rust/tantivy
1
0
Fork 0
mirror of https://github.com/quickwit-oss/tantivy.git synced 2026-01-03 07:42:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Compare commits

base: rust:main
Branches Tags
rust:main rust:paul.masurel/skip-first-doc-on-intersection rust:paul.masurel/search_into_danger_zone_fastfield rust:bucket_id_agg rust:paradedb/fix-agg-validation rust:flatheadmill-geo rust:paradedb-paradedb/fix-overflow-issue rust:low_card_optimization rust:paul.masurel/clippy rust:qw-airmail-20250916 rust:unit-test-aggregation-sum-other rust:clippy-and-cleanups rust:trinity.pointard/opti-sstable-sorted-ords-to-term rust:release_0.24 rust:dependabot/cargo/rand-0.9.1 rust:dependabot/cargo/lru-0.14.0 rust:dependabot/cargo/fs4-0.13.1 rust:tagged-user-input-ast rust:1686a/parse-agg-res rust:dependabot/cargo/rand_distr-0.5.1 rust:issue/2411-adding-panic-handler-to-rayon-merge-thread-pool rust:dependabot/cargo/nom-8 rust:upgrade_deps rust:test_parser rust:dependabot/github_actions/codecov/codecov-action-5 rust:binggan-0.12.0 rust:agg_format rust:path_collector rust:test_order_bug rust:test_columnar_multi rust:columnar_optionalmask_in_multivalues rust:prefix-phrase-query-optim rust:tantivy-with-oneshot rust:update_example rust:store-reader-enumerate rust:document-deserialize-seed rust:check_doc_mem rust:experimental-do-not-merge rust:expdonotmerge rust:cherry-pick-vec-instead-of-btreemap rust:validate_length rust:raphael_bench_avx_filter rust:raphael_optim_memset2 rust:query-field-name-avoid-alloc rust:check_bug rust:missing_test rust:list_fields rust:block-cache-control rust:trinity--multilayer-sstable rust:tracing-tantivy2 rust:tracing-tantivy rust:index_writer_err rust:default_fast_field_tokenizer rust:fmassot/add-preserve-original-ascii-folding rust:numeric_with_str_col rust:straightforward-simplify rust:fmassot/fake-test-branch rust:fmassot/dynamic-tokenizer-poc-follow-up rust:fmassot/add-box-token-filter-and-refactor rust:dynamic-tokenizer-proof-of-concept rust:fmassot/add-box-tokenizer-and-token-filter rust:tokenizer-api-0.1.1 rust:addconversion rust:removed-extra-enum-alternative-for-datetime-precision rust:issue/1981-reproduce-bm25-bug rust:tokenizer_docs rust:merge-proptests rust:columnar-todo rust:quickwit-0.5 rust:remove-byteorder rust:remove_dyn rust:pr/alexcole/1855-inject-bm25-statistics rust:evance-expose-build-query-from-user-input-ast rust:missing-specialized-method-arcdyncolumn rust:fast-u64-range-query rust:0.19.2 rust:columnar-cli2 rust:trinity--pub-term-as-slice rust:typed-column3 rust:columnar-merge3 rust:columnar-merge-exploring-fn rust:use_columnar rust:typed-column2 rust:term_agg_bucket_limit rust:range rust:typed-column rust:columnar-main rust:forced-types rust:u128-columnar rust:use_stats rust:truncation_comment rust:0.19.1 rust:trinity/yoke-documents rust:column-handle rust:columnar-cli rust:columnar rust:revert-1711-sparse_dense_index rust:commit-changes rust:sparse_dense_index_headroom rust:optional_column rust:no-column-for-multivalued rust:fmassot--bench-hdfs-with-array rust:bugfix-position-broken rust:float rust:simplify rust:debug_time rust:sparse_codecs rust:fix_estimate rust:column-reader rust:column-reader-b rust:columnreader-simple rust:col-trait-refact rust:ip_fastfield rust:column-trait-2 rust:refact-dyn-column rust:column-trait rust:refact-codec-trait rust:pr/PSeitz/1485-1 rust:fastfieldcodec rust:termmap_per_field rust:fastfield-codecs-tests