".to_string()
-            };
-            println!("Codec {codec_name}, DataSet {dataset_name}, {result}");
-        }
-    }
-    #[test]
-    fn test_codec_bitpacking() {
-        test_codec::<BitpackedCodec>();
-    }
-    #[test]
-    fn test_codec_interpolation() {
-        test_codec::<LinearCodec>();
-    }
-    #[test]
-    fn test_codec_multi_interpolation() {
-        test_codec::<BlockwiseLinearCodec>();
-    }
-
-    use super::*;
-
-    #[test]
-    fn estimation_good_interpolation_case() {
-        let data = (10..=20000_u64).collect::<Vec<_>>();
-        let data: VecColumn = data.as_slice().into();
-
-        let linear_interpol_estimation = LinearCodec::estimate(&data).unwrap();
-        assert_le!(linear_interpol_estimation, 0.01);
-
-        let multi_linear_interpol_estimation = BlockwiseLinearCodec::estimate(&data).unwrap();
-        assert_le!(multi_linear_interpol_estimation, 0.2);
-        assert_lt!(linear_interpol_estimation, multi_linear_interpol_estimation);
-
-        let bitpacked_estimation = BitpackedCodec::estimate(&data).unwrap();
-        assert_lt!(linear_interpol_estimation, bitpacked_estimation);
-    }
-    #[test]
-    fn estimation_test_bad_interpolation_case() {
-        let data: &[u64] = &[200, 10, 10, 10, 10, 1000, 20];
-
-        let data: VecColumn = data.into();
-        let linear_interpol_estimation = LinearCodec::estimate(&data).unwrap();
-        assert_le!(linear_interpol_estimation, 0.34);
-
-        let bitpacked_estimation = BitpackedCodec::estimate(&data).unwrap();
-        assert_lt!(bitpacked_estimation, linear_interpol_estimation);
-    }
-
-    #[test]
-    fn estimation_prefer_bitpacked() {
-        let data = VecColumn::from(&[10, 10, 10, 10]);
-        let linear_interpol_estimation = LinearCodec::estimate(&data).unwrap();
-        let bitpacked_estimation = BitpackedCodec::estimate(&data).unwrap();
-        assert_lt!(bitpacked_estimation, linear_interpol_estimation);
-    }
-
-    #[test]
-    fn estimation_test_bad_interpolation_case_monotonically_increasing() {
-        let mut data: Vec<u64> = (201..=20000_u64).collect();
-        data.push(1_000_000);
-        let data: VecColumn = data.as_slice().into();
-
-        // in this case the linear interpolation can't in fact not be worse than bitpacking,
-        // but the estimator adds some threshold, which leads to estimated worse behavior
-        let linear_interpol_estimation = LinearCodec::estimate(&data).unwrap();
-        assert_le!(linear_interpol_estimation, 0.35);
-
-        let bitpacked_estimation = BitpackedCodec::estimate(&data).unwrap();
-        assert_le!(bitpacked_estimation, 0.32);
-        assert_le!(bitpacked_estimation, linear_interpol_estimation);
-    }
-
-    #[test]
-    fn test_fast_field_codec_type_to_code() {
-        let mut count_codec = 0;
-        for code in 0..=255 {
-            if let Some(codec_type) = FastFieldCodecType::from_code(code) {
-                assert_eq!(codec_type.to_code(), code);
-                count_codec += 1;
-            }
-        }
-        assert_eq!(count_codec, 3);
-    }
-}
-
-#[cfg(all(test, feature = "unstable"))]
-mod bench {
-    use std::sync::Arc;
-
-    use ownedbytes::OwnedBytes;
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
-    use test::{self, Bencher};
-
-    use super::*;
-    use crate::Column;
-
-    fn get_data() -> Vec<u64> {
-        let mut rng = StdRng::seed_from_u64(2u64);
-        let mut data: Vec<_> = (100..55000_u64)
-            .map(|num| num + rng.gen::<u8>() as u64)
-            .collect();
-        data.push(99_000);
-        data.insert(1000, 2000);
-        data.insert(2000, 100);
-        data.insert(3000, 4100);
-        data.insert(4000, 100);
-        data.insert(5000, 800);
-        data
-    }
-
-    #[inline(never)]
-    fn value_iter() -> impl Iterator<Item = u64> {
-        0..20_000
-    }
-    fn get_reader_for_bench<Codec: FastFieldCodec>(data: &[u64]) -> Codec::Reader {
-        let mut bytes = Vec::new();
-        let min_value = *data.iter().min().unwrap();
-        let data = data.iter().map(|el| *el - min_value).collect::<Vec<_>>();
-        let col = VecColumn::from(&data);
-        let normalized_header = crate::NormalizedHeader {
-            num_vals: col.num_vals(),
-            max_value: col.max_value(),
-        };
-        Codec::serialize(&VecColumn::from(&data), &mut bytes).unwrap();
-        Codec::open_from_bytes(OwnedBytes::new(bytes), normalized_header).unwrap()
-    }
-    fn bench_get<Codec: FastFieldCodec>(b: &mut Bencher, data: &[u64]) {
-        let col = get_reader_for_bench::<Codec>(data);
-        b.iter(|| {
-            let mut sum = 0u64;
-            for pos in value_iter() {
-                let val = col.get_val(pos as u32);
-                sum = sum.wrapping_add(val);
-            }
-            sum
-        });
-    }
-
-    #[inline(never)]
-    fn bench_get_dynamic_helper(b: &mut Bencher, col: Arc<dyn Column>) {
-        b.iter(|| {
-            let mut sum = 0u64;
-            for pos in value_iter() {
-                let val = col.get_val(pos as u32);
-                sum = sum.wrapping_add(val);
-            }
-            sum
-        });
-    }
-
-    fn bench_get_dynamic<Codec: FastFieldCodec>(b: &mut Bencher, data: &[u64]) {
-        let col = Arc::new(get_reader_for_bench::<Codec>(data));
-        bench_get_dynamic_helper(b, col);
-    }
-    fn bench_create<Codec: FastFieldCodec>(b: &mut Bencher, data: &[u64]) {
-        let min_value = *data.iter().min().unwrap();
-        let data = data.iter().map(|el| *el - min_value).collect::<Vec<_>>();
-
-        let mut bytes = Vec::new();
-        b.iter(|| {
-            bytes.clear();
-            Codec::serialize(&VecColumn::from(&data), &mut bytes).unwrap();
-        });
-    }
-
-    #[bench]
-    fn bench_fastfield_bitpack_create(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_create::<BitpackedCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_linearinterpol_create(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_create::<LinearCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_multilinearinterpol_create(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_create::<BlockwiseLinearCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_bitpack_get(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get::<BitpackedCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_bitpack_get_dynamic(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get_dynamic::<BitpackedCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_linearinterpol_get(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get::<LinearCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_linearinterpol_get_dynamic(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get_dynamic::<LinearCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_multilinearinterpol_get(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get::<BlockwiseLinearCodec>(b, &data);
-    }
-    #[bench]
-    fn bench_fastfield_multilinearinterpol_get_dynamic(b: &mut Bencher) {
-        let data: Vec<_> = get_data();
-        bench_get_dynamic::<BlockwiseLinearCodec>(b, &data);
-    }
-}

fastfield_codecs/src/line.rs

@@ -1,222 +0,0 @@
-use std::io;
-use std::num::NonZeroU32;
-
-use common::{BinarySerializable, VInt};
-
-use crate::Column;
-
-const MID_POINT: u64 = (1u64 << 32) - 1u64;
-
-/// `Line` describes a line function `y: ax + b` using integer
-/// arithmetics.
-///
-/// The slope is in fact a decimal split into a 32 bit integer value,
-/// and a 32-bit decimal value.
-///
-/// The multiplication then becomes.
-/// `y = m * x >> 32 + b`
-#[derive(Debug, Clone, Copy, Default)]
-pub struct Line {
-    slope: u64,
-    intercept: u64,
-}
-
-/// Compute the line slope.
-///
-/// This function has the nice property of being
-/// invariant by translation.
-/// `
-///   compute_slope(y0, y1)
-/// = compute_slope(y0 + X % 2^64, y1 + X % 2^64)
-/// `
-fn compute_slope(y0: u64, y1: u64, num_vals: NonZeroU32) -> u64 {
-    let dy = y1.wrapping_sub(y0);
-    let sign = dy <= (1 << 63);
-    let abs_dy = if sign {
-        y1.wrapping_sub(y0)
-    } else {
-        y0.wrapping_sub(y1)
-    };
-    if abs_dy >= 1 << 32 {
-        // This is outside of realm we handle.
-        // Let's just bail.
-        return 0u64;
-    }
-
-    let abs_slope = (abs_dy << 32) / num_vals.get() as u64;
-    if sign {
-        abs_slope
-    } else {
-        // The complement does indeed create the
-        // opposite decreasing slope...
-        //
-        // Intuitively (without the bitshifts and % u64::MAX)
-        // ```
-        //    (x + shift)*(u64::MAX - abs_slope)
-        // -  (x * (u64::MAX - abs_slope))
-        // = - shift * abs_slope
-        // ```
-        u64::MAX - abs_slope
-    }
-}
-
-impl Line {
-    #[inline(always)]
-    pub fn eval(&self, x: u32) -> u64 {
-        let linear_part = ((x as u64).wrapping_mul(self.slope) >> 32) as i32 as u64;
-        self.intercept.wrapping_add(linear_part)
-    }
-
-    // Same as train, but the intercept is only estimated from provided sample positions
-    pub fn estimate(sample_positions_and_values: &[(u64, u64)]) -> Self {
-        let first_val = sample_positions_and_values[0].1;
-        let last_val = sample_positions_and_values[sample_positions_and_values.len() - 1].1;
-        let num_vals = sample_positions_and_values[sample_positions_and_values.len() - 1].0 + 1;
-        Self::train_from(
-            first_val,
-            last_val,
-            num_vals as u32,
-            sample_positions_and_values.iter().cloned(),
-        )
-    }
-
-    // Intercept is only computed from provided positions
-    fn train_from(
-        first_val: u64,
-        last_val: u64,
-        num_vals: u32,
-        positions_and_values: impl Iterator<Item = (u64, u64)>,
-    ) -> Self {
-        // TODO replace with let else
-        let idx_last_val = if let Some(idx_last_val) = NonZeroU32::new(num_vals - 1) {
-            idx_last_val
-        } else {
-            return Line::default();
-        };
-
-        let y0 = first_val;
-        let y1 = last_val;
-
-        // We first independently pick our slope.
-        let slope = compute_slope(y0, y1, idx_last_val);
-
-        // We picked our slope. Note that it does not have to be perfect.
-        // Now we need to compute the best intercept.
-        //
-        // Intuitively, the best intercept is such that line passes through one of the
-        // `(i, ys[])`.
-        //
-        // The best intercept therefore has the form
-        // `y[i] - line.eval(i)` (using wrapping arithmetics).
-        // In other words, the best intercept is one of the `y - Line::eval(ys[i])`
-        // and our task is just to pick the one that minimizes our error.
-        //
-        // Without sorting our values, this is a difficult problem.
-        // We however rely on the following trick...
-        //
-        // We only focus on the case where the interpolation is half decent.
-        // If the line interpolation is doing its job on a dataset suited for it,
-        // we can hope that the maximum error won't be larger than `u64::MAX / 2`.
-        //
-        // In other words, even without the intercept the values `y - Line::eval(ys[i])` will all be
-        // within an interval that takes less than half of the modulo space of `u64`.
-        //
-        // Our task is therefore to identify this interval.
-        // Here we simply translate all of our values by `y0 - 2^63` and pick the min.
-        let mut line = Line {
-            slope,
-            intercept: 0,
-        };
-        let heuristic_shift = y0.wrapping_sub(MID_POINT);
-        line.intercept = positions_and_values
-            .map(|(pos, y)| y.wrapping_sub(line.eval(pos as u32)))
-            .min_by_key(|&val| val.wrapping_sub(heuristic_shift))
-            .unwrap_or(0u64); //< Never happens.
-        line
-    }
-
-    /// Returns a line that attemps to approximate a function
-    /// f: i in 0..[ys.num_vals()) -> ys[i].
-    ///
-    /// - The approximation is always lower than the actual value.
-    /// Or more rigorously, formally `f(i).wrapping_sub(ys[i])` is small
-    /// for any i in [0..ys.len()).
-    /// - It computes without panicking for any value of it.
-    ///
-    /// This function is only invariable by translation if all of the
-    /// `ys` are packaged into half of the space. (See heuristic below)
-    pub fn train(ys: &dyn Column) -> Self {
-        let first_val = ys.iter().next().unwrap();
-        let last_val = ys.iter().nth(ys.num_vals() as usize - 1).unwrap();
-        Self::train_from(
-            first_val,
-            last_val,
-            ys.num_vals(),
-            ys.iter().enumerate().map(|(pos, val)| (pos as u64, val)),
-        )
-    }
-}
-
-impl BinarySerializable for Line {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        VInt(self.slope).serialize(writer)?;
-        VInt(self.intercept).serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let slope = VInt::deserialize(reader)?.0;
-        let intercept = VInt::deserialize(reader)?.0;
-        Ok(Line { slope, intercept })
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use crate::VecColumn;
-
-    /// Test training a line and ensuring that the maximum difference between
-    /// the data points and the line is `expected`.
-    ///
-    /// This function operates translation over the data for better coverage.
-    #[track_caller]
-    fn test_line_interpol_with_translation(ys: &[u64], expected: Option<u64>) {
-        let mut translations = vec![0, 100, u64::MAX / 2, u64::MAX, u64::MAX - 1];
-        translations.extend_from_slice(ys);
-        for translation in translations {
-            let translated_ys: Vec<u64> = ys
-                .iter()
-                .copied()
-                .map(|y| y.wrapping_add(translation))
-                .collect();
-            let largest_err = test_eval_max_err(&translated_ys);
-            assert_eq!(largest_err, expected);
-        }
-    }
-
-    fn test_eval_max_err(ys: &[u64]) -> Option<u64> {
-        let line = Line::train(&VecColumn::from(&ys));
-        ys.iter()
-            .enumerate()
-            .map(|(x, y)| y.wrapping_sub(line.eval(x as u32)))
-            .max()
-    }
-
-    #[test]
-    fn test_train() {
-        test_line_interpol_with_translation(&[11, 11, 11, 12, 12, 13], Some(1));
-        test_line_interpol_with_translation(&[13, 12, 12, 11, 11, 11], Some(1));
-        test_line_interpol_with_translation(&[13, 13, 12, 11, 11, 11], Some(1));
-        test_line_interpol_with_translation(&[13, 13, 12, 11, 11, 11], Some(1));
-        test_line_interpol_with_translation(&[u64::MAX - 1, 0, 0, 1], Some(1));
-        test_line_interpol_with_translation(&[u64::MAX - 1, u64::MAX, 0, 1], Some(0));
-        test_line_interpol_with_translation(&[0, 1, 2, 3, 5], Some(0));
-        test_line_interpol_with_translation(&[1, 2, 3, 4], Some(0));
-
-        let data: Vec<u64> = (0..255).collect();
-        test_line_interpol_with_translation(&data, Some(0));
-        let data: Vec<u64> = (0..255).map(|el| el * 2).collect();
-        test_line_interpol_with_translation(&data, Some(0));
-    }
-}