rust:rayon-indexing-writers rust:issue/1251 rust:index-bench rust:fix_open_bytes_index rust:warming rust:revert-1222-issue/1195 rust:add-random-bench-for-fast-field rust:wasm-friendly-exploratory rust:revert-1159-patch-1 rust:debug-position-panic-0.15.3 rust:fix-issue-1111 rust:remove_rand rust:unused-deps rust:hfb/phrase_match_slop rust:0.15.3 rust:expose-min-max-on-multi-fast-field-reader rust:revert-1094-expose-min-max-on-multi-fast-field-reader rust:hotfix-1088 rust:owned-bytes-instead-of-rawdoc rust:rihardsk-date-ranges rust:issue/997 rust:0.13.3 rust:fail-fast-store-checkpoint rust:debugging-store rust:issue/reproduce-897 rust:issue/969-reproduce rust:troublescooter-static rust:intermediate-mergeable-troublescooter rust:barrotsteindev-filter-collector-tpredicatevalue rust:issue/938b rust:issue/938 rust:barrotsteindev-filter-collector rust:issue/922 rust:issue/896 rust:quickfix-explain rust:slog rust:limit-random-seeks rust:0.13.1 rust:issue/866b rust:bugfix-unittest-macos rust:githubactions rust:fieldnorm_readers rust:blockwand rust:nodeffeatfails rust:removedaliveiterator rust:robyoung-introduce-offset-to-top-docs rust:specialization-of-foreach-for-union-scorer rust:elshize-bmw rust:cutt rust:nrt rust:storewriter-out rust:segment_writer_in_ram2 rust:polling rust:audunhalland-query-boost rust:poll-meta rust:segment_writer_in_ram rust:bundle rust:petr-tik-n662_footer_is_incompatible_impl rust:race-condition-in-tests rust:hotfix-0.10.3-issue681 rust:remove-tru rust:kkoziara-pre-tokenized-text rust:issue/681 rust:issue/680 rust:refactoring/field rust:updating/uuid rust:fix-bench rust:hotfix-656 rust:readd-macos-ci rust:incremental-search rust:crate-division rust:streamer-with-state rust:criterion rust:python-binding-changes rust:bump-version rust:gh-pages rust:issue/554 rust:perf/experimental-union-for_each rust:issue/526b rust:softcommits rust:0.9.1-hotfix rust:revert-521-fix-non-english-stemmers rust:issue/500-hotfix rust:hotfix/0.8.2 rust:issue/483 rust:bug/merge-deted rust:issue/407 rust:issue/457 rust:segment_fruits rust:doc rust:dds/lenient rust:issue/webasm rust:issue/396 rust:broken_collector_with_result rust:dss/automaton-weight-suggestion rust:wasm rust:common-crawl rust:tantivy-imhotep
rust:0.25.0 rust:0.22.1 rust:0.24.2 rust:0.24.1 rust:tantivy-query-grammar-v0.24.0 rust:tantivy-columnar-v0.5.0 rust:tantivy-bitpacker-v0.8.0 rust:ownedbytes-v0.9.0 rust:tantivy-stacker-v0.5.0 rust:tantivy-sstable-v0.5.0 rust:tantivy-common-v0.9.0 rust:tantivy-tokenizer-api-v0.5.0 rust:0.24 rust:qw-airmail-20250224-hotfix-merge-panic rust:qw-airmail-20250219-hotfix-merge-panic rust:qw-airmail-202406 rust:qw-airmail-20240611 rust:qw-airmail-20240606 rust:quickwit-rev rust:0.22.0 rust:tantivy-stacker-v0.3.0 rust:ownedbytes-v0.7.0 rust:tantivy-tokenizer-api-v0.3.0 rust:tantivy-columnar-v0.3.0 rust:tantivy-common-v0.7.0 rust:tantivy-query-grammar-v0.22.0 rust:tantivy-sstable-v0.3.0 rust:tantivy-bitpacker-v0.6.0 rust:0.21.1 rust:0.21 rust:0.20.2 rust:0.20.1 rust:0.20 rust:quickwit-0.5-rev rust:0.19.2 rust:0.19.1 rust:0.19 rust:0.18.1 rust:0.18 rust:0.17 rust:0.16.1 rust:0.15.3 rust:0.15.2 rust:0.15.1 rust:0.15 rust:0.14 rust:0.13.3 rust:0.13.2 rust:0.13.1 rust:0.13 rust:0.12 rust:0.11.3 rust:0.11.1 rust:0.11 rust:0.10.3 rust:0.1 rust:0.10.2 rust:0.10.1 rust:0.10.0 rust:0.9.1 rust:0.9 rust:0.8.2 rust:0.8.2b rust:0.8.1 rust:0.8.0 rust:0.7.2 rust:0.7.1 rust:0.7.0 rust:0.6.1 rust:0.6.0 rust:0.5.2 rust:0.5.1 rust:0.5.0 rust:0.4.3 rust:0.4.2 rust:0.4.0 rust:0.3.1 rust:0.3.0 rust:0.2.0
..
compare: rust:paul.masurel/clippy
Branches Tags
rust:paul.masurel/skip-first-doc-on-intersection rust:main rust:paul.masurel/search_into_danger_zone_fastfield rust:bucket_id_agg rust:paradedb/fix-agg-validation rust:flatheadmill-geo rust:paradedb-paradedb/fix-overflow-issue rust:low_card_optimization rust:paul.masurel/clippy rust:qw-airmail-20250916 rust:unit-test-aggregation-sum-other rust:clippy-and-cleanups rust:trinity.pointard/opti-sstable-sorted-ords-to-term rust:release_0.24 rust:dependabot/cargo/rand-0.9.1 rust:dependabot/cargo/lru-0.14.0 rust:dependabot/cargo/fs4-0.13.1 rust:tagged-user-input-ast rust:1686a/parse-agg-res rust:dependabot/cargo/rand_distr-0.5.1 rust:issue/2411-adding-panic-handler-to-rayon-merge-thread-pool rust:dependabot/cargo/nom-8 rust:upgrade_deps rust:test_parser rust:dependabot/github_actions/codecov/codecov-action-5 rust:binggan-0.12.0 rust:agg_format rust:path_collector rust:test_order_bug rust:test_columnar_multi rust:columnar_optionalmask_in_multivalues rust:prefix-phrase-query-optim rust:tantivy-with-oneshot rust:update_example rust:store-reader-enumerate rust:document-deserialize-seed rust:check_doc_mem rust:experimental-do-not-merge rust:expdonotmerge rust:cherry-pick-vec-instead-of-btreemap rust:validate_length rust:raphael_bench_avx_filter rust:raphael_optim_memset2 rust:query-field-name-avoid-alloc rust:check_bug rust:missing_test rust:list_fields rust:block-cache-control rust:trinity--multilayer-sstable rust:tracing-tantivy2 rust:tracing-tantivy rust:index_writer_err rust:default_fast_field_tokenizer rust:fmassot/add-preserve-original-ascii-folding rust:numeric_with_str_col rust:straightforward-simplify rust:fmassot/fake-test-branch rust:fmassot/dynamic-tokenizer-poc-follow-up rust:fmassot/add-box-token-filter-and-refactor rust:dynamic-tokenizer-proof-of-concept rust:fmassot/add-box-tokenizer-and-token-filter rust:tokenizer-api-0.1.1 rust:addconversion rust:removed-extra-enum-alternative-for-datetime-precision rust:issue/1981-reproduce-bm25-bug rust:tokenizer_docs rust:merge-proptests rust:columnar-todo rust:quickwit-0.5 rust:remove-byteorder rust:remove_dyn rust:pr/alexcole/1855-inject-bm25-statistics rust:evance-expose-build-query-from-user-input-ast rust:missing-specialized-method-arcdyncolumn rust:fast-u64-range-query rust:0.19.2 rust:columnar-cli2 rust:trinity--pub-term-as-slice rust:typed-column3 rust:columnar-merge3 rust:columnar-merge-exploring-fn rust:use_columnar rust:typed-column2 rust:term_agg_bucket_limit rust:range rust:typed-column rust:columnar-main rust:forced-types rust:u128-columnar rust:use_stats rust:truncation_comment rust:0.19.1 rust:trinity/yoke-documents rust:column-handle rust:columnar-cli rust:columnar rust:revert-1711-sparse_dense_index rust:commit-changes rust:sparse_dense_index_headroom rust:optional_column rust:no-column-for-multivalued rust:fmassot--bench-hdfs-with-array rust:bugfix-position-broken rust:float rust:simplify rust:debug_time rust:sparse_codecs rust:fix_estimate rust:column-reader rust:column-reader-b rust:columnreader-simple rust:col-trait-refact rust:ip_fastfield rust:column-trait-2 rust:refact-dyn-column rust:column-trait rust:refact-codec-trait rust:pr/PSeitz/1485-1 rust:fastfieldcodec rust:termmap_per_field rust:fastfield-codecs-tests rust:rayon-indexing-writers rust:issue/1251 rust:index-bench rust:fix_open_bytes_index