fastfield_codecs/src/linear.rs

@@ -1,231 +0,0 @@
-use std::io::{self, Write};
-
-use common::BinarySerializable;
-use ownedbytes::OwnedBytes;
-use tantivy_bitpacker::{compute_num_bits, BitPacker, BitUnpacker};
-
-use crate::line::Line;
-use crate::serialize::NormalizedHeader;
-use crate::{Column, FastFieldCodec, FastFieldCodecType};
-
-/// Depending on the field type, a different
-/// fast field is required.
-#[derive(Clone)]
-pub struct LinearReader {
-    data: OwnedBytes,
-    linear_params: LinearParams,
-    header: NormalizedHeader,
-}
-
-impl Column for LinearReader {
-    #[inline]
-    fn get_val(&self, doc: u32) -> u64 {
-        let interpoled_val: u64 = self.linear_params.line.eval(doc);
-        let bitpacked_diff = self.linear_params.bit_unpacker.get(doc, &self.data);
-        interpoled_val.wrapping_add(bitpacked_diff)
-    }
-
-    #[inline]
-    fn min_value(&self) -> u64 {
-        // The LinearReader assumes a normalized vector.
-        0u64
-    }
-
-    #[inline]
-    fn max_value(&self) -> u64 {
-        self.header.max_value
-    }
-
-    #[inline]
-    fn num_vals(&self) -> u32 {
-        self.header.num_vals
-    }
-}
-
-/// Fastfield serializer, which tries to guess values by linear interpolation
-/// and stores the difference bitpacked.
-pub struct LinearCodec;
-
-#[derive(Debug, Clone)]
-struct LinearParams {
-    line: Line,
-    bit_unpacker: BitUnpacker,
-}
-
-impl BinarySerializable for LinearParams {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        self.line.serialize(writer)?;
-        self.bit_unpacker.bit_width().serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let line = Line::deserialize(reader)?;
-        let bit_width = u8::deserialize(reader)?;
-        Ok(Self {
-            line,
-            bit_unpacker: BitUnpacker::new(bit_width),
-        })
-    }
-}
-
-impl FastFieldCodec for LinearCodec {
-    const CODEC_TYPE: FastFieldCodecType = FastFieldCodecType::Linear;
-
-    type Reader = LinearReader;
-
-    /// Opens a fast field given a file.
-    fn open_from_bytes(mut data: OwnedBytes, header: NormalizedHeader) -> io::Result<Self::Reader> {
-        let linear_params = LinearParams::deserialize(&mut data)?;
-        Ok(LinearReader {
-            data,
-            linear_params,
-            header,
-        })
-    }
-
-    /// Creates a new fast field serializer.
-    fn serialize(column: &dyn Column, write: &mut impl Write) -> io::Result<()> {
-        assert_eq!(column.min_value(), 0);
-        let line = Line::train(column);
-
-        let max_offset_from_line = column
-            .iter()
-            .enumerate()
-            .map(|(pos, actual_value)| {
-                let calculated_value = line.eval(pos as u32);
-                actual_value.wrapping_sub(calculated_value)
-            })
-            .max()
-            .unwrap();
-
-        let num_bits = compute_num_bits(max_offset_from_line);
-        let linear_params = LinearParams {
-            line,
-            bit_unpacker: BitUnpacker::new(num_bits),
-        };
-        linear_params.serialize(write)?;
-
-        let mut bit_packer = BitPacker::new();
-        for (pos, actual_value) in column.iter().enumerate() {
-            let calculated_value = line.eval(pos as u32);
-            let offset = actual_value.wrapping_sub(calculated_value);
-            bit_packer.write(offset, num_bits, write)?;
-        }
-        bit_packer.close(write)?;
-
-        Ok(())
-    }
-
-    /// estimation for linear interpolation is hard because, you don't know
-    /// where the local maxima for the deviation of the calculated value are and
-    /// the offset to shift all values to >=0 is also unknown.
-    #[allow(clippy::question_mark)]
-    fn estimate(column: &dyn Column) -> Option<f32> {
-        if column.num_vals() < 3 {
-            return None; // disable compressor for this case
-        }
-
-        let limit_num_vals = column.num_vals().min(100_000);
-
-        let num_samples = 100;
-        let step_size = (limit_num_vals / num_samples).max(1); // 20 samples
-        let mut sample_positions_and_values: Vec<_> = Vec::new();
-        for (pos, val) in column.iter().enumerate().step_by(step_size as usize) {
-            sample_positions_and_values.push((pos as u64, val));
-        }
-
-        let line = Line::estimate(&sample_positions_and_values);
-
-        let estimated_bit_width = sample_positions_and_values
-            .into_iter()
-            .map(|(pos, actual_value)| {
-                let interpolated_val = line.eval(pos as u32);
-                actual_value.wrapping_sub(interpolated_val)
-            })
-            .map(|diff| ((diff as f32 * 1.5) * 2.0) as u64)
-            .map(compute_num_bits)
-            .max()
-            .unwrap_or(0);
-
-        // Extrapolate to whole column
-        let num_bits = (estimated_bit_width as u64 * column.num_vals() as u64) + 64;
-        let num_bits_uncompressed = 64 * column.num_vals();
-        Some(num_bits as f32 / num_bits_uncompressed as f32)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use rand::RngCore;
-
-    use super::*;
-    use crate::tests::get_codec_test_datasets;
-
-    fn create_and_validate(data: &[u64], name: &str) -> Option<(f32, f32)> {
-        crate::tests::create_and_validate::<LinearCodec>(data, name)
-    }
-
-    #[test]
-    fn test_compression() {
-        let data = (10..=6_000_u64).collect::<Vec<_>>();
-        let (estimate, actual_compression) =
-            create_and_validate(&data, "simple monotonically large").unwrap();
-
-        assert_le!(actual_compression, 0.001);
-        assert_le!(estimate, 0.02);
-    }
-
-    #[test]
-    fn test_with_codec_datasets() {
-        let data_sets = get_codec_test_datasets();
-        for (mut data, name) in data_sets {
-            create_and_validate(&data, name);
-            data.reverse();
-            create_and_validate(&data, name);
-        }
-    }
-    #[test]
-    fn linear_interpol_fast_field_test_large_amplitude() {
-        let data = vec![
-            i64::MAX as u64 / 2,
-            i64::MAX as u64 / 3,
-            i64::MAX as u64 / 2,
-        ];
-
-        create_and_validate(&data, "large amplitude");
-    }
-
-    #[test]
-    fn overflow_error_test() {
-        let data = vec![1572656989877777, 1170935903116329, 720575940379279, 0];
-        create_and_validate(&data, "overflow test");
-    }
-
-    #[test]
-    fn linear_interpol_fast_concave_data() {
-        let data = vec![0, 1, 2, 5, 8, 10, 20, 50];
-        create_and_validate(&data, "concave data");
-    }
-    #[test]
-    fn linear_interpol_fast_convex_data() {
-        let data = vec![0, 40, 60, 70, 75, 77];
-        create_and_validate(&data, "convex data");
-    }
-    #[test]
-    fn linear_interpol_fast_field_test_simple() {
-        let data = (10..=20_u64).collect::<Vec<_>>();
-        create_and_validate(&data, "simple monotonically");
-    }
-
-    #[test]
-    fn linear_interpol_fast_field_rand() {
-        let mut rng = rand::thread_rng();
-        for _ in 0..50 {
-            let mut data = (0..10_000).map(|_| rng.next_u64()).collect::<Vec<_>>();
-            create_and_validate(&data, "random");
-            data.reverse();
-            create_and_validate(&data, "random");
-        }
-    }
-}

fastfield_codecs/src/main.rs

@@ -1,222 +0,0 @@
-#[macro_use]
-extern crate prettytable;
-use std::collections::HashSet;
-use std::env;
-use std::io::BufRead;
-use std::net::{IpAddr, Ipv6Addr};
-use std::str::FromStr;
-
-use fastfield_codecs::{open_u128, serialize_u128, Column, FastFieldCodecType, VecColumn};
-use itertools::Itertools;
-use measure_time::print_time;
-use ownedbytes::OwnedBytes;
-use prettytable::{Cell, Row, Table};
-
-fn print_set_stats(ip_addrs: &[u128]) {
-    println!("NumIps\t{}", ip_addrs.len());
-    let ip_addr_set: HashSet<u128> = ip_addrs.iter().cloned().collect();
-    println!("NumUniqueIps\t{}", ip_addr_set.len());
-    let ratio_unique = ip_addr_set.len() as f64 / ip_addrs.len() as f64;
-    println!("RatioUniqueOverTotal\t{ratio_unique:.4}");
-
-    // histogram
-    let mut ip_addrs = ip_addrs.to_vec();
-    ip_addrs.sort();
-    let mut cnts: Vec<usize> = ip_addrs
-        .into_iter()
-        .dedup_with_count()
-        .map(|(cnt, _)| cnt)
-        .collect();
-    cnts.sort();
-
-    let top_256_cnt: usize = cnts.iter().rev().take(256).sum();
-    let top_128_cnt: usize = cnts.iter().rev().take(128).sum();
-    let top_64_cnt: usize = cnts.iter().rev().take(64).sum();
-    let top_8_cnt: usize = cnts.iter().rev().take(8).sum();
-    let total: usize = cnts.iter().sum();
-
-    println!("{}", total);
-    println!("{}", top_256_cnt);
-    println!("{}", top_128_cnt);
-    println!("Percentage Top8 {:02}", top_8_cnt as f32 / total as f32);
-    println!("Percentage Top64 {:02}", top_64_cnt as f32 / total as f32);
-    println!("Percentage Top128 {:02}", top_128_cnt as f32 / total as f32);
-    println!("Percentage Top256 {:02}", top_256_cnt as f32 / total as f32);
-
-    let mut cnts: Vec<(usize, usize)> = cnts.into_iter().dedup_with_count().collect();
-    cnts.sort_by(|a, b| {
-        if a.1 == b.1 {
-            a.0.cmp(&b.0)
-        } else {
-            b.1.cmp(&a.1)
-        }
-    });
-}
-
-fn ip_dataset() -> Vec<u128> {
-    let mut ip_addr_v4 = 0;
-
-    let stdin = std::io::stdin();
-    let ip_addrs: Vec<u128> = stdin
-        .lock()
-        .lines()
-        .flat_map(|line| {
-            let line = line.unwrap();
-            let line = line.trim();
-            let ip_addr = IpAddr::from_str(line.trim()).ok()?;
-            if ip_addr.is_ipv4() {
-                ip_addr_v4 += 1;
-            }
-            let ip_addr_v6: Ipv6Addr = match ip_addr {
-                IpAddr::V4(v4) => v4.to_ipv6_mapped(),
-                IpAddr::V6(v6) => v6,
-            };
-            Some(ip_addr_v6)
-        })
-        .map(|ip_v6| u128::from_be_bytes(ip_v6.octets()))
-        .collect();
-
-    println!("IpAddrsAny\t{}", ip_addrs.len());
-    println!("IpAddrsV4\t{}", ip_addr_v4);
-
-    ip_addrs
-}
-
-fn bench_ip() {
-    let dataset = ip_dataset();
-    print_set_stats(&dataset);
-
-    // Chunks
-    {
-        let mut data = vec![];
-        for dataset in dataset.chunks(500_000) {
-            serialize_u128(|| dataset.iter().cloned(), dataset.len() as u32, &mut data).unwrap();
-        }
-        let compression = data.len() as f64 / (dataset.len() * 16) as f64;
-        println!("Compression 50_000 chunks {:.4}", compression);
-        println!(
-            "Num Bits per elem {:.2}",
-            (data.len() * 8) as f32 / dataset.len() as f32
-        );
-    }
-
-    let mut data = vec![];
-    {
-        print_time!("creation");
-        serialize_u128(|| dataset.iter().cloned(), dataset.len() as u32, &mut data).unwrap();
-    }
-
-    let compression = data.len() as f64 / (dataset.len() * 16) as f64;
-    println!("Compression {:.2}", compression);
-    println!(
-        "Num Bits per elem {:.2}",
-        (data.len() * 8) as f32 / dataset.len() as f32
-    );
-
-    let decompressor = open_u128::<u128>(OwnedBytes::new(data)).unwrap();
-    // Sample some ranges
-    let mut doc_values = Vec::new();
-    for value in dataset.iter().take(1110).skip(1100).cloned() {
-        doc_values.clear();
-        print_time!("get range");
-        decompressor.get_docids_for_value_range(
-            value..=value,
-            0..decompressor.num_vals(),
-            &mut doc_values,
-        );
-        println!("{:?}", doc_values.len());
-    }
-}
-
-fn main() {
-    if env::args().nth(1).unwrap() == "bench_ip" {
-        bench_ip();
-        return;
-    }
-
-    let mut table = Table::new();
-
-    // Add a row per time
-    table.add_row(row!["", "Compression Ratio", "Compression Estimation"]);
-
-    for (data, data_set_name) in get_codec_test_data_sets() {
-        let results: Vec<(f32, f32, FastFieldCodecType)> = [
-            serialize_with_codec(&data, FastFieldCodecType::Bitpacked),
-            serialize_with_codec(&data, FastFieldCodecType::Linear),
-            serialize_with_codec(&data, FastFieldCodecType::BlockwiseLinear),
-        ]
-        .into_iter()
-        .flatten()
-        .collect();
-        let best_compression_ratio_codec = results
-            .iter()
-            .min_by(|&res1, &res2| res1.partial_cmp(res2).unwrap())
-            .cloned()
-            .unwrap();
-
-        table.add_row(Row::new(vec![Cell::new(data_set_name).style_spec("Bbb")]));
-        for (est, comp, codec_type) in results {
-            let est_cell = est.to_string();
-            let ratio_cell = comp.to_string();
-            let style = if comp == best_compression_ratio_codec.1 {
-                "Fb"
-            } else {
-                ""
-            };
-            table.add_row(Row::new(vec![
-                Cell::new(&format!("{codec_type:?}")).style_spec("bFg"),
-                Cell::new(&ratio_cell).style_spec(style),
-                Cell::new(&est_cell).style_spec(""),
-            ]));
-        }
-    }
-
-    table.printstd();
-}
-
-pub fn get_codec_test_data_sets() -> Vec<(Vec<u64>, &'static str)> {
-    let mut data_and_names = vec![];
-
-    let data = (1000..=200_000_u64).collect::<Vec<_>>();
-    data_and_names.push((data, "Autoincrement"));
-
-    let mut current_cumulative = 0;
-    let data = (1..=200_000_u64)
-        .map(|num| {
-            let num = (num as f32 + num as f32).log10() as u64;
-            current_cumulative += num;
-            current_cumulative
-        })
-        .collect::<Vec<_>>();
-    // let data = (1..=200000_u64).map(|num| num + num).collect::<Vec<_>>();
-    data_and_names.push((data, "Monotonically increasing concave"));
-
-    let mut current_cumulative = 0;
-    let data = (1..=200_000_u64)
-        .map(|num| {
-            let num = (200_000.0 - num as f32).log10() as u64;
-            current_cumulative += num;
-            current_cumulative
-        })
-        .collect::<Vec<_>>();
-    data_and_names.push((data, "Monotonically increasing convex"));
-
-    let data = (1000..=200_000_u64)
-        .map(|num| num + rand::random::<u8>() as u64)
-        .collect::<Vec<_>>();
-    data_and_names.push((data, "Almost monotonically increasing"));
-
-    data_and_names
-}
-
-pub fn serialize_with_codec(
-    data: &[u64],
-    codec_type: FastFieldCodecType,
-) -> Option<(f32, f32, FastFieldCodecType)> {
-    let col = VecColumn::from(data);
-    let estimation = fastfield_codecs::estimate(&col, codec_type)?;
-    let mut out = Vec::new();
-    fastfield_codecs::serialize(&col, &mut out, &[codec_type]).ok()?;
-    let actual_compression = out.len() as f32 / (col.num_vals() * 8) as f32;
-    Some((estimation, actual_compression, codec_type))
-}

fastfield_codecs/src/monotonic_mapping.rs

@@ -1,267 +0,0 @@
-use std::marker::PhantomData;
-
-use fastdivide::DividerU64;
-use ordered_float::NotNan;
-
-use crate::MonotonicallyMappableToU128;
-
-/// Monotonic maps a value to u64 value space.
-/// Monotonic mapping enables `PartialOrd` on u64 space without conversion to original space.
-pub trait MonotonicallyMappableToU64: 'static + PartialOrd + Copy + Send + Sync {
-    /// Converts a value to u64.
-    ///
-    /// Internally all fast field values are encoded as u64.
-    fn to_u64(self) -> u64;
-
-    /// Converts a value from u64
-    ///
-    /// Internally all fast field values are encoded as u64.
-    /// **Note: To be used for converting encoded Term, Posting values.**
-    fn from_u64(val: u64) -> Self;
-}
-
-/// Values need to be strictly monotonic mapped to a `Internal` value (u64 or u128) that can be
-/// used in fast field codecs.
-///
-/// The monotonic mapping is required so that `PartialOrd` can be used on `Internal` without
-/// converting to `External`.
-///
-/// All strictly monotonic functions are invertible because they are guaranteed to have a one-to-one
-/// mapping from their range to their domain. The `inverse` method is required when opening a codec,
-/// so a value can be converted back to its original domain (e.g. ip address or f64) from its
-/// internal representation.
-pub trait StrictlyMonotonicFn<External, Internal> {
-    /// Strictly monotonically maps the value from External to Internal.
-    fn mapping(&self, inp: External) -> Internal;
-    /// Inverse of `mapping`. Maps the value from Internal to External.
-    fn inverse(&self, out: Internal) -> External;
-}
-
-/// Inverts a strictly monotonic mapping from `StrictlyMonotonicFn<A, B>` to
-/// `StrictlyMonotonicFn<B, A>`.
-///
-/// # Warning
-///
-/// This type comes with a footgun. A type being strictly monotonic does not impose that the inverse
-/// mapping is strictly monotonic over the entire space External. e.g. a -> a * 2. Use at your own
-/// risks.
-pub(crate) struct StrictlyMonotonicMappingInverter<T> {
-    orig_mapping: T,
-}
-impl<T> From<T> for StrictlyMonotonicMappingInverter<T> {
-    fn from(orig_mapping: T) -> Self {
-        Self { orig_mapping }
-    }
-}
-
-impl<From, To, T> StrictlyMonotonicFn<To, From> for StrictlyMonotonicMappingInverter<T>
-where T: StrictlyMonotonicFn<From, To>
-{
-    fn mapping(&self, val: To) -> From {
-        self.orig_mapping.inverse(val)
-    }
-
-    fn inverse(&self, val: From) -> To {
-        self.orig_mapping.mapping(val)
-    }
-}
-
-/// Applies the strictly monotonic mapping from `T` without any additional changes.
-pub(crate) struct StrictlyMonotonicMappingToInternal<T> {
-    _phantom: PhantomData<T>,
-}
-
-impl<T> StrictlyMonotonicMappingToInternal<T> {
-    pub(crate) fn new() -> StrictlyMonotonicMappingToInternal<T> {
-        Self {
-            _phantom: PhantomData,
-        }
-    }
-}
-
-impl<External: MonotonicallyMappableToU128, T: MonotonicallyMappableToU128>
-    StrictlyMonotonicFn<External, u128> for StrictlyMonotonicMappingToInternal<T>
-where T: MonotonicallyMappableToU128
-{
-    fn mapping(&self, inp: External) -> u128 {
-        External::to_u128(inp)
-    }
-
-    fn inverse(&self, out: u128) -> External {
-        External::from_u128(out)
-    }
-}
-
-impl<External: MonotonicallyMappableToU64, T: MonotonicallyMappableToU64>
-    StrictlyMonotonicFn<External, u64> for StrictlyMonotonicMappingToInternal<T>
-where T: MonotonicallyMappableToU64
-{
-    fn mapping(&self, inp: External) -> u64 {
-        External::to_u64(inp)
-    }
-
-    fn inverse(&self, out: u64) -> External {
-        External::from_u64(out)
-    }
-}
-
-/// Mapping dividing by  gcd and a base value.
-///
-/// The function is assumed to be only called on values divided by passed
-/// gcd value. (It is necessary for the function to be monotonic.)
-pub(crate) struct StrictlyMonotonicMappingToInternalGCDBaseval {
-    gcd_divider: DividerU64,
-    gcd: u64,
-    min_value: u64,
-}
-impl StrictlyMonotonicMappingToInternalGCDBaseval {
-    pub(crate) fn new(gcd: u64, min_value: u64) -> Self {
-        let gcd_divider = DividerU64::divide_by(gcd);
-        Self {
-            gcd_divider,
-            gcd,
-            min_value,
-        }
-    }
-}
-impl<External: MonotonicallyMappableToU64> StrictlyMonotonicFn<External, u64>
-    for StrictlyMonotonicMappingToInternalGCDBaseval
-{
-    fn mapping(&self, inp: External) -> u64 {
-        self.gcd_divider
-            .divide(External::to_u64(inp) - self.min_value)
-    }
-
-    fn inverse(&self, out: u64) -> External {
-        External::from_u64(self.min_value + out * self.gcd)
-    }
-}
-
-/// Strictly monotonic mapping with a base value.
-pub(crate) struct StrictlyMonotonicMappingToInternalBaseval {
-    min_value: u64,
-}
-impl StrictlyMonotonicMappingToInternalBaseval {
-    pub(crate) fn new(min_value: u64) -> Self {
-        Self { min_value }
-    }
-}
-
-impl<External: MonotonicallyMappableToU64> StrictlyMonotonicFn<External, u64>
-    for StrictlyMonotonicMappingToInternalBaseval
-{
-    fn mapping(&self, val: External) -> u64 {
-        External::to_u64(val) - self.min_value
-    }
-
-    fn inverse(&self, val: u64) -> External {
-        External::from_u64(self.min_value + val)
-    }
-}
-
-impl MonotonicallyMappableToU64 for u64 {
-    fn to_u64(self) -> u64 {
-        self
-    }
-
-    fn from_u64(val: u64) -> Self {
-        val
-    }
-}
-
-impl MonotonicallyMappableToU64 for i64 {
-    #[inline(always)]
-    fn to_u64(self) -> u64 {
-        common::i64_to_u64(self)
-    }
-
-    #[inline(always)]
-    fn from_u64(val: u64) -> Self {
-        common::u64_to_i64(val)
-    }
-}
-
-impl MonotonicallyMappableToU64 for bool {
-    #[inline(always)]
-    fn to_u64(self) -> u64 {
-        u64::from(self)
-    }
-
-    #[inline(always)]
-    fn from_u64(val: u64) -> Self {
-        val > 0
-    }
-}
-
-// TODO remove me.
-// Tantivy should refuse NaN values and work with NotNaN internally.
-impl MonotonicallyMappableToU64 for f64 {
-    fn to_u64(self) -> u64 {
-        common::f64_to_u64(self)
-    }
-
-    fn from_u64(val: u64) -> Self {
-        common::u64_to_f64(val)
-    }
-}
-
-impl MonotonicallyMappableToU64 for ordered_float::NotNan<f64> {
-    fn to_u64(self) -> u64 {
-        common::f64_to_u64(self.into_inner())
-    }
-
-    fn from_u64(val: u64) -> Self {
-        NotNan::new(common::u64_to_f64(val)).expect("Invalid NotNaN f64 value.")
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use super::*;
-
-    #[test]
-    fn test_from_u64_pos_inf() {
-        let inf_as_u64 = common::f64_to_u64(f64::INFINITY);
-        let inf_back_to_f64 = NotNan::from_u64(inf_as_u64);
-        assert_eq!(inf_back_to_f64, NotNan::new(f64::INFINITY).unwrap());
-    }
-
-    #[test]
-    fn test_from_u64_neg_inf() {
-        let inf_as_u64 = common::f64_to_u64(-f64::INFINITY);
-        let inf_back_to_f64 = NotNan::from_u64(inf_as_u64);
-        assert_eq!(inf_back_to_f64, NotNan::new(-f64::INFINITY).unwrap());
-    }
-
-    #[test]
-    #[should_panic(expected = "Invalid NotNaN")]
-    fn test_from_u64_nan_panics() {
-        let nan_as_u64 = common::f64_to_u64(f64::NAN);
-        NotNan::from_u64(nan_as_u64);
-    }
-
-    #[test]
-    fn strictly_monotonic_test() {
-        // identity mapping
-        test_round_trip(&StrictlyMonotonicMappingToInternal::<u64>::new(), 100u64);
-        // round trip to i64
-        test_round_trip(&StrictlyMonotonicMappingToInternal::<i64>::new(), 100u64);
-        // identity mapping
-        test_round_trip(&StrictlyMonotonicMappingToInternal::<u128>::new(), 100u128);
-
-        // base value to i64 round trip
-        let mapping = StrictlyMonotonicMappingToInternalBaseval::new(100);
-        test_round_trip::<_, _, u64>(&mapping, 100i64);
-        // base value and gcd to u64 round trip
-        let mapping = StrictlyMonotonicMappingToInternalGCDBaseval::new(10, 100);
-        test_round_trip::<_, _, u64>(&mapping, 100u64);
-    }
-
-    fn test_round_trip<T: StrictlyMonotonicFn<K, L>, K: std::fmt::Debug + Eq + Copy, L>(
-        mapping: &T,
-        test_val: K,
-    ) {
-        assert_eq!(mapping.inverse(mapping.mapping(test_val)), test_val);
-    }
-}