rust:warming rust:revert-1222-issue/1195 rust:add-random-bench-for-fast-field rust:wasm-friendly-exploratory rust:revert-1159-patch-1 rust:debug-position-panic-0.15.3 rust:fix-issue-1111 rust:remove_rand rust:unused-deps rust:hfb/phrase_match_slop rust:0.15.3 rust:expose-min-max-on-multi-fast-field-reader rust:revert-1094-expose-min-max-on-multi-fast-field-reader rust:hotfix-1088 rust:owned-bytes-instead-of-rawdoc rust:rihardsk-date-ranges rust:issue/997 rust:0.13.3 rust:fail-fast-store-checkpoint rust:debugging-store rust:issue/reproduce-897 rust:issue/969-reproduce rust:troublescooter-static rust:intermediate-mergeable-troublescooter rust:barrotsteindev-filter-collector-tpredicatevalue rust:issue/938b rust:issue/938 rust:barrotsteindev-filter-collector rust:issue/922 rust:issue/896 rust:quickfix-explain rust:slog rust:limit-random-seeks rust:0.13.1 rust:issue/866b rust:bugfix-unittest-macos rust:githubactions rust:fieldnorm_readers rust:blockwand rust:nodeffeatfails rust:removedaliveiterator rust:robyoung-introduce-offset-to-top-docs rust:specialization-of-foreach-for-union-scorer rust:elshize-bmw rust:cutt rust:nrt rust:storewriter-out rust:segment_writer_in_ram2 rust:polling rust:audunhalland-query-boost rust:poll-meta rust:segment_writer_in_ram rust:bundle rust:petr-tik-n662_footer_is_incompatible_impl rust:race-condition-in-tests rust:hotfix-0.10.3-issue681 rust:remove-tru rust:kkoziara-pre-tokenized-text rust:issue/681 rust:issue/680 rust:refactoring/field rust:updating/uuid rust:fix-bench rust:hotfix-656 rust:readd-macos-ci rust:incremental-search rust:crate-division rust:streamer-with-state rust:criterion rust:python-binding-changes rust:bump-version rust:gh-pages rust:issue/554 rust:perf/experimental-union-for_each rust:issue/526b rust:softcommits rust:0.9.1-hotfix rust:revert-521-fix-non-english-stemmers rust:issue/500-hotfix rust:hotfix/0.8.2 rust:issue/483 rust:bug/merge-deted rust:issue/407 rust:issue/457 rust:segment_fruits rust:doc rust:dds/lenient rust:issue/webasm rust:issue/396 rust:broken_collector_with_result rust:dss/automaton-weight-suggestion rust:wasm rust:common-crawl rust:tantivy-imhotep
rust:0.25.0 rust:0.22.1 rust:0.24.2 rust:0.24.1 rust:tantivy-query-grammar-v0.24.0 rust:tantivy-columnar-v0.5.0 rust:tantivy-bitpacker-v0.8.0 rust:ownedbytes-v0.9.0 rust:tantivy-stacker-v0.5.0 rust:tantivy-sstable-v0.5.0 rust:tantivy-common-v0.9.0 rust:tantivy-tokenizer-api-v0.5.0 rust:0.24 rust:qw-airmail-20250224-hotfix-merge-panic rust:qw-airmail-20250219-hotfix-merge-panic rust:qw-airmail-202406 rust:qw-airmail-20240611 rust:qw-airmail-20240606 rust:quickwit-rev rust:0.22.0 rust:tantivy-stacker-v0.3.0 rust:ownedbytes-v0.7.0 rust:tantivy-tokenizer-api-v0.3.0 rust:tantivy-columnar-v0.3.0 rust:tantivy-common-v0.7.0 rust:tantivy-query-grammar-v0.22.0 rust:tantivy-sstable-v0.3.0 rust:tantivy-bitpacker-v0.6.0 rust:0.21.1 rust:0.21 rust:0.20.2 rust:0.20.1 rust:0.20 rust:quickwit-0.5-rev rust:0.19.2 rust:0.19.1 rust:0.19 rust:0.18.1 rust:0.18 rust:0.17 rust:0.16.1 rust:0.15.3 rust:0.15.2 rust:0.15.1 rust:0.15 rust:0.14 rust:0.13.3 rust:0.13.2 rust:0.13.1 rust:0.13 rust:0.12 rust:0.11.3 rust:0.11.1 rust:0.11 rust:0.10.3 rust:0.1 rust:0.10.2 rust:0.10.1 rust:0.10.0 rust:0.9.1 rust:0.9 rust:0.8.2 rust:0.8.2b rust:0.8.1 rust:0.8.0 rust:0.7.2 rust:0.7.1 rust:0.7.0 rust:0.6.1 rust:0.6.0 rust:0.5.2 rust:0.5.1 rust:0.5.0 rust:0.4.3 rust:0.4.2 rust:0.4.0 rust:0.3.1 rust:0.3.0 rust:0.2.0