fastfield_codecs/src/monotonic_mapping_u128.rs

@@ -1,40 +0,0 @@
-use std::net::Ipv6Addr;
-
-/// Montonic maps a value to u128 value space
-/// Monotonic mapping enables `PartialOrd` on u128 space without conversion to original space.
-pub trait MonotonicallyMappableToU128: 'static + PartialOrd + Copy + Send + Sync {
-    /// Converts a value to u128.
-    ///
-    /// Internally all fast field values are encoded as u64.
-    fn to_u128(self) -> u128;
-
-    /// Converts a value from u128
-    ///
-    /// Internally all fast field values are encoded as u64.
-    /// **Note: To be used for converting encoded Term, Posting values.**
-    fn from_u128(val: u128) -> Self;
-}
-
-impl MonotonicallyMappableToU128 for u128 {
-    fn to_u128(self) -> u128 {
-        self
-    }
-
-    fn from_u128(val: u128) -> Self {
-        val
-    }
-}
-
-impl MonotonicallyMappableToU128 for Ipv6Addr {
-    fn to_u128(self) -> u128 {
-        ip_to_u128(self)
-    }
-
-    fn from_u128(val: u128) -> Self {
-        Ipv6Addr::from(val.to_be_bytes())
-    }
-}
-
-fn ip_to_u128(ip_addr: Ipv6Addr) -> u128 {
-    u128::from_be_bytes(ip_addr.octets())
-}

fastfield_codecs/src/null_index/dense.rs

@@ -1,454 +0,0 @@
-use std::convert::TryInto;
-use std::io::{self, Write};
-
-use common::BinarySerializable;
-use itertools::Itertools;
-use ownedbytes::OwnedBytes;
-
-use super::{get_bit_at, set_bit_at};
-
-/// For the `DenseCodec`, `data` which contains the encoded blocks.
-/// Each block consists of [u8; 12]. The first 8 bytes is a bitvec for 64 elements.
-/// The last 4 bytes are the offset, the number of set bits so far.
-///
-/// When translating the original index to a dense index, the correct block can be computed
-/// directly `orig_idx/64`. Inside the block the position is `orig_idx%64`.
-///
-/// When translating a dense index to the original index, we can use the offset to find the correct
-/// block. Direct computation is not possible, but we can employ a linear or binary search.
-#[derive(Clone)]
-pub struct DenseCodec {
-    // data consists of blocks of 64 bits.
-    //
-    // The format is &[(u64, u32)]
-    // u64 is the bitvec
-    // u32 is the offset of the block, the number of set bits so far.
-    //
-    // At the end one block is appended, to store the number of values in the index in offset.
-    data: OwnedBytes,
-}
-const ELEMENTS_PER_BLOCK: u32 = 64;
-const BLOCK_BITVEC_SIZE: usize = 8;
-const BLOCK_OFFSET_SIZE: usize = 4;
-const SERIALIZED_BLOCK_SIZE: usize = BLOCK_BITVEC_SIZE + BLOCK_OFFSET_SIZE;
-
-#[inline]
-fn count_ones(bitvec: u64, pos_in_bitvec: u32) -> u32 {
-    if pos_in_bitvec == 63 {
-        bitvec.count_ones()
-    } else {
-        let mask = (1u64 << (pos_in_bitvec + 1)) - 1;
-        let masked_bitvec = bitvec & mask;
-        masked_bitvec.count_ones()
-    }
-}
-
-#[derive(Clone, Copy)]
-struct DenseIndexBlock {
-    bitvec: u64,
-    offset: u32,
-}
-
-impl From<[u8; SERIALIZED_BLOCK_SIZE]> for DenseIndexBlock {
-    fn from(data: [u8; SERIALIZED_BLOCK_SIZE]) -> Self {
-        let bitvec = u64::from_le_bytes(data[..BLOCK_BITVEC_SIZE].try_into().unwrap());
-        let offset = u32::from_le_bytes(data[BLOCK_BITVEC_SIZE..].try_into().unwrap());
-        Self { bitvec, offset }
-    }
-}
-
-impl DenseCodec {
-    /// Open the DenseCodec from OwnedBytes
-    pub fn open(data: OwnedBytes) -> Self {
-        Self { data }
-    }
-    #[inline]
-    /// Check if value at position is not null.
-    pub fn exists(&self, idx: u32) -> bool {
-        let block_pos = idx / ELEMENTS_PER_BLOCK;
-        let bitvec = self.dense_index_block(block_pos).bitvec;
-
-        let pos_in_bitvec = idx % ELEMENTS_PER_BLOCK;
-
-        get_bit_at(bitvec, pos_in_bitvec)
-    }
-    #[inline]
-    fn dense_index_block(&self, block_pos: u32) -> DenseIndexBlock {
-        dense_index_block(&self.data, block_pos)
-    }
-
-    /// Return the number of non-null values in an index
-    pub fn num_non_nulls(&self) -> u32 {
-        let last_block = (self.data.len() / SERIALIZED_BLOCK_SIZE) - 1;
-        self.dense_index_block(last_block as u32).offset
-    }
-
-    #[inline]
-    /// Translate from the original index to the codec index.
-    pub fn translate_to_codec_idx(&self, idx: u32) -> Option<u32> {
-        let block_pos = idx / ELEMENTS_PER_BLOCK;
-        let index_block = self.dense_index_block(block_pos);
-        let pos_in_block_bit_vec = idx % ELEMENTS_PER_BLOCK;
-        let ones_in_block = count_ones(index_block.bitvec, pos_in_block_bit_vec);
-        if get_bit_at(index_block.bitvec, pos_in_block_bit_vec) {
-            // -1 is ok, since idx does exist, so there's at least one
-            Some(index_block.offset + ones_in_block - 1)
-        } else {
-            None
-        }
-    }
-
-    /// Translate positions from the codec index to the original index.
-    ///
-    /// # Panics
-    ///
-    /// May panic if any `idx` is greater than the max codec index.
-    pub fn translate_codec_idx_to_original_idx<'a>(
-        &'a self,
-        iter: impl Iterator<Item = u32> + 'a,
-    ) -> impl Iterator<Item = u32> + 'a {
-        let mut block_pos = 0u32;
-        iter.map(move |dense_idx| {
-            // update block_pos to limit search scope
-            block_pos = find_block(dense_idx, block_pos, &self.data);
-            let index_block = self.dense_index_block(block_pos);
-
-            // The next offset is higher than dense_idx and therefore:
-            // dense_idx <= offset + num_set_bits in block
-            let mut num_set_bits = 0;
-            for idx_in_bitvec in 0..ELEMENTS_PER_BLOCK {
-                if get_bit_at(index_block.bitvec, idx_in_bitvec) {
-                    num_set_bits += 1;
-                }
-                if num_set_bits == (dense_idx - index_block.offset + 1) {
-                    let orig_idx = block_pos * ELEMENTS_PER_BLOCK + idx_in_bitvec;
-                    return orig_idx;
-                }
-            }
-            panic!("Internal Error: Offset calculation in dense idx seems to be wrong.");
-        })
-    }
-}
-
-#[inline]
-fn dense_index_block(data: &[u8], block_pos: u32) -> DenseIndexBlock {
-    let data_start_pos = block_pos as usize * SERIALIZED_BLOCK_SIZE;
-    let block_data: [u8; SERIALIZED_BLOCK_SIZE] = data[data_start_pos..][..SERIALIZED_BLOCK_SIZE]
-        .try_into()
-        .unwrap();
-    block_data.into()
-}
-
-#[inline]
-/// Finds the block position containing the dense_idx.
-///
-/// # Correctness
-/// dense_idx needs to be smaller than the number of values in the index
-///
-/// The last offset number is equal to the number of values in the index.
-fn find_block(dense_idx: u32, mut block_pos: u32, data: &[u8]) -> u32 {
-    loop {
-        let offset = dense_index_block(data, block_pos).offset;
-        if offset > dense_idx {
-            return block_pos - 1;
-        }
-        block_pos += 1;
-    }
-}
-
-/// Iterator over all values, true if set, otherwise false
-pub fn serialize_dense_codec(
-    iter: impl Iterator<Item = bool>,
-    mut out: impl Write,
-) -> io::Result<()> {
-    let mut offset: u32 = 0;
-
-    for chunk in &iter.chunks(ELEMENTS_PER_BLOCK as usize) {
-        let mut block: u64 = 0;
-        for (pos, is_bit_set) in chunk.enumerate() {
-            if is_bit_set {
-                set_bit_at(&mut block, pos as u64);
-            }
-        }
-
-        block.serialize(&mut out)?;
-        offset.serialize(&mut out)?;
-
-        offset += block.count_ones();
-    }
-    // Add sentinal block for the offset
-    let block: u64 = 0;
-    block.serialize(&mut out)?;
-    offset.serialize(&mut out)?;
-
-    Ok(())
-}
-
-#[cfg(test)]
-mod tests {
-    use proptest::prelude::{any, prop, *};
-    use proptest::strategy::Strategy;
-    use proptest::{prop_oneof, proptest};
-
-    use super::*;
-
-    fn random_bitvec() -> BoxedStrategy<Vec<bool>> {
-        prop_oneof![
-            1 => prop::collection::vec(proptest::bool::weighted(1.0), 0..100),
-            1 => prop::collection::vec(proptest::bool::weighted(1.0), 0..64),
-            1 => prop::collection::vec(proptest::bool::weighted(0.0), 0..100),
-            1 => prop::collection::vec(proptest::bool::weighted(0.0), 0..64),
-            8 => vec![any::<bool>()],
-            2 => prop::collection::vec(any::<bool>(), 0..50),
-        ]
-        .boxed()
-    }
-
-    proptest! {
-        #![proptest_config(ProptestConfig::with_cases(500))]
-        #[test]
-        fn test_with_random_bitvecs(bitvec1 in random_bitvec(), bitvec2 in random_bitvec(), bitvec3 in random_bitvec()) {
-            let mut bitvec = Vec::new();
-            bitvec.extend_from_slice(&bitvec1);
-            bitvec.extend_from_slice(&bitvec2);
-            bitvec.extend_from_slice(&bitvec3);
-            test_null_index(bitvec);
-        }
-    }
-
-    #[test]
-    fn dense_codec_test_one_block_false() {
-        let mut iter = vec![false; 64];
-        iter.push(true);
-        test_null_index(iter);
-    }
-
-    fn test_null_index(data: Vec<bool>) {
-        let mut out = vec![];
-
-        serialize_dense_codec(data.iter().cloned(), &mut out).unwrap();
-        let null_index = DenseCodec::open(OwnedBytes::new(out));
-
-        let orig_idx_with_value: Vec<u32> = data
-            .iter()
-            .enumerate()
-            .filter(|(_pos, val)| **val)
-            .map(|(pos, _val)| pos as u32)
-            .collect();
-
-        assert_eq!(
-            null_index
-                .translate_codec_idx_to_original_idx(0..orig_idx_with_value.len() as u32)
-                .collect_vec(),
-            orig_idx_with_value
-        );
-
-        for (dense_idx, orig_idx) in orig_idx_with_value.iter().enumerate() {
-            assert_eq!(
-                null_index.translate_to_codec_idx(*orig_idx),
-                Some(dense_idx as u32)
-            );
-        }
-
-        for (pos, value) in data.iter().enumerate() {
-            assert_eq!(null_index.exists(pos as u32), *value);
-        }
-    }
-
-    #[test]
-    fn dense_codec_test_translation() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_dense_codec(iter, &mut out).unwrap();
-        let null_index = DenseCodec::open(OwnedBytes::new(out));
-
-        assert_eq!(
-            null_index
-                .translate_codec_idx_to_original_idx(0..2)
-                .collect_vec(),
-            vec![0, 2]
-        );
-    }
-
-    #[test]
-    fn dense_codec_translate() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_dense_codec(iter, &mut out).unwrap();
-        let null_index = DenseCodec::open(OwnedBytes::new(out));
-        assert_eq!(null_index.translate_to_codec_idx(0), Some(0));
-        assert_eq!(null_index.translate_to_codec_idx(2), Some(1));
-    }
-
-    #[test]
-    fn dense_codec_test_small() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_dense_codec(iter, &mut out).unwrap();
-        let null_index = DenseCodec::open(OwnedBytes::new(out));
-        assert!(null_index.exists(0));
-        assert!(!null_index.exists(1));
-        assert!(null_index.exists(2));
-        assert!(!null_index.exists(3));
-    }
-
-    #[test]
-    fn dense_codec_test_large() {
-        let mut docs = vec![];
-        docs.extend((0..1000).map(|_idx| false));
-        docs.extend((0..=1000).map(|_idx| true));
-
-        let iter = docs.iter().cloned();
-        let mut out = vec![];
-        serialize_dense_codec(iter, &mut out).unwrap();
-        let null_index = DenseCodec::open(OwnedBytes::new(out));
-        assert!(!null_index.exists(0));
-        assert!(!null_index.exists(100));
-        assert!(!null_index.exists(999));
-        assert!(null_index.exists(1000));
-        assert!(null_index.exists(1999));
-        assert!(null_index.exists(2000));
-        assert!(!null_index.exists(2001));
-    }
-
-    #[test]
-    fn test_count_ones() {
-        let mut block = 0;
-        set_bit_at(&mut block, 0);
-        set_bit_at(&mut block, 2);
-
-        assert_eq!(count_ones(block, 0), 1);
-        assert_eq!(count_ones(block, 1), 1);
-        assert_eq!(count_ones(block, 2), 2);
-    }
-}
-
-#[cfg(all(test, feature = "unstable"))]
-mod bench {
-
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
-    use test::Bencher;
-
-    use super::*;
-
-    const TOTAL_NUM_VALUES: u32 = 1_000_000;
-    fn gen_bools(fill_ratio: f64) -> DenseCodec {
-        let mut out = Vec::new();
-        let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
-        let bools: Vec<_> = (0..TOTAL_NUM_VALUES)
-            .map(|_| rng.gen_bool(fill_ratio))
-            .collect();
-        serialize_dense_codec(bools.into_iter(), &mut out).unwrap();
-
-        let codec = DenseCodec::open(OwnedBytes::new(out));
-        codec
-    }
-
-    fn random_range_iterator(start: u32, end: u32, step_size: u32) -> impl Iterator<Item = u32> {
-        let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
-        let mut current = start;
-        std::iter::from_fn(move || {
-            current += rng.gen_range(1..step_size + 1);
-            if current >= end {
-                None
-            } else {
-                Some(current)
-            }
-        })
-    }
-
-    fn walk_over_data(codec: &DenseCodec, max_step_size: u32) -> Option<u32> {
-        walk_over_data_from_positions(
-            codec,
-            random_range_iterator(0, TOTAL_NUM_VALUES, max_step_size),
-        )
-    }
-
-    fn walk_over_data_from_positions(
-        codec: &DenseCodec,
-        positions: impl Iterator<Item = u32>,
-    ) -> Option<u32> {
-        let mut dense_idx: Option<u32> = None;
-        for idx in positions {
-            dense_idx = dense_idx.or(codec.translate_to_codec_idx(idx));
-        }
-        dense_idx
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_orig_to_codec_90percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.9f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_orig_to_codec_50percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.5f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_orig_to_codec_full_scan_10percent(bench: &mut Bencher) {
-        let codec = gen_bools(0.1f64);
-        bench.iter(|| walk_over_data_from_positions(&codec, 0..TOTAL_NUM_VALUES));
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_orig_to_codec_full_scan_90percent(bench: &mut Bencher) {
-        let codec = gen_bools(0.9f64);
-        bench.iter(|| walk_over_data_from_positions(&codec, 0..TOTAL_NUM_VALUES));
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_orig_to_codec_10percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.1f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_codec_to_orig_90percent_filled_random_stride_big_step(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.9f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 50_000))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_codec_to_orig_90percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.9f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 100))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_dense_codec_translate_codec_to_orig_90percent_filled_full_scan(bench: &mut Bencher) {
-        let codec = gen_bools(0.9f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(0..num_vals)
-                .last()
-        });
-    }
-}

fastfield_codecs/src/null_index/mod.rs

@@ -1,14 +0,0 @@
-pub use dense::{serialize_dense_codec, DenseCodec};
-
-mod dense;
-mod sparse;
-
-#[inline]
-fn get_bit_at(input: u64, n: u32) -> bool {
-    input & (1 << n) != 0
-}
-
-#[inline]
-fn set_bit_at(input: &mut u64, n: u64) {
-    *input |= 1 << n;
-}