1 Commits
main ... paul.masur

Author SHA1 Message Date
Paul Masurel
ebb82dc549 clippy 2025-10-08 17:07:07 +02:00
166 changed files with 4016 additions and 12102 deletions
Expand all files Collapse all files

4
.github/workflows/coverage.yml vendored
View File

@@ -15,11 +15,11 @@ jobs:
steps:
- uses: actions/checkout@v4
- name: Install Rust
run: rustup toolchain install nightly-2025-12-01 --profile minimal --component llvm-tools-preview
run: rustup toolchain install nightly-2024-07-01 --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-2025-12-01 llvm-cov --all-features --workspace --doctests --lcov --output-path lcov.info
run: cargo +nightly-2024-07-01 llvm-cov --all-features --workspace --doctests --lcov --output-path lcov.info
- name: Upload coverage to Codecov
uses: codecov/codecov-action@v3
continue-on-error: true

30
.github/workflows/test.yml vendored
View File

@@ -39,11 +39,11 @@ jobs:
- name: Check Formatting
run: cargo +nightly fmt --all -- --check
- name: Check Stable Compilation
run: cargo build --all-features
- name: Check Bench Compilation
run: cargo +nightly bench --no-run --profile=dev --all-features
@@ -59,10 +59,10 @@ jobs:
strategy:
matrix:
features:
- { label: "all", flags: "mmap,stopwords,lz4-compression,zstd-compression,failpoints,stemmer" }
- { label: "quickwit", flags: "mmap,quickwit,failpoints" }
- { label: "none", flags: "" }
features: [
{ label: "all", flags: "mmap,stopwords,lz4-compression,zstd-compression,failpoints" },
{ label: "quickwit", flags: "mmap,quickwit,failpoints" }
]
name: test-${{ matrix.features.label}}
@@ -80,21 +80,7 @@ jobs:
- uses: Swatinem/rust-cache@v2
- name: Run tests
run: |
# if matrix.feature.flags is empty then run on --lib to avoid compiling examples
# (as most of them rely on mmap) otherwise run all
if [ -z "${{ matrix.features.flags }}" ]; then
cargo +stable nextest run --lib --no-default-features --verbose --workspace
else
cargo +stable nextest run --features ${{ matrix.features.flags }} --no-default-features --verbose --workspace
fi
run: cargo +stable nextest run --features ${{ matrix.features.flags }} --verbose --workspace
- name: Run doctests
run: |
# if matrix.feature.flags is empty then run on --lib to avoid compiling examples
# (as most of them rely on mmap) otherwise run all
if [ -z "${{ matrix.features.flags }}" ]; then
echo "no doctest for no feature flag"
else
cargo +stable test --doc --features ${{ matrix.features.flags }} --verbose --workspace
fi
run: cargo +stable test --doc --features ${{ matrix.features.flags }} --verbose --workspace

2
CHANGELOG.md
View File

@@ -78,7 +78,7 @@ This will slightly increase space and access time. [#2439](https://github.com/qu
- **Store DateTime as nanoseconds in doc store** DateTime in the doc store was truncated to microseconds previously. This removes this truncation, while still keeping backwards compatibility. [#2486](https://github.com/quickwit-oss/tantivy/pull/2486)(@PSeitz)
- **Performance/Memory**
- **Performace/Memory**
- lift clauses in LogicalAst for optimized ast during execution [#2449](https://github.com/quickwit-oss/tantivy/pull/2449)(@PSeitz)
- Use Vec instead of BTreeMap to back OwnedValue object [#2364](https://github.com/quickwit-oss/tantivy/pull/2364)(@fulmicoton)
- Replace TantivyDocument with CompactDoc. CompactDoc is much smaller and provides similar performance. [#2402](https://github.com/quickwit-oss/tantivy/pull/2402)(@PSeitz)

25
Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy"
version = "0.26.0"
version = "0.25.0"
authors = ["Paul Masurel <paul.masurel@gmail.com>"]
license = "MIT"
categories = ["database-implementations", "data-structures"]
@@ -37,7 +37,7 @@ fs4 = { version = "0.13.1", optional = true }
levenshtein_automata = "0.2.1"
uuid = { version = "1.0.0", features = ["v4", "serde"] }
crossbeam-channel = "0.5.4"
rust-stemmers = { version = "1.2.0", optional = true }
rust-stemmers = "1.2.0"
downcast-rs = "2.0.1"
bitpacking = { version = "0.9.2", default-features = false, features = [
"bitpacker4x",
@@ -69,18 +69,17 @@ hyperloglogplus = { version = "0.4.1", features = ["const-loop"] }
futures-util = { version = "0.3.28", optional = true }
futures-channel = { version = "0.3.28", optional = true }
fnv = "1.0.7"
typetag = "0.2.21"
[target.'cfg(windows)'.dependencies]
winapi = "0.3.9"
[dev-dependencies]
binggan = "0.14.2"
binggan = "0.14.0"
rand = "0.8.5"
maplit = "1.0.2"
matches = "0.1.9"
pretty_assertions = "1.2.1"
proptest = "1.7.0"
proptest = "1.0.0"
test-log = "0.2.10"
futures = "0.3.21"
paste = "1.0.11"
@@ -88,7 +87,7 @@ more-asserts = "0.3.1"
rand_distr = "0.4.3"
time = { version = "0.3.10", features = ["serde-well-known", "macros"] }
postcard = { version = "1.0.4", features = [
"use-std",
"use-std",
], default-features = false }
[target.'cfg(not(windows))'.dev-dependencies]
@@ -113,8 +112,7 @@ debug-assertions = true
overflow-checks = true
[features]
default = ["mmap", "stopwords", "lz4-compression", "columnar-zstd-compression", "stemmer"]
stemmer = ["rust-stemmers"]
default = ["mmap", "stopwords", "lz4-compression", "columnar-zstd-compression"]
mmap = ["fs4", "tempfile", "memmap2"]
stopwords = []
@@ -174,18 +172,7 @@ harness = false
name = "exists_json"
harness = false
[[bench]]
name = "range_query"
harness = false
[[bench]]
name = "and_or_queries"
harness = false
[[bench]]
name = "range_queries"
harness = false
[[bench]]
name = "bool_queries_with_range"
harness = false