fastfield_codecs/src/null_index/sparse.rs

@@ -1,752 +0,0 @@
-use std::io::{self, Write};
-
-use common::BitSet;
-use ownedbytes::OwnedBytes;
-
-use super::{serialize_dense_codec, DenseCodec};
-
-/// `SparseCodec` is the codec for data, when only few documents have values.
-/// In contrast to `DenseCodec` opening a `SparseCodec` causes runtime data to be produced, for
-/// faster access.
-///
-/// The lower 16 bits of doc ids are stored as u16 while the upper 16 bits are given by the block
-/// id. Each block contains 1<<16 docids.
-///
-/// # Serialized Data Layout
-/// The data starts with the block data. Each block is either dense or sparse encoded, depending on
-/// the number of values in the block. A block is sparse when it contains less than
-/// DENSE_BLOCK_THRESHOLD (6144) values.
-/// [Sparse data block | dense data block, .. #repeat*; Desc: Either a sparse or dense encoded
-/// block]
-/// ### Sparse block data
-/// [u16 LE, .. #repeat*; Desc: Positions with values in a block]
-/// ### Dense block data
-/// [Dense codec for the whole block; Desc: Similar to a bitvec(0..ELEMENTS_PER_BLOCK) + Metadata
-/// for faster lookups. See dense.rs]
-///
-/// The data is followed by block metadata, to know which area of the raw block data belongs to
-/// which block. Only metadata for blocks with elements is recorded to
-/// keep the overhead low for scenarios with many very sparse columns. The block metadata consists
-/// of the block index and the number of values in the block. Since we don't store empty blocks
-/// num_vals is incremented by 1, e.g. 0 means 1 value.
-///
-/// The last u16 is storing the number of metadata blocks.
-/// [u16 LE, .. #repeat*; Desc: Positions with values in a block][(u16 LE, u16 LE), .. #repeat*;
-/// Desc: (Block Id u16, Num Elements u16)][u16 LE; Desc: num blocks with values u16]
-///
-/// # Opening
-/// When opening the data layout, the data is expanded to `Vec<SparseCodecBlockVariant>`, where the
-/// index is the block index. For each block `byte_start` and `offset` is computed.
-pub struct SparseCodec {
-    data: OwnedBytes,
-    blocks: Vec<SparseCodecBlockVariant>,
-}
-
-/// The threshold for for number of elements after which we switch to dense block encoding
-const DENSE_BLOCK_THRESHOLD: u32 = 6144;
-
-const ELEMENTS_PER_BLOCK: u32 = u16::MAX as u32 + 1;
-
-/// 1.5 bit per Element + 12 bytes for the sentinal block
-const NUM_BYTES_DENSE_BLOCK: u32 = (ELEMENTS_PER_BLOCK + ELEMENTS_PER_BLOCK / 2 + 64 + 32) / 8;
-
-#[derive(Clone)]
-enum SparseCodecBlockVariant {
-    Empty { offset: u32 },
-    Dense(DenseBlock),
-    Sparse(SparseBlock),
-}
-
-impl SparseCodecBlockVariant {
-    /// The number of non-null values that preceeded that block.
-    fn offset(&self) -> u32 {
-        match self {
-            SparseCodecBlockVariant::Empty { offset } => *offset,
-            SparseCodecBlockVariant::Dense(dense) => dense.offset,
-            SparseCodecBlockVariant::Sparse(sparse) => sparse.offset,
-        }
-    }
-}
-
-/// A block consists of max u16 values
-#[derive(Clone)]
-struct DenseBlock {
-    /// The number of values set before the block
-    offset: u32,
-    /// The data for the dense encoding
-    codec: DenseCodec,
-}
-
-impl DenseBlock {
-    pub fn exists(&self, idx: u32) -> bool {
-        self.codec.exists(idx)
-    }
-    pub fn translate_to_codec_idx(&self, idx: u32) -> Option<u32> {
-        self.codec.translate_to_codec_idx(idx)
-    }
-    pub fn translate_codec_idx_to_original_idx(&self, idx: u32) -> u32 {
-        self.codec
-            .translate_codec_idx_to_original_idx(idx..=idx)
-            .next()
-            .unwrap()
-    }
-}
-
-/// A block consists of max u16 values
-#[derive(Debug, Copy, Clone)]
-struct SparseBlock {
-    /// The number of values in the block
-    num_vals: u32,
-    /// The number of values set before the block
-    offset: u32,
-    /// The start position of the data for the block
-    byte_start: u32,
-}
-
-impl SparseBlock {
-    fn empty_block(offset: u32) -> Self {
-        Self {
-            num_vals: 0,
-            byte_start: 0,
-            offset,
-        }
-    }
-
-    #[inline]
-    fn value_at_idx(&self, data: &[u8], idx: u16) -> u16 {
-        let start_offset: usize = self.byte_start as usize + (idx as u32 as usize * 2);
-        get_u16(data, start_offset)
-    }
-
-    #[inline]
-    #[allow(clippy::comparison_chain)]
-    // Looks for the element in the block. Returns the positions if found.
-    fn binary_search(&self, data: &[u8], target: u16) -> Option<u16> {
-        let mut size = self.num_vals as u16;
-        let mut left = 0;
-        let mut right = size;
-        // TODO try different implem.
-        //  e.g. exponential search into binary search
-        while left < right {
-            let mid = left + size / 2;
-
-            // TODO do boundary check only once, and then use an
-            // unsafe `value_at_idx`
-            let mid_val = self.value_at_idx(data, mid);
-
-            if target > mid_val {
-                left = mid + 1;
-            } else if target < mid_val {
-                right = mid;
-            } else {
-                return Some(mid);
-            }
-
-            size = right - left;
-        }
-        None
-    }
-}
-
-#[inline]
-fn get_u16(data: &[u8], byte_position: usize) -> u16 {
-    let bytes: [u8; 2] = data[byte_position..byte_position + 2].try_into().unwrap();
-    u16::from_le_bytes(bytes)
-}
-
-const SERIALIZED_BLOCK_METADATA_SIZE: usize = 4;
-
-fn deserialize_sparse_codec_block(data: &OwnedBytes) -> Vec<SparseCodecBlockVariant> {
-    // The number of vals so far
-    let mut offset = 0;
-    let mut sparse_codec_blocks = Vec::new();
-    let num_blocks = get_u16(data, data.len() - 2);
-    let block_data_index_start =
-        data.len() - 2 - num_blocks as usize * SERIALIZED_BLOCK_METADATA_SIZE;
-    let mut byte_start = 0;
-    for block_num in 0..num_blocks as usize {
-        let block_data_index = block_data_index_start + SERIALIZED_BLOCK_METADATA_SIZE * block_num;
-        let block_idx = get_u16(data, block_data_index);
-        let num_vals = get_u16(data, block_data_index + 2) as u32 + 1;
-        sparse_codec_blocks.resize(
-            block_idx as usize,
-            SparseCodecBlockVariant::Empty { offset },
-        );
-
-        if is_sparse(num_vals) {
-            let block = SparseBlock {
-                num_vals,
-                offset,
-                byte_start,
-            };
-            sparse_codec_blocks.push(SparseCodecBlockVariant::Sparse(block));
-            byte_start += 2 * num_vals;
-        } else {
-            let block = DenseBlock {
-                offset,
-                codec: DenseCodec::open(data.slice(byte_start as usize..data.len()).clone()),
-            };
-            sparse_codec_blocks.push(SparseCodecBlockVariant::Dense(block));
-            // Dense blocks have a fixed size spanning ELEMENTS_PER_BLOCK.
-            byte_start += NUM_BYTES_DENSE_BLOCK;
-        }
-
-        offset += num_vals;
-    }
-    sparse_codec_blocks.push(SparseCodecBlockVariant::Empty { offset });
-    sparse_codec_blocks
-}
-
-/// Splits a value address into lower and upper 16bits.
-/// The lower 16 bits are the value in the block
-/// The upper 16 bits are the block index
-#[derive(Debug, Clone, Copy)]
-struct ValueAddr {
-    block_idx: u16,
-    value_in_block: u16,
-}
-
-/// Splits a idx into block index and value in the block
-fn value_addr(idx: u32) -> ValueAddr {
-    /// Static assert number elements per block this method expects
-    #[allow(clippy::assertions_on_constants)]
-    const _: () = assert!(ELEMENTS_PER_BLOCK == (1 << 16));
-
-    let value_in_block = idx as u16;
-    let block_idx = (idx >> 16) as u16;
-    ValueAddr {
-        block_idx,
-        value_in_block,
-    }
-}
-
-impl SparseCodec {
-    /// Open the SparseCodec from OwnedBytes
-    pub fn open(data: OwnedBytes) -> Self {
-        let blocks = deserialize_sparse_codec_block(&data);
-        Self { data, blocks }
-    }
-
-    #[inline]
-    /// Check if value at position is not null.
-    pub fn exists(&self, idx: u32) -> bool {
-        let value_addr = value_addr(idx);
-        // There may be trailing nulls without data, those are not stored as blocks. It would be
-        // possible to create empty blocks, but for that we would need to serialize the number of
-        // values or pass them when opening
-
-        if let Some(block) = self.blocks.get(value_addr.block_idx as usize) {
-            match block {
-                SparseCodecBlockVariant::Empty { offset: _ } => false,
-                SparseCodecBlockVariant::Dense(block) => {
-                    block.exists(value_addr.value_in_block as u32)
-                }
-                SparseCodecBlockVariant::Sparse(block) => block
-                    .binary_search(&self.data, value_addr.value_in_block)
-                    .is_some(),
-            }
-        } else {
-            false
-        }
-    }
-
-    /// Return the number of non-null values in an index
-    pub fn num_non_nulls(&self) -> u32 {
-        self.blocks.last().map(|block| block.offset()).unwrap_or(0)
-    }
-
-    #[inline]
-    /// Translate from the original index to the codec index.
-    pub fn translate_to_codec_idx(&self, idx: u32) -> Option<u32> {
-        let value_addr = value_addr(idx);
-        let block = self.blocks.get(value_addr.block_idx as usize)?;
-
-        match block {
-            SparseCodecBlockVariant::Empty { offset: _ } => None,
-            SparseCodecBlockVariant::Dense(block) => block
-                .translate_to_codec_idx(value_addr.value_in_block as u32)
-                .map(|pos_in_block| pos_in_block + block.offset),
-            SparseCodecBlockVariant::Sparse(block) => {
-                let pos_in_block = block.binary_search(&self.data, value_addr.value_in_block);
-                pos_in_block.map(|pos_in_block: u16| block.offset + pos_in_block as u32)
-            }
-        }
-    }
-
-    fn find_block(&self, dense_idx: u32, mut block_pos: u32) -> u32 {
-        loop {
-            let offset = self.blocks[block_pos as usize].offset();
-            if offset > dense_idx {
-                return block_pos - 1;
-            }
-            block_pos += 1;
-        }
-    }
-
-    /// Translate positions from the codec index to the original index.
-    ///
-    /// # Panics
-    ///
-    /// May panic if any `idx` is greater than the max codec index.
-    pub fn translate_codec_idx_to_original_idx<'a>(
-        &'a self,
-        iter: impl Iterator<Item = u32> + 'a,
-    ) -> impl Iterator<Item = u32> + 'a {
-        // TODO: There's a big potential performance gain, by using iterators per block instead of
-        // random access for each element in a block
-        // group_by itertools won't help though, since it requires a temporary local variable
-        let mut block_pos = 0u32;
-        iter.map(move |codec_idx| {
-            // update block_pos to limit search scope
-            block_pos = self.find_block(codec_idx, block_pos);
-            let block_doc_idx_start = block_pos * ELEMENTS_PER_BLOCK;
-            let block = &self.blocks[block_pos as usize];
-            let idx_in_block = codec_idx - block.offset();
-            match block {
-                SparseCodecBlockVariant::Empty { offset: _ } => {
-                    panic!(
-                        "invalid input, cannot translate to original index. associated empty \
-                         block with dense idx. block_pos {}, idx_in_block {}",
-                        block_pos, idx_in_block
-                    )
-                }
-                SparseCodecBlockVariant::Dense(dense) => {
-                    dense.translate_codec_idx_to_original_idx(idx_in_block) + block_doc_idx_start
-                }
-                SparseCodecBlockVariant::Sparse(block) => {
-                    block.value_at_idx(&self.data, idx_in_block as u16) as u32 + block_doc_idx_start
-                }
-            }
-        })
-    }
-}
-
-fn is_sparse(num_elem_in_block: u32) -> bool {
-    num_elem_in_block < DENSE_BLOCK_THRESHOLD
-}
-
-#[derive(Default)]
-struct BlockDataSerialized {
-    block_idx: u16,
-    num_vals: u32,
-}
-
-/// Iterator over positions of set values.
-pub fn serialize_sparse_codec<W: Write>(
-    mut iter: impl Iterator<Item = u32>,
-    mut out: W,
-) -> io::Result<()> {
-    let mut block_metadata: Vec<BlockDataSerialized> = Vec::new();
-    let mut current_block = Vec::new();
-    // This if-statement for the first element ensures that
-    // `block_metadata` is not empty in the loop below.
-    if let Some(idx) = iter.next() {
-        let value_addr = value_addr(idx);
-        block_metadata.push(BlockDataSerialized {
-            block_idx: value_addr.block_idx,
-            num_vals: 1,
-        });
-        current_block.push(value_addr.value_in_block);
-    }
-    let flush_block = |current_block: &mut Vec<u16>, out: &mut W| -> io::Result<()> {
-        let is_sparse = is_sparse(current_block.len() as u32);
-        if is_sparse {
-            for val_in_block in current_block.iter() {
-                out.write_all(val_in_block.to_le_bytes().as_ref())?;
-            }
-        } else {
-            let mut bitset = BitSet::with_max_value(ELEMENTS_PER_BLOCK + 1);
-            for val_in_block in current_block.iter() {
-                bitset.insert(*val_in_block as u32);
-            }
-
-            let iter = (0..ELEMENTS_PER_BLOCK).map(|idx| bitset.contains(idx));
-            serialize_dense_codec(iter, out)?;
-        }
-        current_block.clear();
-        Ok(())
-    };
-    for idx in iter {
-        let value_addr = value_addr(idx);
-        if block_metadata[block_metadata.len() - 1].block_idx == value_addr.block_idx {
-            let last_idx_metadata = block_metadata.len() - 1;
-            block_metadata[last_idx_metadata].num_vals += 1;
-        } else {
-            // flush prev block
-            flush_block(&mut current_block, &mut out)?;
-
-            block_metadata.push(BlockDataSerialized {
-                block_idx: value_addr.block_idx,
-                num_vals: 1,
-            });
-        }
-        current_block.push(value_addr.value_in_block);
-    }
-    // handle last block
-    flush_block(&mut current_block, &mut out)?;
-
-    for block in &block_metadata {
-        out.write_all(block.block_idx.to_le_bytes().as_ref())?;
-        // We don't store empty blocks, therefore we can subtract 1.
-        // This way we will be able to use u16 when the number of elements is 1 << 16 or u16::MAX+1
-        out.write_all(((block.num_vals - 1) as u16).to_le_bytes().as_ref())?;
-    }
-    out.write_all((block_metadata.len() as u16).to_le_bytes().as_ref())?;
-
-    Ok(())
-}
-
-#[cfg(test)]
-mod tests {
-    use itertools::Itertools;
-    use proptest::prelude::{any, prop, *};
-    use proptest::strategy::Strategy;
-    use proptest::{prop_oneof, proptest};
-
-    use super::*;
-
-    fn random_bitvec() -> BoxedStrategy<Vec<bool>> {
-        prop_oneof![
-            1 => prop::collection::vec(proptest::bool::weighted(1.0), 0..100),
-            1 => prop::collection::vec(proptest::bool::weighted(0.00), 0..(ELEMENTS_PER_BLOCK as usize * 3)), // empty blocks
-            1 => prop::collection::vec(proptest::bool::weighted(1.00), 0..(ELEMENTS_PER_BLOCK as usize + 10)), // full block
-            1 => prop::collection::vec(proptest::bool::weighted(0.01), 0..100),
-            1 => prop::collection::vec(proptest::bool::weighted(0.01), 0..u16::MAX as usize),
-            8 => vec![any::<bool>()],
-        ]
-        .boxed()
-    }
-
-    proptest! {
-        #![proptest_config(ProptestConfig::with_cases(50))]
-        #[test]
-        fn test_with_random_bitvecs(bitvec1 in random_bitvec(), bitvec2 in random_bitvec(), bitvec3 in random_bitvec()) {
-            let mut bitvec = Vec::new();
-            bitvec.extend_from_slice(&bitvec1);
-            bitvec.extend_from_slice(&bitvec2);
-            bitvec.extend_from_slice(&bitvec3);
-            test_null_index(bitvec);
-        }
-    }
-
-    #[test]
-    fn sparse_codec_test_one_block_false() {
-        let mut iter = vec![false; ELEMENTS_PER_BLOCK as usize];
-        iter.push(true);
-        test_null_index(iter);
-    }
-
-    #[test]
-    fn sparse_codec_test_one_block_true() {
-        let mut iter = vec![true; ELEMENTS_PER_BLOCK as usize];
-        iter.push(true);
-        test_null_index(iter);
-    }
-
-    fn test_null_index(data: Vec<bool>) {
-        let mut out = vec![];
-
-        serialize_sparse_codec(
-            data.iter()
-                .cloned()
-                .enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-        let null_index = SparseCodec::open(OwnedBytes::new(out));
-
-        let orig_idx_with_value: Vec<u32> = data
-            .iter()
-            .enumerate()
-            .filter(|(_pos, val)| **val)
-            .map(|(pos, _val)| pos as u32)
-            .collect();
-
-        assert_eq!(
-            null_index
-                .translate_codec_idx_to_original_idx(0..orig_idx_with_value.len() as u32)
-                .collect_vec(),
-            orig_idx_with_value
-        );
-
-        let step_size = (orig_idx_with_value.len() / 100).max(1);
-        for (dense_idx, orig_idx) in orig_idx_with_value.iter().enumerate().step_by(step_size) {
-            assert_eq!(
-                null_index.translate_to_codec_idx(*orig_idx),
-                Some(dense_idx as u32)
-            );
-        }
-
-        // 100 samples
-        let step_size = (data.len() / 100).max(1);
-        for (pos, value) in data.iter().enumerate().step_by(step_size) {
-            assert_eq!(null_index.exists(pos as u32), *value);
-        }
-    }
-
-    #[test]
-    fn sparse_codec_test_translation() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_sparse_codec(
-            iter.enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-        let null_index = SparseCodec::open(OwnedBytes::new(out));
-
-        assert_eq!(
-            null_index
-                .translate_codec_idx_to_original_idx(0..2)
-                .collect_vec(),
-            vec![0, 2]
-        );
-    }
-
-    #[test]
-    fn sparse_codec_translate() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_sparse_codec(
-            iter.enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-        let null_index = SparseCodec::open(OwnedBytes::new(out));
-        assert_eq!(null_index.translate_to_codec_idx(0), Some(0));
-        assert_eq!(null_index.translate_to_codec_idx(2), Some(1));
-    }
-
-    #[test]
-    fn sparse_codec_test_small() {
-        let mut out = vec![];
-
-        let iter = ([true, false, true, false]).iter().cloned();
-        serialize_sparse_codec(
-            iter.enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-        let null_index = SparseCodec::open(OwnedBytes::new(out));
-        assert!(null_index.exists(0));
-        assert!(!null_index.exists(1));
-        assert!(null_index.exists(2));
-        assert!(!null_index.exists(3));
-    }
-
-    #[test]
-    fn sparse_codec_test_large() {
-        let mut docs = vec![];
-        docs.extend((0..ELEMENTS_PER_BLOCK).map(|_idx| false));
-        docs.extend((0..=1).map(|_idx| true));
-
-        let iter = docs.iter().cloned();
-        let mut out = vec![];
-        serialize_sparse_codec(
-            iter.enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-        let null_index = SparseCodec::open(OwnedBytes::new(out));
-        assert!(!null_index.exists(0));
-        assert!(!null_index.exists(100));
-        assert!(!null_index.exists(ELEMENTS_PER_BLOCK - 1));
-        assert!(null_index.exists(ELEMENTS_PER_BLOCK));
-        assert!(null_index.exists(ELEMENTS_PER_BLOCK + 1));
-    }
-}
-
-#[cfg(all(test, feature = "unstable"))]
-mod bench {
-
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
-    use test::Bencher;
-
-    use super::*;
-
-    const TOTAL_NUM_VALUES: u32 = 1_000_000;
-    fn gen_bools(fill_ratio: f64) -> SparseCodec {
-        let mut out = Vec::new();
-        let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
-        serialize_sparse_codec(
-            (0..TOTAL_NUM_VALUES)
-                .map(|_| rng.gen_bool(fill_ratio))
-                .enumerate()
-                .filter(|(_pos, val)| *val)
-                .map(|(pos, _val)| pos as u32),
-            &mut out,
-        )
-        .unwrap();
-
-        let codec = SparseCodec::open(OwnedBytes::new(out));
-        codec
-    }
-
-    fn random_range_iterator(start: u32, end: u32, step_size: u32) -> impl Iterator<Item = u32> {
-        let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
-        let mut current = start;
-        std::iter::from_fn(move || {
-            current += rng.gen_range(1..step_size + 1);
-            if current >= end {
-                None
-            } else {
-                Some(current)
-            }
-        })
-    }
-
-    fn walk_over_data(codec: &SparseCodec, max_step_size: u32) -> Option<u32> {
-        walk_over_data_from_positions(
-            codec,
-            random_range_iterator(0, TOTAL_NUM_VALUES, max_step_size),
-        )
-    }
-
-    fn walk_over_data_from_positions(
-        codec: &SparseCodec,
-        positions: impl Iterator<Item = u32>,
-    ) -> Option<u32> {
-        let mut dense_idx: Option<u32> = None;
-        for idx in positions {
-            dense_idx = dense_idx.or(codec.translate_to_codec_idx(idx));
-        }
-        dense_idx
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_1percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.01f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_5percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.05f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_full_scan_10percent(bench: &mut Bencher) {
-        let codec = gen_bools(0.1f64);
-        bench.iter(|| walk_over_data_from_positions(&codec, 0..TOTAL_NUM_VALUES));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_full_scan_90percent(bench: &mut Bencher) {
-        let codec = gen_bools(0.9f64);
-        bench.iter(|| walk_over_data_from_positions(&codec, 0..TOTAL_NUM_VALUES));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_full_scan_1percent(bench: &mut Bencher) {
-        let codec = gen_bools(0.01f64);
-        bench.iter(|| walk_over_data_from_positions(&codec, 0..TOTAL_NUM_VALUES));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_10percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.1f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_orig_to_codec_90percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.9f64);
-        bench.iter(|| walk_over_data(&codec, 100));
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_1percent_filled_random_stride_big_step(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.01f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 50_000))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_1percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.01f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 100))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_1percent_filled_full_scan(bench: &mut Bencher) {
-        let codec = gen_bools(0.01f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(0..num_vals)
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_90percent_filled_random_stride_big_step(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.90f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 50_000))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_90percent_filled_random_stride(
-        bench: &mut Bencher,
-    ) {
-        let codec = gen_bools(0.9f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(random_range_iterator(0, num_vals, 100))
-                .last()
-        });
-    }
-
-    #[bench]
-    fn bench_sparse_codec_translate_codec_to_orig_90percent_filled_full_scan(bench: &mut Bencher) {
-        let codec = gen_bools(0.9f64);
-        let num_vals = codec.num_non_nulls();
-        bench.iter(|| {
-            codec
-                .translate_codec_idx_to_original_idx(0..num_vals)
-                .last()
-        });
-    }
-}