3
README.md
View File

@@ -23,6 +23,8 @@ performance for different types of queries/collections.
Your mileage WILL vary depending on the nature of queries and their load.
<img src="doc/assets/images/searchbenchmark.png">
Details about the benchmark can be found at this [repository](https://github.com/quickwit-oss/search-benchmark-game).
## Features
@@ -123,7 +125,6 @@ You can also find other bindings on [GitHub](https://github.com/search?q=tantivy
- [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
- [Bichon](https://github.com/rustmailer/bichon): A lightweight, high-performance Rust email archiver with WebUI
- and [more](https://github.com/search?q=tantivy)!
### On average, how much faster is Tantivy compared to Lucene?

2
TODO.txt
View File

@@ -10,7 +10,7 @@ rename FastFieldReaders::open to load
remove fast field reader
find a way to unify the two DateTime.
re-add type check in the filter wrapper
readd type check in the filter wrapper
add unit test on columnar list columns.

180
benches/agg_bench.rs
View File

@@ -1,6 +1,5 @@
use binggan::plugins::PeakMemAllocPlugin;
use binggan::{black_box, InputGroup, PeakMemAlloc, INSTRUMENTED_SYSTEM};
use rand::distributions::WeightedIndex;
use rand::prelude::SliceRandom;
use rand::rngs::StdRng;
use rand::{Rng, SeedableRng};
@@ -55,19 +54,11 @@ fn bench_agg(mut group: InputGroup<Index>) {
register!(group, extendedstats_f64);
register!(group, percentiles_f64);
register!(group, terms_few);
register!(group, terms_all_unique);
register!(group, terms_many);
register!(group, terms_many_top_1000);
register!(group, terms_many_order_by_term);
register!(group, terms_many_with_top_hits);
register!(group, terms_all_unique_with_avg_sub_agg);
register!(group, terms_many_with_avg_sub_agg);
register!(group, terms_few_with_avg_sub_agg);
register!(group, terms_status_with_avg_sub_agg);
register!(group, terms_status);
register!(group, terms_few_with_histogram);
register!(group, terms_status_with_histogram);
register!(group, terms_many_json_mixed_type_with_avg_sub_agg);
register!(group, cardinality_agg);
@@ -80,15 +71,8 @@ fn bench_agg(mut group: InputGroup<Index>) {
register!(group, histogram);
register!(group, histogram_hard_bounds);
register!(group, histogram_with_avg_sub_agg);
register!(group, histogram_with_term_agg_few);
register!(group, avg_and_range_with_avg_sub_agg);
// Filter aggregation benchmarks
register!(group, filter_agg_all_query_count_agg);
register!(group, filter_agg_term_query_count_agg);
register!(group, filter_agg_all_query_with_sub_aggs);
register!(group, filter_agg_term_query_with_sub_aggs);
group.run();
}
@@ -139,12 +123,12 @@ fn extendedstats_f64(index: &Index) {
}
fn percentiles_f64(index: &Index) {
let agg_req = json!({
"mypercentiles": {
"percentiles": {
"field": "score_f64",
"percents": [ 95, 99, 99.9 ]
}
"mypercentiles": {
"percentiles": {
"field": "score_f64",
"percents": [ 95, 99, 99.9 ]
}
}
});
execute_agg(index, agg_req);
}
@@ -181,19 +165,6 @@ fn terms_few(index: &Index) {
});
execute_agg(index, agg_req);
}
fn terms_status(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_few_terms_status" } },
});
execute_agg(index, agg_req);
}
fn terms_all_unique(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_all_unique_terms" } },
});
execute_agg(index, agg_req);
}
fn terms_many(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_many_terms" } },
@@ -242,63 +213,6 @@ fn terms_many_with_avg_sub_agg(index: &Index) {
});
execute_agg(index, agg_req);
}
fn terms_all_unique_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_all_unique_terms" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn terms_few_with_histogram(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_few_terms" },
"aggs": {
"histo": {"histogram": { "field": "score_f64", "interval": 10 }}
}
}
});
execute_agg(index, agg_req);
}
fn terms_status_with_histogram(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_few_terms_status" },
"aggs": {
"histo": {"histogram": { "field": "score_f64", "interval": 10 }}
}
}
});
execute_agg(index, agg_req);
}
fn terms_few_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_few_terms" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn terms_status_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_few_terms_status" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn terms_many_json_mixed_type_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"my_texts": {