fastfield_codecs/src/null_index_footer.rs

@@ -1,146 +0,0 @@
-use std::io::{self, Write};
-use std::ops::Range;
-
-use common::{BinarySerializable, CountingWriter, VInt};
-use ownedbytes::OwnedBytes;
-
-#[derive(Debug, Clone, Copy, Eq, PartialEq)]
-pub(crate) enum FastFieldCardinality {
-    Single = 1,
-    Multi = 2,
-}
-
-impl BinarySerializable for FastFieldCardinality {
-    fn serialize<W: Write>(&self, wrt: &mut W) -> io::Result<()> {
-        self.to_code().serialize(wrt)
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let code = u8::deserialize(reader)?;
-        let codec_type: Self = Self::from_code(code)
-            .ok_or_else(|| io::Error::new(io::ErrorKind::InvalidData, "Unknown code `{code}.`"))?;
-        Ok(codec_type)
-    }
-}
-
-impl FastFieldCardinality {
-    pub(crate) fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub(crate) fn from_code(code: u8) -> Option<Self> {
-        match code {
-            1 => Some(Self::Single),
-            2 => Some(Self::Multi),
-            _ => None,
-        }
-    }
-}
-
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
-pub(crate) enum NullIndexCodec {
-    Full = 1,
-}
-
-impl BinarySerializable for NullIndexCodec {
-    fn serialize<W: Write>(&self, wrt: &mut W) -> io::Result<()> {
-        self.to_code().serialize(wrt)
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let code = u8::deserialize(reader)?;
-        let codec_type: Self = Self::from_code(code)
-            .ok_or_else(|| io::Error::new(io::ErrorKind::InvalidData, "Unknown code `{code}.`"))?;
-        Ok(codec_type)
-    }
-}
-
-impl NullIndexCodec {
-    pub(crate) fn to_code(self) -> u8 {
-        self as u8
-    }
-
-    pub(crate) fn from_code(code: u8) -> Option<Self> {
-        match code {
-            1 => Some(Self::Full),
-            _ => None,
-        }
-    }
-}
-
-#[derive(Debug, Clone, Eq, PartialEq)]
-pub(crate) struct NullIndexFooter {
-    pub(crate) cardinality: FastFieldCardinality,
-    pub(crate) null_index_codec: NullIndexCodec,
-    // Unused for NullIndexCodec::Full
-    pub(crate) null_index_byte_range: Range<u64>,
-}
-
-impl BinarySerializable for NullIndexFooter {
-    fn serialize<W: Write>(&self, writer: &mut W) -> io::Result<()> {
-        self.cardinality.serialize(writer)?;
-        self.null_index_codec.serialize(writer)?;
-        VInt(self.null_index_byte_range.start).serialize(writer)?;
-        VInt(self.null_index_byte_range.end - self.null_index_byte_range.start)
-            .serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let cardinality = FastFieldCardinality::deserialize(reader)?;
-        let null_index_codec = NullIndexCodec::deserialize(reader)?;
-        let null_index_byte_range_start = VInt::deserialize(reader)?.0;
-        let null_index_byte_range_end = VInt::deserialize(reader)?.0 + null_index_byte_range_start;
-        Ok(Self {
-            cardinality,
-            null_index_codec,
-            null_index_byte_range: null_index_byte_range_start..null_index_byte_range_end,
-        })
-    }
-}
-
-pub(crate) fn append_null_index_footer(
-    output: &mut impl io::Write,
-    null_index_footer: NullIndexFooter,
-) -> io::Result<()> {
-    let mut counting_write = CountingWriter::wrap(output);
-    null_index_footer.serialize(&mut counting_write)?;
-    let footer_payload_len = counting_write.written_bytes();
-    BinarySerializable::serialize(&(footer_payload_len as u16), &mut counting_write)?;
-
-    Ok(())
-}
-
-pub(crate) fn read_null_index_footer(
-    data: OwnedBytes,
-) -> io::Result<(OwnedBytes, NullIndexFooter)> {
-    let (data, null_footer_length_bytes) = data.rsplit(2);
-
-    let footer_length = u16::deserialize(&mut null_footer_length_bytes.as_slice())?;
-    let (data, null_index_footer_bytes) = data.rsplit(footer_length as usize);
-    let null_index_footer = NullIndexFooter::deserialize(&mut null_index_footer_bytes.as_ref())?;
-
-    Ok((data, null_index_footer))
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn null_index_footer_deser_test() {
-        let null_index_footer = NullIndexFooter {
-            cardinality: FastFieldCardinality::Single,
-            null_index_codec: NullIndexCodec::Full,
-            null_index_byte_range: 100..120,
-        };
-
-        let mut out = vec![];
-        null_index_footer.serialize(&mut out).unwrap();
-
-        assert_eq!(
-            null_index_footer,
-            NullIndexFooter::deserialize(&mut &out[..]).unwrap()
-        );
-    }
-}

fastfield_codecs/src/serialize.rs

@@ -1,408 +0,0 @@
-use std::io;
-use std::num::NonZeroU64;
-use std::sync::Arc;
-
-use common::{BinarySerializable, VInt};
-use log::warn;
-use ownedbytes::OwnedBytes;
-
-use crate::bitpacked::BitpackedCodec;
-use crate::blockwise_linear::BlockwiseLinearCodec;
-use crate::compact_space::CompactSpaceCompressor;
-use crate::format_version::append_format_version;
-use crate::linear::LinearCodec;
-use crate::monotonic_mapping::{
-    StrictlyMonotonicFn, StrictlyMonotonicMappingToInternal,
-    StrictlyMonotonicMappingToInternalGCDBaseval,
-};
-use crate::null_index_footer::{
-    append_null_index_footer, FastFieldCardinality, NullIndexCodec, NullIndexFooter,
-};
-use crate::{
-    monotonic_map_column, Column, FastFieldCodec, FastFieldCodecType, MonotonicallyMappableToU64,
-    U128FastFieldCodecType, VecColumn, ALL_CODEC_TYPES,
-};
-
-/// The normalized header gives some parameters after applying the following
-/// normalization of the vector:
-/// `val -> (val - min_value) / gcd`
-///
-/// By design, after normalization, `min_value = 0` and `gcd = 1`.
-#[derive(Debug, Copy, Clone)]
-pub struct NormalizedHeader {
-    /// The number of values in the underlying column.
-    pub num_vals: u32,
-    /// The max value of the underlying column.
-    pub max_value: u64,
-}
-
-#[derive(Debug, Copy, Clone)]
-pub(crate) struct Header {
-    pub num_vals: u32,
-    pub min_value: u64,
-    pub max_value: u64,
-    pub gcd: Option<NonZeroU64>,
-    pub codec_type: FastFieldCodecType,
-}
-
-impl Header {
-    pub fn normalized(self) -> NormalizedHeader {
-        let gcd = self.gcd.map(|gcd| gcd.get()).unwrap_or(1);
-        let gcd_min_val_mapping =
-            StrictlyMonotonicMappingToInternalGCDBaseval::new(gcd, self.min_value);
-
-        let max_value = gcd_min_val_mapping.mapping(self.max_value);
-        NormalizedHeader {
-            num_vals: self.num_vals,
-            max_value,
-        }
-    }
-
-    pub fn normalize_column<C: Column>(&self, from_column: C) -> impl Column {
-        normalize_column(from_column, self.min_value, self.gcd)
-    }
-
-    pub fn compute_header(
-        column: impl Column<u64>,
-        codecs: &[FastFieldCodecType],
-    ) -> Option<Header> {
-        let num_vals = column.num_vals();
-        let min_value = column.min_value();
-        let max_value = column.max_value();
-        let gcd = crate::gcd::find_gcd(column.iter().map(|val| val - min_value))
-            .filter(|gcd| gcd.get() > 1u64);
-        let normalized_column = normalize_column(column, min_value, gcd);
-        let codec_type = detect_codec(normalized_column, codecs)?;
-        Some(Header {
-            num_vals,
-            min_value,
-            max_value,
-            gcd,
-            codec_type,
-        })
-    }
-}
-
-#[derive(Debug, Copy, Clone, PartialEq, Eq)]
-pub(crate) struct U128Header {
-    pub num_vals: u32,
-    pub codec_type: U128FastFieldCodecType,
-}
-
-impl BinarySerializable for U128Header {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        VInt(self.num_vals as u64).serialize(writer)?;
-        self.codec_type.serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let num_vals = VInt::deserialize(reader)?.0 as u32;
-        let codec_type = U128FastFieldCodecType::deserialize(reader)?;
-        Ok(U128Header {
-            num_vals,
-            codec_type,
-        })
-    }
-}
-
-pub fn normalize_column<C: Column>(
-    from_column: C,
-    min_value: u64,
-    gcd: Option<NonZeroU64>,
-) -> impl Column {
-    let gcd = gcd.map(|gcd| gcd.get()).unwrap_or(1);
-    let mapping = StrictlyMonotonicMappingToInternalGCDBaseval::new(gcd, min_value);
-    monotonic_map_column(from_column, mapping)
-}
-
-impl BinarySerializable for Header {
-    fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
-        VInt(self.num_vals as u64).serialize(writer)?;
-        VInt(self.min_value).serialize(writer)?;
-        VInt(self.max_value - self.min_value).serialize(writer)?;
-        if let Some(gcd) = self.gcd {
-            VInt(gcd.get()).serialize(writer)?;
-        } else {
-            VInt(0u64).serialize(writer)?;
-        }
-        self.codec_type.serialize(writer)?;
-        Ok(())
-    }
-
-    fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
-        let num_vals = VInt::deserialize(reader)?.0 as u32;
-        let min_value = VInt::deserialize(reader)?.0;
-        let amplitude = VInt::deserialize(reader)?.0;
-        let max_value = min_value + amplitude;
-        let gcd_u64 = VInt::deserialize(reader)?.0;
-        let codec_type = FastFieldCodecType::deserialize(reader)?;
-        Ok(Header {
-            num_vals,
-            min_value,
-            max_value,
-            gcd: NonZeroU64::new(gcd_u64),
-            codec_type,
-        })
-    }
-}
-
-/// Return estimated compression for given codec in the value range [0.0..1.0], where 1.0 means no
-/// compression.
-pub fn estimate<T: MonotonicallyMappableToU64>(
-    typed_column: impl Column<T>,
-    codec_type: FastFieldCodecType,
-) -> Option<f32> {
-    let column = monotonic_map_column(typed_column, StrictlyMonotonicMappingToInternal::<T>::new());
-    let min_value = column.min_value();
-    let gcd = crate::gcd::find_gcd(column.iter().map(|val| val - min_value))
-        .filter(|gcd| gcd.get() > 1u64);
-    let mapping = StrictlyMonotonicMappingToInternalGCDBaseval::new(
-        gcd.map(|gcd| gcd.get()).unwrap_or(1u64),
-        min_value,
-    );
-    let normalized_column = monotonic_map_column(&column, mapping);
-    match codec_type {
-        FastFieldCodecType::Bitpacked => BitpackedCodec::estimate(&normalized_column),
-        FastFieldCodecType::Linear => LinearCodec::estimate(&normalized_column),
-        FastFieldCodecType::BlockwiseLinear => BlockwiseLinearCodec::estimate(&normalized_column),
-    }
-}
-
-/// Serializes u128 values with the compact space codec.
-pub fn serialize_u128<F: Fn() -> I, I: Iterator<Item = u128>>(
-    iter_gen: F,
-    num_vals: u32,
-    output: &mut impl io::Write,
-) -> io::Result<()> {
-    serialize_u128_new(ValueIndexInfo::default(), iter_gen, num_vals, output)
-}
-
-#[allow(dead_code)]
-pub enum ValueIndexInfo<'a> {
-    MultiValue(Box<dyn MultiValueIndexInfo + 'a>),
-    SingleValue(Box<dyn SingleValueIndexInfo + 'a>),
-}
-
-impl Default for ValueIndexInfo<'static> {
-    fn default() -> Self {
-        struct Dummy {}
-        impl SingleValueIndexInfo for Dummy {
-            fn num_vals(&self) -> u32 {
-                todo!()
-            }
-            fn num_non_nulls(&self) -> u32 {
-                todo!()
-            }
-            fn iter(&self) -> Box<dyn Iterator<Item = u32>> {
-                todo!()
-            }
-        }
-
-        Self::SingleValue(Box::new(Dummy {}))
-    }
-}
-
-impl<'a> ValueIndexInfo<'a> {
-    fn get_cardinality(&self) -> FastFieldCardinality {
-        match self {
-            ValueIndexInfo::MultiValue(_) => FastFieldCardinality::Multi,
-            ValueIndexInfo::SingleValue(_) => FastFieldCardinality::Single,
-        }
-    }
-}
-
-pub trait MultiValueIndexInfo {
-    /// The number of docs in the column.
-    fn num_docs(&self) -> u32;
-    /// The number of values in the column.
-    fn num_vals(&self) -> u32;
-    /// Return the start index of the values for each doc
-    fn iter(&self) -> Box<dyn Iterator<Item = u32> + '_>;
-}
-
-pub trait SingleValueIndexInfo {
-    /// The number of values including nulls in the column.
-    fn num_vals(&self) -> u32;
-    /// The number of non-null values in the column.
-    fn num_non_nulls(&self) -> u32;
-    /// Return a iterator of the positions of docs with a value
-    fn iter(&self) -> Box<dyn Iterator<Item = u32> + '_>;
-}
-
-/// Serializes u128 values with the compact space codec.
-pub fn serialize_u128_new<F: Fn() -> I, I: Iterator<Item = u128>>(
-    value_index: ValueIndexInfo,
-    iter_gen: F,
-    num_vals: u32,
-    output: &mut impl io::Write,
-) -> io::Result<()> {
-    let header = U128Header {
-        num_vals,
-        codec_type: U128FastFieldCodecType::CompactSpace,
-    };
-    header.serialize(output)?;
-    let compressor = CompactSpaceCompressor::train_from(iter_gen(), num_vals);
-    compressor.compress_into(iter_gen(), output).unwrap();
-
-    let null_index_footer = NullIndexFooter {
-        cardinality: value_index.get_cardinality(),
-        null_index_codec: NullIndexCodec::Full,
-        null_index_byte_range: 0..0,
-    };
-    append_null_index_footer(output, null_index_footer)?;
-    append_format_version(output)?;
-
-    Ok(())
-}
-
-/// Serializes the column with the codec with the best estimate on the data.
-pub fn serialize<T: MonotonicallyMappableToU64>(
-    typed_column: impl Column<T>,
-    output: &mut impl io::Write,
-    codecs: &[FastFieldCodecType],
-) -> io::Result<()> {
-    serialize_new(ValueIndexInfo::default(), typed_column, output, codecs)
-}
-
-/// Serializes the column with the codec with the best estimate on the data.
-pub fn serialize_new<T: MonotonicallyMappableToU64>(
-    value_index: ValueIndexInfo,
-    typed_column: impl Column<T>,
-    output: &mut impl io::Write,
-    codecs: &[FastFieldCodecType],
-) -> io::Result<()> {
-    let column = monotonic_map_column(typed_column, StrictlyMonotonicMappingToInternal::<T>::new());
-    let header = Header::compute_header(&column, codecs).ok_or_else(|| {
-        io::Error::new(
-            io::ErrorKind::InvalidInput,
-            format!(
-                "Data cannot be serialized with this list of codec. {:?}",
-                codecs
-            ),
-        )
-    })?;
-    header.serialize(output)?;
-    let normalized_column = header.normalize_column(column);
-    assert_eq!(normalized_column.min_value(), 0u64);
-    serialize_given_codec(normalized_column, header.codec_type, output)?;
-
-    let null_index_footer = NullIndexFooter {
-        cardinality: value_index.get_cardinality(),
-        null_index_codec: NullIndexCodec::Full,
-        null_index_byte_range: 0..0,
-    };
-    append_null_index_footer(output, null_index_footer)?;
-    append_format_version(output)?;
-
-    Ok(())
-}
-
-fn detect_codec(
-    column: impl Column<u64>,
-    codecs: &[FastFieldCodecType],
-) -> Option<FastFieldCodecType> {
-    let mut estimations = Vec::new();
-    for &codec in codecs {
-        let estimation_opt = match codec {
-            FastFieldCodecType::Bitpacked => BitpackedCodec::estimate(&column),
-            FastFieldCodecType::Linear => LinearCodec::estimate(&column),
-            FastFieldCodecType::BlockwiseLinear => BlockwiseLinearCodec::estimate(&column),
-        };
-        if let Some(estimation) = estimation_opt {
-            estimations.push((estimation, codec));
-        }
-    }
-    if let Some(broken_estimation) = estimations.iter().find(|estimation| estimation.0.is_nan()) {
-        warn!(
-            "broken estimation for fast field codec {:?}",
-            broken_estimation.1
-        );
-    }
-    // removing nan values for codecs with broken calculations, and max values which disables
-    // codecs
-    estimations.retain(|estimation| !estimation.0.is_nan() && estimation.0 != f32::MAX);
-    estimations.sort_by(|(score_left, _), (score_right, _)| score_left.total_cmp(score_right));
-    Some(estimations.first()?.1)
-}
-
-fn serialize_given_codec(
-    column: impl Column<u64>,
-    codec_type: FastFieldCodecType,
-    output: &mut impl io::Write,
-) -> io::Result<()> {
-    match codec_type {
-        FastFieldCodecType::Bitpacked => {
-            BitpackedCodec::serialize(&column, output)?;
-        }
-        FastFieldCodecType::Linear => {
-            LinearCodec::serialize(&column, output)?;
-        }
-        FastFieldCodecType::BlockwiseLinear => {
-            BlockwiseLinearCodec::serialize(&column, output)?;
-        }
-    }
-    output.flush()?;
-    Ok(())
-}
-
-/// Helper function to serialize a column (autodetect from all codecs) and then open it
-pub fn serialize_and_load<T: MonotonicallyMappableToU64 + Ord + Default>(
-    column: &[T],
-) -> Arc<dyn Column<T>> {
-    let mut buffer = Vec::new();
-    super::serialize(VecColumn::from(&column), &mut buffer, &ALL_CODEC_TYPES).unwrap();
-    super::open(OwnedBytes::new(buffer)).unwrap()
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_serialize_deserialize_u128_header() {
-        let original = U128Header {
-            num_vals: 11,
-            codec_type: U128FastFieldCodecType::CompactSpace,
-        };
-        let mut out = Vec::new();
-        original.serialize(&mut out).unwrap();
-        let restored = U128Header::deserialize(&mut &out[..]).unwrap();
-        assert_eq!(restored, original);
-    }
-
-    #[test]
-    fn test_serialize_deserialize() {
-        let original = [1u64, 5u64, 10u64];
-        let restored: Vec<u64> = serialize_and_load(&original[..]).iter().collect();
-        assert_eq!(&restored, &original[..]);
-    }
-
-    #[test]
-    fn test_fastfield_bool_size_bitwidth_1() {
-        let mut buffer = Vec::new();
-        let col = VecColumn::from(&[false, true][..]);
-        serialize(col, &mut buffer, &ALL_CODEC_TYPES).unwrap();
-        // 5 bytes of header, 1 byte of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 3 + 5 + 8 + 4 + 2);
-    }
-
-    #[test]
-    fn test_fastfield_bool_bit_size_bitwidth_0() {
-        let mut buffer = Vec::new();
-        let col = VecColumn::from(&[true][..]);
-        serialize(col, &mut buffer, &ALL_CODEC_TYPES).unwrap();
-        // 5 bytes of header, 0 bytes of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 3 + 5 + 7 + 4 + 2);
-    }
-
-    #[test]
-    fn test_fastfield_gcd() {
-        let mut buffer = Vec::new();
-        let vals: Vec<u64> = (0..80).map(|val| (val % 7) * 1_000u64).collect();
-        let col = VecColumn::from(&vals[..]);
-        serialize(col, &mut buffer, &[FastFieldCodecType::Bitpacked]).unwrap();
-        // Values are stored over 3 bits.
-        assert_eq!(buffer.len(), 3 + 7 + (3 * 80 / 8) + 7 + 4 + 2);
-    }
-}

ownedbytes/Cargo.toml

@@ -1,15 +0,0 @@
-[package]
-authors = ["Paul Masurel <paul@quickwit.io>", "Pascal Seitz <pascal@quickwit.io>"]
-name = "ownedbytes"
-version = "0.5.0"
-edition = "2021"
-description = "Expose data as static slice"
-license = "MIT"
-documentation = "https://docs.rs/ownedbytes/"
-homepage = "https://github.com/quickwit-oss/tantivy"
-repository = "https://github.com/quickwit-oss/tantivy"
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-stable_deref_trait = "1.2.0"

ownedbytes/src/lib.rs

@@ -1,358 +0,0 @@
-use std::convert::TryInto;
-use std::ops::{Deref, Range};
-use std::sync::Arc;
-use std::{fmt, io, mem};
-
-pub use stable_deref_trait::StableDeref;
-
-/// An OwnedBytes simply wraps an object that owns a slice of data and exposes
-/// this data as a slice.
-///
-/// The backing object is required to be `StableDeref`.
-#[derive(Clone)]
-pub struct OwnedBytes {
-    data: &'static [u8],
-    box_stable_deref: Arc<dyn Deref<Target = [u8]> + Sync + Send>,
-}
-
-impl OwnedBytes {
-    /// Creates an empty `OwnedBytes`.
-    pub fn empty() -> OwnedBytes {
-        OwnedBytes::new(&[][..])
-    }
-
-    /// Creates an `OwnedBytes` instance given a `StableDeref` object.
-    pub fn new<T: StableDeref + Deref<Target = [u8]> + 'static + Send + Sync>(
-        data_holder: T,
-    ) -> OwnedBytes {
-        let box_stable_deref = Arc::new(data_holder);
-        let bytes: &[u8] = box_stable_deref.as_ref();
-        let data = unsafe { mem::transmute::<_, &'static [u8]>(bytes.deref()) };
-        OwnedBytes {
-            data,
-            box_stable_deref,
-        }
-    }
-
-    /// creates a fileslice that is just a view over a slice of the data.
-    #[must_use]
-    #[inline]
-    pub fn slice(&self, range: Range<usize>) -> Self {
-        OwnedBytes {
-            data: &self.data[range],
-            box_stable_deref: self.box_stable_deref.clone(),
-        }
-    }
-
-    /// Returns the underlying slice of data.
-    /// `Deref` and `AsRef` are also available.
-    #[inline]
-    pub fn as_slice(&self) -> &[u8] {
-        self.data
-    }
-
-    /// Returns the len of the slice.
-    #[inline]
-    pub fn len(&self) -> usize {
-        self.data.len()
-    }
-
-    /// Splits the OwnedBytes into two OwnedBytes `(left, right)`.
-    ///
-    /// Left will hold `split_len` bytes.
-    ///
-    /// This operation is cheap and does not require to copy any memory.
-    /// On the other hand, both `left` and `right` retain a handle over
-    /// the entire slice of memory. In other words, the memory will only
-    /// be released when both left and right are dropped.
-    #[inline]
-    #[must_use]
-    pub fn split(self, split_len: usize) -> (OwnedBytes, OwnedBytes) {
-        let right_box_stable_deref = self.box_stable_deref.clone();
-        let left = OwnedBytes {
-            data: &self.data[..split_len],
-            box_stable_deref: self.box_stable_deref,
-        };
-        let right = OwnedBytes {
-            data: &self.data[split_len..],
-            box_stable_deref: right_box_stable_deref,
-        };
-        (left, right)
-    }
-
-    /// Splits the OwnedBytes into two OwnedBytes `(left, right)`.
-    ///
-    /// Right will hold `split_len` bytes.
-    ///
-    /// This operation is cheap and does not require to copy any memory.
-    /// On the other hand, both `left` and `right` retain a handle over
-    /// the entire slice of memory. In other words, the memory will only
-    /// be released when both left and right are dropped.
-    #[inline]
-    #[must_use]
-    pub fn rsplit(self, split_len: usize) -> (OwnedBytes, OwnedBytes) {
-        let data_len = self.data.len();
-        self.split(data_len - split_len)
-    }
-
-    /// Splits the right part of the `OwnedBytes` at the given offset.
-    ///
-    /// `self` is truncated to `split_len`, left with the remaining bytes.
-    pub fn split_off(&mut self, split_len: usize) -> OwnedBytes {
-        let right_box_stable_deref = self.box_stable_deref.clone();
-        let right_piece = OwnedBytes {
-            data: &self.data[split_len..],
-            box_stable_deref: right_box_stable_deref,
-        };
-        self.data = &self.data[..split_len];
-        right_piece
-    }
-
-    /// Returns true iff this `OwnedBytes` is empty.
-    #[inline]
-    pub fn is_empty(&self) -> bool {
-        self.as_slice().is_empty()
-    }
-
-    /// Drops the left most `advance_len` bytes.
-    #[inline]
-    pub fn advance(&mut self, advance_len: usize) {
-        self.data = &self.data[advance_len..]
-    }
-
-    /// Reads an `u8` from the `OwnedBytes` and advance by one byte.
-    #[inline]
-    pub fn read_u8(&mut self) -> u8 {
-        assert!(!self.is_empty());
-
-        let byte = self.as_slice()[0];
-        self.advance(1);
-        byte
-    }
-
-    /// Reads an `u64` encoded as little-endian from the `OwnedBytes` and advance by 8 bytes.
-    #[inline]
-    pub fn read_u64(&mut self) -> u64 {
-        assert!(self.len() > 7);
-
-        let octlet: [u8; 8] = self.as_slice()[..8].try_into().unwrap();
-        self.advance(8);
-        u64::from_le_bytes(octlet)
-    }
-}
-
-impl fmt::Debug for OwnedBytes {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        // We truncate the bytes in order to make sure the debug string
-        // is not too long.
-        let bytes_truncated: &[u8] = if self.len() > 8 {
-            &self.as_slice()[..10]
-        } else {
-            self.as_slice()
-        };
-        write!(f, "OwnedBytes({:?}, len={})", bytes_truncated, self.len())
-    }
-}
-
-impl PartialEq for OwnedBytes {
-    fn eq(&self, other: &OwnedBytes) -> bool {
-        self.as_slice() == other.as_slice()
-    }
-}
-
-impl Eq for OwnedBytes {}
-
-impl PartialEq<[u8]> for OwnedBytes {
-    fn eq(&self, other: &[u8]) -> bool {
-        self.as_slice() == other
-    }
-}
-
-impl PartialEq<str> for OwnedBytes {
-    fn eq(&self, other: &str) -> bool {
-        self.as_slice() == other.as_bytes()
-    }
-}
-
-impl<'a, T: ?Sized> PartialEq<&'a T> for OwnedBytes
-where OwnedBytes: PartialEq<T>
-{
-    fn eq(&self, other: &&'a T) -> bool {
-        *self == **other
-    }
-}
-
-impl Deref for OwnedBytes {
-    type Target = [u8];
-
-    #[inline]
-    fn deref(&self) -> &Self::Target {
-        self.as_slice()
-    }
-}
-
-impl io::Read for OwnedBytes {
-    #[inline]
-    fn read(&mut self, buf: &mut [u8]) -> io::Result<usize> {
-        let read_len = {
-            let data = self.as_slice();
-            if data.len() >= buf.len() {
-                let buf_len = buf.len();
-                buf.copy_from_slice(&data[..buf_len]);
-                buf.len()
-            } else {
-                let data_len = data.len();
-                buf[..data_len].copy_from_slice(data);
-                data_len
-            }
-        };
-        self.advance(read_len);
-        Ok(read_len)
-    }
-    #[inline]
-    fn read_to_end(&mut self, buf: &mut Vec<u8>) -> io::Result<usize> {
-        let read_len = {
-            let data = self.as_slice();
-            buf.extend(data);
-            data.len()
-        };
-        self.advance(read_len);
-        Ok(read_len)
-    }
-    #[inline]
-    fn read_exact(&mut self, buf: &mut [u8]) -> io::Result<()> {
-        let read_len = self.read(buf)?;
-        if read_len != buf.len() {
-            return Err(io::Error::new(
-                io::ErrorKind::UnexpectedEof,
-                "failed to fill whole buffer",
-            ));
-        }
-        Ok(())
-    }
-}
-
-impl AsRef<[u8]> for OwnedBytes {
-    #[inline]
-    fn as_ref(&self) -> &[u8] {
-        self.as_slice()
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use std::io::{self, Read};
-
-    use super::OwnedBytes;
-
-    #[test]
-    fn test_owned_bytes_debug() {
-        let short_bytes = OwnedBytes::new(b"abcd".as_ref());
-        assert_eq!(
-            format!("{:?}", short_bytes),
-            "OwnedBytes([97, 98, 99, 100], len=4)"
-        );
-        let long_bytes = OwnedBytes::new(b"abcdefghijklmnopq".as_ref());
-        assert_eq!(
-            format!("{:?}", long_bytes),
-            "OwnedBytes([97, 98, 99, 100, 101, 102, 103, 104, 105, 106], len=17)"
-        );
-    }
-
-    #[test]
-    fn test_owned_bytes_read() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"abcdefghiklmnopqrstuvwxyz".as_ref());
-        {
-            let mut buf = [0u8; 5];
-            bytes.read_exact(&mut buf[..]).unwrap();
-            assert_eq!(&buf, b"abcde");
-            assert_eq!(bytes.as_slice(), b"fghiklmnopqrstuvwxyz")
-        }
-        {
-            let mut buf = [0u8; 2];
-            bytes.read_exact(&mut buf[..]).unwrap();
-            assert_eq!(&buf, b"fg");
-            assert_eq!(bytes.as_slice(), b"hiklmnopqrstuvwxyz")
-        }
-        Ok(())
-    }
-
-    #[test]
-    fn test_owned_bytes_read_right_at_the_end() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"abcde".as_ref());
-        let mut buf = [0u8; 5];
-        assert_eq!(bytes.read(&mut buf[..]).unwrap(), 5);
-        assert_eq!(&buf, b"abcde");
-        assert_eq!(bytes.as_slice(), b"");
-        assert_eq!(bytes.read(&mut buf[..]).unwrap(), 0);
-        assert_eq!(&buf, b"abcde");
-        Ok(())
-    }
-    #[test]
-    fn test_owned_bytes_read_incomplete() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"abcde".as_ref());
-        let mut buf = [0u8; 7];
-        assert_eq!(bytes.read(&mut buf[..]).unwrap(), 5);
-        assert_eq!(&buf[..5], b"abcde");
-        assert_eq!(bytes.read(&mut buf[..]).unwrap(), 0);
-        Ok(())
-    }
-
-    #[test]
-    fn test_owned_bytes_read_to_end() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"abcde".as_ref());
-        let mut buf = Vec::new();
-        bytes.read_to_end(&mut buf)?;
-        assert_eq!(buf.as_slice(), b"abcde".as_ref());
-        Ok(())
-    }
-
-    #[test]
-    fn test_owned_bytes_read_u8() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"\xFF".as_ref());
-        assert_eq!(bytes.read_u8(), 255);
-        assert_eq!(bytes.len(), 0);
-        Ok(())
-    }
-
-    #[test]
-    fn test_owned_bytes_read_u64() -> io::Result<()> {
-        let mut bytes = OwnedBytes::new(b"\0\xFF\xFF\xFF\xFF\xFF\xFF\xFF".as_ref());
-        assert_eq!(bytes.read_u64(), u64::MAX - 255);
-        assert_eq!(bytes.len(), 0);
-        Ok(())
-    }
-
-    #[test]
-    fn test_owned_bytes_split() {
-        let bytes = OwnedBytes::new(b"abcdefghi".as_ref());
-        let (left, right) = bytes.split(3);
-        assert_eq!(left.as_slice(), b"abc");
-        assert_eq!(right.as_slice(), b"defghi");
-    }
-
-    #[test]
-    fn test_owned_bytes_split_boundary() {
-        let bytes = OwnedBytes::new(b"abcdefghi".as_ref());
-        {
-            let (left, right) = bytes.clone().split(0);
-            assert_eq!(left.as_slice(), b"");
-            assert_eq!(right.as_slice(), b"abcdefghi");
-        }
-        {
-            let (left, right) = bytes.split(9);
-            assert_eq!(left.as_slice(), b"abcdefghi");
-            assert_eq!(right.as_slice(), b"");
-        }
-    }
-
-    #[test]
-    fn test_split_off() {
-        let mut data = OwnedBytes::new(b"abcdef".as_ref());
-        assert_eq!(data, "abcdef");
-        assert_eq!(data.split_off(2), "cdef");
-        assert_eq!(data, "ab");
-        assert_eq!(data.split_off(1), "b");
-        assert_eq!(data, "a");
-    }
-}

query-grammar/Cargo.toml

@@ -1,17 +0,0 @@
-[package]
-name = "tantivy-query-grammar"
-version = "0.19.0"
-authors = ["Paul Masurel <paul.masurel@gmail.com>"]
-license = "MIT"
-categories = ["database-implementations", "data-structures"]
-description = """Search engine library"""
-homepage = "https://github.com/quickwit-oss/tantivy"
-repository = "https://github.com/quickwit-oss/tantivy"
-readme = "README.md"
-keywords = ["search", "information", "retrieval"]
-edition = "2021"
-
-[dependencies]
-combine = {version="4", default-features=false, features=[] }
-once_cell = "1.7.2"
-regex ={ version = "1.5.4", default-features = false, features = ["std", "unicode"] }

query-grammar/README.md

@@ -1,3 +0,0 @@
-# Tantivy Query Grammar
-
-This crate is used by tantivy to parse queries.