@@ -425,17 +339,6 @@ fn histogram_with_avg_sub_agg(index: &Index) {
});
execute_agg(index, agg_req);
}
fn histogram_with_term_agg_few(index: &Index) {
let agg_req = json!({
"rangef64": {
"histogram": { "field": "score_f64", "interval": 10 },
"aggs": {
"my_texts": { "terms": { "field": "text_few_terms" } }
}
}
});
execute_agg(index, agg_req);
}
fn avg_and_range_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"rangef64": {
@@ -483,21 +386,14 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
.set_stored();
let text_field = schema_builder.add_text_field("text", text_fieldtype);
let json_field = schema_builder.add_json_field("json", FAST);
let text_field_all_unique_terms =
schema_builder.add_text_field("text_all_unique_terms", STRING | FAST);
let text_field_many_terms = schema_builder.add_text_field("text_many_terms", STRING | FAST);
let text_field_many_terms = schema_builder.add_text_field("text_many_terms", STRING | FAST);
let text_field_few_terms = schema_builder.add_text_field("text_few_terms", STRING | FAST);
let text_field_few_terms_status =
schema_builder.add_text_field("text_few_terms_status", STRING | FAST);
let score_fieldtype = tantivy::schema::NumericOptions::default().set_fast();
let score_field = schema_builder.add_u64_field("score", score_fieldtype.clone());
let score_field_f64 = schema_builder.add_f64_field("score_f64", score_fieldtype.clone());
let score_field_i64 = schema_builder.add_i64_field("score_i64", score_fieldtype);
let index = Index::create_from_tempdir(schema_builder.build())?;
let few_terms_data = ["INFO", "ERROR", "WARN", "DEBUG"];
// Approximate production log proportions: INFO dominant, WARN and DEBUG occasional, ERROR rare.
let log_level_distribution = WeightedIndex::new([80u32, 3, 12, 5]).unwrap();
let lg_norm = rand_distr::LogNormal::new(2.996f64, 0.979f64).unwrap();
@@ -513,21 +409,15 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
index_writer.add_document(doc!())?;
}
if cardinality == Cardinality::Multivalued {
let log_level_sample_a = few_terms_data[log_level_distribution.sample(&mut rng)];
let log_level_sample_b = few_terms_data[log_level_distribution.sample(&mut rng)];
index_writer.add_document(doc!(
json_field => json!({"mixed_type": 10.0}),
json_field => json!({"mixed_type": 10.0}),
text_field => "cool",
text_field => "cool",
text_field_all_unique_terms => "cool",
text_field_all_unique_terms => "coolo",
text_field_many_terms => "cool",
text_field_many_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms_status => log_level_sample_a,
text_field_few_terms_status => log_level_sample_b,
score_field => 1u64,
score_field => 1u64,
score_field_f64 => lg_norm.sample(&mut rng),
@@ -552,10 +442,8 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
index_writer.add_document(doc!(
text_field => "cool",
json_field => json,
text_field_all_unique_terms => format!("unique_term_{}", rng.gen::<u64>()),
text_field_many_terms => many_terms_data.choose(&mut rng).unwrap().to_string(),
text_field_few_terms => few_terms_data.choose(&mut rng).unwrap().to_string(),
text_field_few_terms_status => few_terms_data[log_level_distribution.sample(&mut rng)],
score_field => val as u64,
score_field_f64 => lg_norm.sample(&mut rng),
score_field_i64 => val as i64,
@@ -572,61 +460,3 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
Ok(index)
}
// Filter aggregation benchmarks
fn filter_agg_all_query_count_agg(index: &Index) {
let agg_req = json!({
"filtered": {
"filter": "*",
"aggs": {
"count": { "value_count": { "field": "score" } }
}
}
});
execute_agg(index, agg_req);
}
fn filter_agg_term_query_count_agg(index: &Index) {
let agg_req = json!({
"filtered": {
"filter": "text:cool",
"aggs": {
"count": { "value_count": { "field": "score" } }
}
}
});
execute_agg(index, agg_req);
}
fn filter_agg_all_query_with_sub_aggs(index: &Index) {
let agg_req = json!({
"filtered": {
"filter": "*",
"aggs": {
"avg_score": { "avg": { "field": "score" } },
"stats_score": { "stats": { "field": "score_f64" } },
"terms_text": {
"terms": { "field": "text_few_terms" }
}
}
}
});
execute_agg(index, agg_req);
}
fn filter_agg_term_query_with_sub_aggs(index: &Index) {
let agg_req = json!({
"filtered": {
"filter": "text:cool",
"aggs": {
"avg_score": { "avg": { "field": "score" } },
"stats_score": { "stats": { "field": "score_f64" } },
"terms_text": {
"terms": { "field": "text_few_terms" }
}
}
}
});
execute_agg(index, agg_req);
}