query-grammar/src/lib.rs

@@ -1,17 +0,0 @@
-#![allow(clippy::derive_partial_eq_without_eq)]
-
-mod occur;
-mod query_grammar;
-mod user_input_ast;
-use combine::parser::Parser;
-
-pub use crate::occur::Occur;
-use crate::query_grammar::parse_to_ast;
-pub use crate::user_input_ast::{UserInputAst, UserInputBound, UserInputLeaf, UserInputLiteral};
-
-pub struct Error;
-
-pub fn parse_query(query: &str) -> Result<UserInputAst, Error> {
-    let (user_input_ast, _remaining) = parse_to_ast().parse(query).map_err(|_| Error)?;
-    Ok(user_input_ast)
-}

query-grammar/src/occur.rs

@@ -1,72 +0,0 @@
-use std::fmt;
-use std::fmt::Write;
-
-/// Defines whether a term in a query must be present,
-/// should be present or must not be present.
-#[derive(Debug, Clone, Hash, Copy, Eq, PartialEq)]
-pub enum Occur {
-    /// For a given document to be considered for scoring,
-    /// at least one of the terms with the Should or the Must
-    /// Occur constraint must be within the document.
-    Should,
-    /// Document without the term are excluded from the search.
-    Must,
-    /// Document that contain the term are excluded from the
-    /// search.
-    MustNot,
-}
-
-impl Occur {
-    /// Returns the one-char prefix symbol for this `Occur`.
-    /// - `Should` => '?',
-    /// - `Must` => '+'
-    /// - `Not` => '-'
-    fn to_char(self) -> char {
-        match self {
-            Occur::Should => '?',
-            Occur::Must => '+',
-            Occur::MustNot => '-',
-        }
-    }
-
-    /// Compose two occur values.
-    pub fn compose(left: Occur, right: Occur) -> Occur {
-        match (left, right) {
-            (Occur::Should, _) => right,
-            (Occur::Must, Occur::MustNot) => Occur::MustNot,
-            (Occur::Must, _) => Occur::Must,
-            (Occur::MustNot, Occur::MustNot) => Occur::Must,
-            (Occur::MustNot, _) => Occur::MustNot,
-        }
-    }
-}
-
-impl fmt::Display for Occur {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        f.write_char(self.to_char())
-    }
-}
-
-#[cfg(test)]
-mod test {
-    use crate::Occur;
-
-    #[test]
-    fn test_occur_compose() {
-        assert_eq!(Occur::compose(Occur::Should, Occur::Should), Occur::Should);
-        assert_eq!(Occur::compose(Occur::Should, Occur::Must), Occur::Must);
-        assert_eq!(
-            Occur::compose(Occur::Should, Occur::MustNot),
-            Occur::MustNot
-        );
-        assert_eq!(Occur::compose(Occur::Must, Occur::Should), Occur::Must);
-        assert_eq!(Occur::compose(Occur::Must, Occur::Must), Occur::Must);
-        assert_eq!(Occur::compose(Occur::Must, Occur::MustNot), Occur::MustNot);
-        assert_eq!(
-            Occur::compose(Occur::MustNot, Occur::Should),
-            Occur::MustNot
-        );
-        assert_eq!(Occur::compose(Occur::MustNot, Occur::Must), Occur::MustNot);
-        assert_eq!(Occur::compose(Occur::MustNot, Occur::MustNot), Occur::Must);
-    }
-}

query-grammar/src/query_grammar.rs

@@ -1,815 +0,0 @@
-use combine::error::StringStreamError;
-use combine::parser::char::{char, digit, space, spaces, string};
-use combine::parser::combinator::recognize;
-use combine::parser::range::{take_while, take_while1};
-use combine::parser::repeat::escaped;
-use combine::parser::Parser;
-use combine::{
-    attempt, between, choice, eof, many, many1, one_of, optional, parser, satisfy, sep_by,
-    skip_many1, value,
-};
-use once_cell::sync::Lazy;
-use regex::Regex;
-
-use super::user_input_ast::{UserInputAst, UserInputBound, UserInputLeaf, UserInputLiteral};
-use crate::Occur;
-
-// Note: '-' char is only forbidden at the beginning of a field name, would be clearer to add it to
-// special characters.
-const SPECIAL_CHARS: &[char] = &[
-    '+', '^', '`', ':', '{', '}', '"', '[', ']', '(', ')', '!', '\\', '*', ' ',
-];
-const ESCAPED_SPECIAL_CHARS_PATTERN: &str = r#"\\(\+|\^|`|:|\{|\}|"|\[|\]|\(|\)|!|\\|\*|\s)"#;
-
-/// Parses a field_name
-/// A field name must have at least one character and be followed by a colon.
-/// All characters are allowed including special characters `SPECIAL_CHARS`, but these
-/// need to be escaped with a backslash character '\'.
-fn field_name<'a>() -> impl Parser<&'a str, Output = String> {
-    static ESCAPED_SPECIAL_CHARS_RE: Lazy<Regex> =
-        Lazy::new(|| Regex::new(ESCAPED_SPECIAL_CHARS_PATTERN).unwrap());
-
-    recognize::<String, _, _>(escaped(
-        (
-            take_while1(|c| !SPECIAL_CHARS.contains(&c) && c != '-'),
-            take_while(|c| !SPECIAL_CHARS.contains(&c)),
-        ),
-        '\\',
-        satisfy(|_| true), /* if the next character is not a special char, the \ will be treated
-                            * as the \ character. */
-    ))
-    .skip(char(':'))
-    .map(|s| ESCAPED_SPECIAL_CHARS_RE.replace_all(&s, "$1").to_string())
-    .and_then(|s: String| match s.is_empty() {
-        true => Err(StringStreamError::UnexpectedParse),
-        _ => Ok(s),
-    })
-}
-
-fn word<'a>() -> impl Parser<&'a str, Output = String> {
-    (
-        satisfy(|c: char| {
-            !c.is_whitespace()
-                && !['-', '^', '`', ':', '{', '}', '"', '[', ']', '(', ')'].contains(&c)
-        }),
-        many(satisfy(|c: char| {
-            !c.is_whitespace() && ![':', '^', '{', '}', '"', '[', ']', '(', ')'].contains(&c)
-        })),
-    )
-        .map(|(s1, s2): (char, String)| format!("{}{}", s1, s2))
-        .and_then(|s: String| match s.as_str() {
-            "OR" | "AND " | "NOT" => Err(StringStreamError::UnexpectedParse),
-            _ => Ok(s),
-        })
-}
-
-// word variant that allows more characters, e.g. for range queries that don't allow field
-// specifier
-fn relaxed_word<'a>() -> impl Parser<&'a str, Output = String> {
-    (
-        satisfy(|c: char| {
-            !c.is_whitespace() && !['`', '{', '}', '"', '[', ']', '(', ')'].contains(&c)
-        }),
-        many(satisfy(|c: char| {
-            !c.is_whitespace() && !['{', '}', '"', '[', ']', '(', ')'].contains(&c)
-        })),
-    )
-        .map(|(s1, s2): (char, String)| format!("{}{}", s1, s2))
-}
-
-/// Parses a date time according to rfc3339
-/// 2015-08-02T18:54:42+02
-/// 2021-04-13T19:46:26.266051969+00:00
-///
-/// NOTE: also accepts 999999-99-99T99:99:99.266051969+99:99
-/// We delegate rejecting such invalid dates to the logical AST computation code
-/// which invokes `time::OffsetDateTime::parse(..., &Rfc3339)` on the value to actually parse
-/// it (instead of merely extracting the datetime value as string as done here).
-fn date_time<'a>() -> impl Parser<&'a str, Output = String> {
-    let two_digits = || recognize::<String, _, _>((digit(), digit()));
-
-    // Parses a time zone
-    // -06:30
-    // Z
-    let time_zone = {
-        let utc = recognize::<String, _, _>(char('Z'));
-        let offset = recognize((
-            choice([char('-'), char('+')]),
-            two_digits(),
-            char(':'),
-            two_digits(),
-        ));
-
-        utc.or(offset)
-    };
-
-    // Parses a date
-    // 2010-01-30
-    let date = {
-        recognize::<String, _, _>((
-            many1::<String, _, _>(digit()),
-            char('-'),
-            two_digits(),
-            char('-'),
-            two_digits(),
-        ))
-    };
-
-    // Parses a time
-    // 12:30:02
-    // 19:46:26.266051969
-    let time = {
-        recognize::<String, _, _>((
-            two_digits(),
-            char(':'),
-            two_digits(),
-            char(':'),
-            two_digits(),
-            optional((char('.'), many1::<String, _, _>(digit()))),
-            time_zone,
-        ))
-    };
-
-    recognize((date, char('T'), time))
-}
-
-fn term_val<'a>() -> impl Parser<&'a str, Output = String> {
-    let phrase = char('"').with(many1(satisfy(|c| c != '"'))).skip(char('"'));
-    negative_number().or(phrase.or(word()))
-}
-
-fn term_query<'a>() -> impl Parser<&'a str, Output = UserInputLiteral> {
-    (field_name(), term_val(), slop_val()).map(|(field_name, phrase, slop)| UserInputLiteral {
-        field_name: Some(field_name),
-        phrase,
-        slop,
-    })
-}
-
-fn slop_val<'a>() -> impl Parser<&'a str, Output = u32> {
-    let slop =
-        (char('~'), many1(digit())).and_then(|(_, slop): (_, String)| match slop.parse::<u32>() {
-            Ok(d) => Ok(d),
-            _ => Err(StringStreamError::UnexpectedParse),
-        });
-    optional(slop).map(|slop| match slop {
-        Some(d) => d,
-        _ => 0,
-    })
-}
-
-fn literal<'a>() -> impl Parser<&'a str, Output = UserInputLeaf> {
-    let term_default_field = (term_val(), slop_val()).map(|(phrase, slop)| UserInputLiteral {
-        field_name: None,
-        phrase,
-        slop,
-    });
-
-    attempt(term_query())
-        .or(term_default_field)
-        .map(UserInputLeaf::from)
-}
-
-fn negative_number<'a>() -> impl Parser<&'a str, Output = String> {
-    (
-        char('-'),
-        many1(digit()),
-        optional((char('.'), many1(digit()))),
-    )
-        .map(|(s1, s2, s3): (char, String, Option<(char, String)>)| {
-            if let Some(('.', s3)) = s3 {
-                format!("{}{}.{}", s1, s2, s3)
-            } else {
-                format!("{}{}", s1, s2)
-            }
-        })
-}
-
-fn spaces1<'a>() -> impl Parser<&'a str, Output = ()> {
-    skip_many1(space())
-}
-
-/// Function that parses a range out of a Stream
-/// Supports ranges like:
-/// [5 TO 10], {5 TO 10}, [* TO 10], [10 TO *], {10 TO *], >5, <=10
-/// [a TO *], [a TO c], [abc TO bcd}
-fn range<'a>() -> impl Parser<&'a str, Output = UserInputLeaf> {
-    let range_term_val = || {
-        attempt(date_time())
-            .or(negative_number())
-            .or(relaxed_word())
-            .or(char('*').with(value("*".to_string())))
-    };
-
-    // check for unbounded range in the form of <5, <=10, >5, >=5
-    let elastic_unbounded_range = (
-        choice([
-            attempt(string(">=")),
-            attempt(string("<=")),
-            attempt(string("<")),
-            attempt(string(">")),
-        ])
-        .skip(spaces()),
-        range_term_val(),
-    )
-        .map(
-            |(comparison_sign, bound): (&str, String)| match comparison_sign {
-                ">=" => (UserInputBound::Inclusive(bound), UserInputBound::Unbounded),
-                "<=" => (UserInputBound::Unbounded, UserInputBound::Inclusive(bound)),
-                "<" => (UserInputBound::Unbounded, UserInputBound::Exclusive(bound)),
-                ">" => (UserInputBound::Exclusive(bound), UserInputBound::Unbounded),
-                // default case
-                _ => (UserInputBound::Unbounded, UserInputBound::Unbounded),
-            },
-        );
-    let lower_bound = (one_of("{[".chars()), range_term_val()).map(
-        |(boundary_char, lower_bound): (char, String)| {
-            if lower_bound == "*" {
-                UserInputBound::Unbounded
-            } else if boundary_char == '{' {
-                UserInputBound::Exclusive(lower_bound)
-            } else {
-                UserInputBound::Inclusive(lower_bound)
-            }
-        },
-    );
-    let upper_bound = (range_term_val(), one_of("}]".chars())).map(
-        |(higher_bound, boundary_char): (String, char)| {
-            if higher_bound == "*" {
-                UserInputBound::Unbounded
-            } else if boundary_char == '}' {
-                UserInputBound::Exclusive(higher_bound)
-            } else {
-                UserInputBound::Inclusive(higher_bound)
-            }
-        },
-    );
-    // return only lower and upper
-    let lower_to_upper = (
-        lower_bound.skip((spaces(), string("TO"), spaces())),
-        upper_bound,
-    );
-
-    (
-        optional(field_name()).skip(spaces()),
-        // try elastic first, if it matches, the range is unbounded
-        attempt(elastic_unbounded_range).or(lower_to_upper),
-    )
-        .map(|(field, (lower, upper))|
-             // Construct the leaf from extracted field (optional)
-             // and bounds
-             UserInputLeaf::Range {
-                 field,
-                 lower,
-                 upper
-    })
-}
-
-/// Function that parses a set out of a Stream
-/// Supports ranges like: `IN [val1 val2 val3]`
-fn set<'a>() -> impl Parser<&'a str, Output = UserInputLeaf> {
-    let term_list = between(char('['), char(']'), sep_by(term_val(), spaces()));
-
-    let set_content = ((string("IN"), spaces()), term_list).map(|(_, elements)| elements);
-
-    (optional(attempt(field_name().skip(spaces()))), set_content)
-        .map(|(field, elements)| UserInputLeaf::Set { field, elements })
-}
-
-fn negate(expr: UserInputAst) -> UserInputAst {
-    expr.unary(Occur::MustNot)
-}
-
-fn leaf<'a>() -> impl Parser<&'a str, Output = UserInputAst> {
-    parser(|input| {
-        char('(')
-            .with(ast())
-            .skip(char(')'))
-            .or(char('*').map(|_| UserInputAst::from(UserInputLeaf::All)))
-            .or(attempt(
-                string("NOT").skip(spaces1()).with(leaf()).map(negate),
-            ))
-            .or(attempt(range().map(UserInputAst::from)))
-            .or(attempt(set().map(UserInputAst::from)))
-            .or(literal().map(UserInputAst::from))
-            .parse_stream(input)
-            .into_result()
-    })
-}
-
-fn occur_symbol<'a>() -> impl Parser<&'a str, Output = Occur> {
-    char('-')
-        .map(|_| Occur::MustNot)
-        .or(char('+').map(|_| Occur::Must))
-}
-
-fn occur_leaf<'a>() -> impl Parser<&'a str, Output = (Option<Occur>, UserInputAst)> {
-    (optional(occur_symbol()), boosted_leaf())
-}
-
-fn positive_float_number<'a>() -> impl Parser<&'a str, Output = f64> {
-    (many1(digit()), optional((char('.'), many1(digit())))).map(
-        |(int_part, decimal_part_opt): (String, Option<(char, String)>)| {
-            let mut float_str = int_part;
-            if let Some((chr, decimal_str)) = decimal_part_opt {
-                float_str.push(chr);
-                float_str.push_str(&decimal_str);
-            }
-            float_str.parse::<f64>().unwrap()
-        },
-    )
-}
-
-fn boost<'a>() -> impl Parser<&'a str, Output = f64> {
-    (char('^'), positive_float_number()).map(|(_, boost)| boost)
-}
-
-fn boosted_leaf<'a>() -> impl Parser<&'a str, Output = UserInputAst> {
-    (leaf(), optional(boost())).map(|(leaf, boost_opt)| match boost_opt {
-        Some(boost) if (boost - 1.0).abs() > f64::EPSILON => {
-            UserInputAst::Boost(Box::new(leaf), boost)
-        }
-        _ => leaf,
-    })
-}
-
-#[derive(Clone, Copy)]
-enum BinaryOperand {
-    Or,
-    And,
-}
-
-fn binary_operand<'a>() -> impl Parser<&'a str, Output = BinaryOperand> {
-    string("AND")
-        .with(value(BinaryOperand::And))
-        .or(string("OR").with(value(BinaryOperand::Or)))
-}
-
-fn aggregate_binary_expressions(
-    left: UserInputAst,
-    others: Vec<(BinaryOperand, UserInputAst)>,
-) -> UserInputAst {
-    let mut dnf: Vec<Vec<UserInputAst>> = vec![vec![left]];
-    for (operator, operand_ast) in others {
-        match operator {
-            BinaryOperand::And => {
-                if let Some(last) = dnf.last_mut() {
-                    last.push(operand_ast);
-                }
-            }
-            BinaryOperand::Or => {
-                dnf.push(vec![operand_ast]);
-            }
-        }
-    }
-    if dnf.len() == 1 {
-        UserInputAst::and(dnf.into_iter().next().unwrap()) //< safe
-    } else {
-        let conjunctions = dnf.into_iter().map(UserInputAst::and).collect();
-        UserInputAst::or(conjunctions)
-    }
-}
-
-fn operand_leaf<'a>() -> impl Parser<&'a str, Output = (BinaryOperand, UserInputAst)> {
-    (
-        binary_operand().skip(spaces()),
-        boosted_leaf().skip(spaces()),
-    )
-}
-
-pub fn ast<'a>() -> impl Parser<&'a str, Output = UserInputAst> {
-    let boolean_expr = (boosted_leaf().skip(spaces()), many1(operand_leaf()))
-        .map(|(left, right)| aggregate_binary_expressions(left, right));
-    let whitespace_separated_leaves = many1(occur_leaf().skip(spaces().silent())).map(
-        |subqueries: Vec<(Option<Occur>, UserInputAst)>| {
-            if subqueries.len() == 1 {
-                let (occur_opt, ast) = subqueries.into_iter().next().unwrap();
-                match occur_opt.unwrap_or(Occur::Should) {
-                    Occur::Must | Occur::Should => ast,
-                    Occur::MustNot => UserInputAst::Clause(vec![(Some(Occur::MustNot), ast)]),
-                }
-            } else {
-                UserInputAst::Clause(subqueries.into_iter().collect())
-            }
-        },
-    );
-    let expr = attempt(boolean_expr).or(whitespace_separated_leaves);
-    spaces().with(expr).skip(spaces())
-}
-
-pub fn parse_to_ast<'a>() -> impl Parser<&'a str, Output = UserInputAst> {
-    spaces()
-        .with(optional(ast()).skip(eof()))
-        .map(|opt_ast| opt_ast.unwrap_or_else(UserInputAst::empty_query))
-}
-
-#[cfg(test)]
-mod test {
-
-    type TestParseResult = Result<(), StringStreamError>;
-
-    use combine::parser::Parser;
-
-    use super::*;
-
-    pub fn nearly_equals(a: f64, b: f64) -> bool {
-        (a - b).abs() < 0.0005 * (a + b).abs()
-    }
-
-    fn assert_nearly_equals(expected: f64, val: f64) {
-        assert!(
-            nearly_equals(val, expected),
-            "Got {}, expected {}.",
-            val,
-            expected
-        );
-    }
-
-    #[test]
-    fn test_occur_symbol() -> TestParseResult {
-        assert_eq!(super::occur_symbol().parse("-")?, (Occur::MustNot, ""));
-        assert_eq!(super::occur_symbol().parse("+")?, (Occur::Must, ""));
-        Ok(())
-    }
-
-    #[test]
-    fn test_positive_float_number() {
-        fn valid_parse(float_str: &str, expected_val: f64, expected_remaining: &str) {
-            let (val, remaining) = positive_float_number().parse(float_str).unwrap();
-            assert_eq!(remaining, expected_remaining);
-            assert_nearly_equals(val, expected_val);
-        }
-        fn error_parse(float_str: &str) {
-            assert!(positive_float_number().parse(float_str).is_err());
-        }
-        valid_parse("1.0", 1.0, "");
-        valid_parse("1", 1.0, "");
-        valid_parse("0.234234 aaa", 0.234234f64, " aaa");
-        error_parse(".3332");
-        error_parse("1.");
-        error_parse("-1.");
-    }
-
-    #[test]
-    fn test_date_time() {
-        let (val, remaining) = date_time()
-            .parse("2015-08-02T18:54:42+02:30")
-            .expect("cannot parse date");
-        assert_eq!(val, "2015-08-02T18:54:42+02:30");
-        assert_eq!(remaining, "");
-        assert!(date_time().parse("2015-08-02T18:54:42+02").is_err());
-
-        let (val, remaining) = date_time()
-            .parse("2021-04-13T19:46:26.266051969+00:00")
-            .expect("cannot parse fractional date");
-        assert_eq!(val, "2021-04-13T19:46:26.266051969+00:00");
-        assert_eq!(remaining, "");
-    }
-
-    fn test_parse_query_to_ast_helper(query: &str, expected: &str) {
-        let query = parse_to_ast().parse(query).unwrap().0;
-        let query_str = format!("{:?}", query);
-        assert_eq!(query_str, expected);
-    }
-
-    fn test_is_parse_err(query: &str) {
-        assert!(parse_to_ast().parse(query).is_err());
-    }
-
-    #[test]
-    fn test_parse_empty_to_ast() {
-        test_parse_query_to_ast_helper("", "<emptyclause>");
-    }
-
-    #[test]
-    fn test_parse_query_to_ast_hyphen() {
-        test_parse_query_to_ast_helper("\"www-form-encoded\"", "\"www-form-encoded\"");
-        test_parse_query_to_ast_helper("www-form-encoded", "\"www-form-encoded\"");
-        test_parse_query_to_ast_helper("www-form-encoded", "\"www-form-encoded\"");
-    }
-
-    #[test]
-    fn test_parse_query_to_ast_not_op() {
-        assert_eq!(
-            format!("{:?}", parse_to_ast().parse("NOT")),
-            "Err(UnexpectedParse)"
-        );
-        test_parse_query_to_ast_helper("NOTa", "\"NOTa\"");
-        test_parse_query_to_ast_helper("NOT a", "(-\"a\")");
-    }
-
-    #[test]
-    fn test_boosting() {
-        assert!(parse_to_ast().parse("a^2^3").is_err());
-        assert!(parse_to_ast().parse("a^2^").is_err());
-        test_parse_query_to_ast_helper("a^3", "(\"a\")^3");
-        test_parse_query_to_ast_helper("a^3 b^2", "(*(\"a\")^3 *(\"b\")^2)");
-        test_parse_query_to_ast_helper("a^1", "\"a\"");
-    }
-
-    #[test]
-    fn test_parse_query_to_ast_binary_op() {
-        test_parse_query_to_ast_helper("a AND b", "(+\"a\" +\"b\")");
-        test_parse_query_to_ast_helper("a OR b", "(?\"a\" ?\"b\")");
-        test_parse_query_to_ast_helper("a OR b AND c", "(?\"a\" ?(+\"b\" +\"c\"))");
-        test_parse_query_to_ast_helper("a AND b         AND c", "(+\"a\" +\"b\" +\"c\")");
-        assert_eq!(
-            format!("{:?}", parse_to_ast().parse("a OR b aaa")),
-            "Err(UnexpectedParse)"
-        );
-        assert_eq!(
-            format!("{:?}", parse_to_ast().parse("a AND b aaa")),
-            "Err(UnexpectedParse)"
-        );
-        assert_eq!(
-            format!("{:?}", parse_to_ast().parse("aaa a OR b ")),
-            "Err(UnexpectedParse)"
-        );
-        assert_eq!(
-            format!("{:?}", parse_to_ast().parse("aaa ccc a OR b ")),
-            "Err(UnexpectedParse)"
-        );
-    }
-
-    #[test]
-    fn test_parse_elastic_query_ranges() {
-        test_parse_query_to_ast_helper("title: >a", "\"title\":{\"a\" TO \"*\"}");
-        test_parse_query_to_ast_helper("title:>=a", "\"title\":[\"a\" TO \"*\"}");
-        test_parse_query_to_ast_helper("title: <a", "\"title\":{\"*\" TO \"a\"}");
-        test_parse_query_to_ast_helper("title:<=a", "\"title\":{\"*\" TO \"a\"]");
-        test_parse_query_to_ast_helper("title:<=bsd", "\"title\":{\"*\" TO \"bsd\"]");
-
-        test_parse_query_to_ast_helper("weight: >70", "\"weight\":{\"70\" TO \"*\"}");
-        test_parse_query_to_ast_helper("weight:>=70", "\"weight\":[\"70\" TO \"*\"}");
-        test_parse_query_to_ast_helper("weight: <70", "\"weight\":{\"*\" TO \"70\"}");
-        test_parse_query_to_ast_helper("weight:<=70", "\"weight\":{\"*\" TO \"70\"]");
-        test_parse_query_to_ast_helper("weight: >60.7", "\"weight\":{\"60.7\" TO \"*\"}");
-
-        test_parse_query_to_ast_helper("weight: <= 70", "\"weight\":{\"*\" TO \"70\"]");
-
-        test_parse_query_to_ast_helper("weight: <= 70.5", "\"weight\":{\"*\" TO \"70.5\"]");
-    }
-
-    #[test]
-    fn test_occur_leaf() {
-        let ((occur, ast), _) = super::occur_leaf().parse("+abc").unwrap();
-        assert_eq!(occur, Some(Occur::Must));
-        assert_eq!(format!("{:?}", ast), "\"abc\"");
-    }
-
-    #[test]
-    fn test_field_name() {
-        assert_eq!(
-            super::field_name().parse(".my.field.name:a"),
-            Ok((".my.field.name".to_string(), "a"))
-        );
-        assert_eq!(
-            super::field_name().parse(r#"にんじん:a"#),
-            Ok(("にんじん".to_string(), "a"))
-        );
-        assert_eq!(
-            super::field_name().parse(r#"my\field:a"#),
-            Ok((r#"my\field"#.to_string(), "a"))
-        );
-        assert!(super::field_name().parse("my field:a").is_err());
-        assert_eq!(
-            super::field_name().parse("\\(1\\+1\\):2"),
-            Ok(("(1+1)".to_string(), "2"))
-        );
-        assert_eq!(
-            super::field_name().parse("my_field_name:a"),
-            Ok(("my_field_name".to_string(), "a"))
-        );
-        assert_eq!(
-            super::field_name().parse("myfield.b:hello").unwrap(),
-            ("myfield.b".to_string(), "hello")
-        );
-        assert_eq!(
-            super::field_name().parse(r#"myfield\.b:hello"#).unwrap(),
-            (r#"myfield\.b"#.to_string(), "hello")
-        );
-        assert!(super::field_name().parse("my_field_name").is_err());
-        assert!(super::field_name().parse(":a").is_err());
-        assert!(super::field_name().parse("-my_field:a").is_err());
-        assert_eq!(
-            super::field_name().parse("_my_field:a"),
-            Ok(("_my_field".to_string(), "a"))
-        );
-        assert_eq!(
-            super::field_name().parse("~my~field:a"),
-            Ok(("~my~field".to_string(), "a"))
-        );
-        for special_char in SPECIAL_CHARS.iter() {
-            let query = &format!("\\{special_char}my\\{special_char}field:a");
-            assert_eq!(
-                super::field_name().parse(query),
-                Ok((format!("{special_char}my{special_char}field"), "a"))
-            );
-        }
-    }
-
-    #[test]
-    fn test_field_name_re() {
-        let escaped_special_chars_re = Regex::new(ESCAPED_SPECIAL_CHARS_PATTERN).unwrap();
-        for special_char in SPECIAL_CHARS.iter() {
-            assert_eq!(
-                escaped_special_chars_re.replace_all(&format!("\\{}", special_char), "$1"),
-                special_char.to_string()
-            );
-        }
-    }
-
-    #[test]
-    fn test_range_parser() {
-        // testing the range() parser separately
-        let res = range()
-            .parse("title: <hello")
-            .expect("Cannot parse felxible bound word")
-            .0;
-        let expected = UserInputLeaf::Range {
-            field: Some("title".to_string()),
-            lower: UserInputBound::Unbounded,
-            upper: UserInputBound::Exclusive("hello".to_string()),
-        };
-        let res2 = range()
-            .parse("title:{* TO hello}")
-            .expect("Cannot parse ununbounded to word")
-            .0;
-        assert_eq!(res, expected);
-        assert_eq!(res2, expected);
-
-        let expected_weight = UserInputLeaf::Range {
-            field: Some("weight".to_string()),
-            lower: UserInputBound::Inclusive("71.2".to_string()),
-            upper: UserInputBound::Unbounded,
-        };
-        let res3 = range()
-            .parse("weight: >=71.2")
-            .expect("Cannot parse flexible bound float")
-            .0;
-        let res4 = range()
-            .parse("weight:[71.2 TO *}")
-            .expect("Cannot parse float to unbounded")
-            .0;
-        assert_eq!(res3, expected_weight);
-        assert_eq!(res4, expected_weight);
-
-        let expected_dates = UserInputLeaf::Range {
-            field: Some("date_field".to_string()),
-            lower: UserInputBound::Exclusive("2015-08-02T18:54:42Z".to_string()),
-            upper: UserInputBound::Inclusive("2021-08-02T18:54:42+02:30".to_string()),
-        };
-        let res5 = range()
-            .parse("date_field:{2015-08-02T18:54:42Z TO 2021-08-02T18:54:42+02:30]")
-            .expect("Cannot parse date range")
-            .0;
-        assert_eq!(res5, expected_dates);
-
-        let expected_flexible_dates = UserInputLeaf::Range {
-            field: Some("date_field".to_string()),
-            lower: UserInputBound::Unbounded,
-            upper: UserInputBound::Inclusive("2021-08-02T18:54:42.12345+02:30".to_string()),
-        };
-
-        let res6 = range()
-            .parse("date_field: <=2021-08-02T18:54:42.12345+02:30")
-            .expect("Cannot parse date range")
-            .0;
-        assert_eq!(res6, expected_flexible_dates);
-        // IP Range Unbounded
-        let expected_weight = UserInputLeaf::Range {
-            field: Some("ip".to_string()),
-            lower: UserInputBound::Inclusive("::1".to_string()),
-            upper: UserInputBound::Unbounded,
-        };
-        let res1 = range()
-            .parse("ip: >=::1")
-            .expect("Cannot parse ip v6 format")
-            .0;
-        let res2 = range()
-            .parse("ip:[::1 TO *}")
-            .expect("Cannot parse ip v6 format")
-            .0;
-        assert_eq!(res1, expected_weight);
-        assert_eq!(res2, expected_weight);
-
-        // IP Range Bounded
-        let expected_weight = UserInputLeaf::Range {
-            field: Some("ip".to_string()),
-            lower: UserInputBound::Inclusive("::0.0.0.50".to_string()),
-            upper: UserInputBound::Exclusive("::0.0.0.52".to_string()),
-        };
-        let res1 = range()
-            .parse("ip:[::0.0.0.50 TO ::0.0.0.52}")
-            .expect("Cannot parse ip v6 format")
-            .0;
-        assert_eq!(res1, expected_weight);
-    }
-
-    #[test]
-    fn test_parse_query_to_triming_spaces() {
-        test_parse_query_to_ast_helper("   abc", "\"abc\"");
-        test_parse_query_to_ast_helper("abc ", "\"abc\"");
-        test_parse_query_to_ast_helper("(  a OR abc)", "(?\"a\" ?\"abc\")");
-        test_parse_query_to_ast_helper("(a  OR abc)", "(?\"a\" ?\"abc\")");
-        test_parse_query_to_ast_helper("(a OR  abc)", "(?\"a\" ?\"abc\")");
-        test_parse_query_to_ast_helper("a OR abc ", "(?\"a\" ?\"abc\")");
-        test_parse_query_to_ast_helper("(a OR abc )", "(?\"a\" ?\"abc\")");
-        test_parse_query_to_ast_helper("(a OR  abc) ", "(?\"a\" ?\"abc\")");
-    }
-
-    #[test]
-    fn test_parse_query_single_term() {
-        test_parse_query_to_ast_helper("abc", "\"abc\"");
-    }
-
-    #[test]
-    fn test_parse_query_default_clause() {
-        test_parse_query_to_ast_helper("a b", "(*\"a\" *\"b\")");
-    }
-
-    #[test]
-    fn test_parse_query_must_default_clause() {
-        test_parse_query_to_ast_helper("+(a b)", "(*\"a\" *\"b\")");
-    }
-
-    #[test]
-    fn test_parse_query_must_single_term() {
-        test_parse_query_to_ast_helper("+d", "\"d\"");
-    }
-
-    #[test]
-    fn test_single_term_with_field() {
-        test_parse_query_to_ast_helper("abc:toto", "\"abc\":\"toto\"");
-    }
-
-    #[test]
-    fn test_single_term_with_float() {
-        test_parse_query_to_ast_helper("abc:1.1", "\"abc\":\"1.1\"");
-        test_parse_query_to_ast_helper("a.b.c:1.1", "\"a.b.c\":\"1.1\"");
-        test_parse_query_to_ast_helper("a\\ b\\ c:1.1", "\"a b c\":\"1.1\"");
-    }
-
-    #[test]
-    fn test_must_clause() {
-        test_parse_query_to_ast_helper("(+a +b)", "(+\"a\" +\"b\")");
-    }
-
-    #[test]
-    fn test_parse_test_query_plus_a_b_plus_d() {
-        test_parse_query_to_ast_helper("+(a b) +d", "(+(*\"a\" *\"b\") +\"d\")");
-    }
-
-    #[test]
-    fn test_parse_test_query_set() {
-        test_parse_query_to_ast_helper("abc: IN [a b c]", r#""abc": IN ["a" "b" "c"]"#);
-        test_parse_query_to_ast_helper("abc: IN [1]", r#""abc": IN ["1"]"#);
-        test_parse_query_to_ast_helper("abc: IN []", r#""abc": IN []"#);
-        test_parse_query_to_ast_helper("IN [1 2]", r#"IN ["1" "2"]"#);
-    }
-
-    #[test]
-    fn test_parse_test_query_other() {
-        test_parse_query_to_ast_helper("(+a +b) d", "(*(+\"a\" +\"b\") *\"d\")");
-        test_parse_query_to_ast_helper("+abc:toto", "\"abc\":\"toto\"");
-        test_parse_query_to_ast_helper("+a\\+b\\+c:toto", "\"a+b+c\":\"toto\"");
-        test_parse_query_to_ast_helper("(+abc:toto -titi)", "(+\"abc\":\"toto\" -\"titi\")");
-        test_parse_query_to_ast_helper("-abc:toto", "(-\"abc\":\"toto\")");
-        test_is_parse_err("--abc:toto");
-        test_parse_query_to_ast_helper("abc:a b", "(*\"abc\":\"a\" *\"b\")");
-        test_parse_query_to_ast_helper("abc:\"a b\"", "\"abc\":\"a b\"");
-        test_parse_query_to_ast_helper("foo:[1 TO 5]", "\"foo\":[\"1\" TO \"5\"]");
-    }
-
-    #[test]
-    fn test_parse_query_with_range() {
-        test_parse_query_to_ast_helper("[1 TO 5]", "[\"1\" TO \"5\"]");
-        test_parse_query_to_ast_helper("foo:{a TO z}", "\"foo\":{\"a\" TO \"z\"}");
-        test_parse_query_to_ast_helper("foo:[1 TO toto}", "\"foo\":[\"1\" TO \"toto\"}");
-        test_parse_query_to_ast_helper("foo:[* TO toto}", "\"foo\":{\"*\" TO \"toto\"}");
-        test_parse_query_to_ast_helper("foo:[1 TO *}", "\"foo\":[\"1\" TO \"*\"}");
-        test_parse_query_to_ast_helper(
-            "1.2.foo.bar:[1.1 TO *}",
-            "\"1.2.foo.bar\":[\"1.1\" TO \"*\"}",
-        );
-        test_is_parse_err("abc +    ");
-    }
-
-    #[test]
-    fn test_slop() {
-        assert!(parse_to_ast().parse("\"a b\"~").is_err());
-        assert!(parse_to_ast().parse("foo:\"a b\"~").is_err());
-        assert!(parse_to_ast().parse("\"a b\"~a").is_err());
-        assert!(parse_to_ast().parse("\"a b\"~100000000000000000").is_err());
-
-        test_parse_query_to_ast_helper("\"a b\"^2~4", "(*(\"a b\")^2 *\"~4\")");
-        test_parse_query_to_ast_helper("\"~Document\"", "\"~Document\"");
-        test_parse_query_to_ast_helper("~Document", "\"~Document\"");
-        test_parse_query_to_ast_helper("a~2", "\"a~2\"");
-        test_parse_query_to_ast_helper("\"a b\"~0", "\"a b\"");
-        test_parse_query_to_ast_helper("\"a b\"~1", "\"a b\"~1");
-        test_parse_query_to_ast_helper("\"a b\"~3", "\"a b\"~3");
-        test_parse_query_to_ast_helper("foo:\"a b\"~300", "\"foo\":\"a b\"~300");
-        test_parse_query_to_ast_helper("\"a b\"~300^2", "(\"a b\"~300)^2");
-    }
-}