160
benches/and_or_queries.rs
View File

@@ -16,15 +16,14 @@
// - This bench isolates boolean iteration speed and intersection/union cost.
// - Use `cargo bench --bench boolean_conjunction` to run.
use binggan::{black_box, BenchGroup, BenchRunner};
use binggan::{black_box, BenchRunner};
use rand::prelude::*;
use rand::rngs::StdRng;
use rand::SeedableRng;
use tantivy::collector::sort_key::SortByStaticFastValue;
use tantivy::collector::{Collector, Count, TopDocs};
use tantivy::query::{Query, QueryParser};
use tantivy::schema::{Schema, FAST, TEXT};
use tantivy::{doc, Index, Order, ReloadPolicy, Searcher};
use tantivy::collector::{Count, TopDocs};
use tantivy::query::QueryParser;
use tantivy::schema::{Schema, TEXT};
use tantivy::{doc, Index, ReloadPolicy, Searcher};
#[derive(Clone)]
struct BenchIndex {
@@ -34,6 +33,23 @@ struct BenchIndex {
query_parser: QueryParser,
}
impl BenchIndex {
#[inline(always)]
fn count_query(&self, query_str: &str) -> usize {
let query = self.query_parser.parse_query(query_str).unwrap();
self.searcher.search(&query, &Count).unwrap()
}
#[inline(always)]
fn topk_len(&self, query_str: &str, k: usize) -> usize {
let query = self.query_parser.parse_query(query_str).unwrap();
self.searcher
.search(&query, &TopDocs::with_limit(k))
.unwrap()
.len()
}
}
/// Build a single index containing both fields (title, body) and
/// return two BenchIndex views:
/// - single_field: QueryParser defaults to only "body"
@@ -43,8 +59,6 @@ fn build_shared_indices(num_docs: usize, p_a: f32, p_b: f32, p_c: f32) -> (Bench
let mut schema_builder = Schema::builder();
let f_title = schema_builder.add_text_field("title", TEXT);
let f_body = schema_builder.add_text_field("body", TEXT);
let f_score = schema_builder.add_u64_field("score", FAST);
let f_score2 = schema_builder.add_u64_field("score2", FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema.clone());
@@ -53,13 +67,11 @@ fn build_shared_indices(num_docs: usize, p_a: f32, p_b: f32, p_c: f32) -> (Bench
// Populate: spread each present token 90/10 to body/title
{
let mut writer = index.writer_with_num_threads(1, 500_000_000).unwrap();
let mut writer = index.writer(500_000_000).unwrap();
for _ in 0..num_docs {
let has_a = rng.gen_bool(p_a as f64);
let has_b = rng.gen_bool(p_b as f64);
let has_c = rng.gen_bool(p_c as f64);
let score = rng.gen_range(0u64..100u64);
let score2 = rng.gen_range(0u64..100_000u64);
let mut title_tokens: Vec<&str> = Vec::new();
let mut body_tokens: Vec<&str> = Vec::new();
if has_a {
@@ -89,9 +101,7 @@ fn build_shared_indices(num_docs: usize, p_a: f32, p_b: f32, p_c: f32) -> (Bench
writer
.add_document(doc!(
f_title=>title_tokens.join(" "),
f_body=>body_tokens.join(" "),
f_score=>score,
f_score2=>score2,
f_body=>body_tokens.join(" ")
))
.unwrap();
}
@@ -143,76 +153,72 @@ fn main() {
),
];
let queries = &["a", "+a +b", "+a +b +c", "a OR b", "a OR b OR c"];
let mut runner = BenchRunner::new();
for (label, n, pa, pb, pc) in scenarios {
let (single_view, multi_view) = build_shared_indices(n, pa, pb, pc);
for (view_name, bench_index) in [("single_field", single_view), ("multi_field", multi_view)]
// Single-field group: default field is body only
{
// Single-field group: default field is body only
let mut group = runner.new_group();
group.set_name(format!("{} — {}", view_name, label));
for query_str in queries {
add_bench_task(&mut group, &bench_index, query_str, Count, "count");
add_bench_task(
&mut group,
&bench_index,
query_str,
TopDocs::with_limit(10).order_by_score(),
"top10",
);
add_bench_task(
&mut group,
&bench_index,
query_str,
TopDocs::with_limit(10).order_by_fast_field::<u64>("score", Order::Asc),
"top10_by_ff",
);
add_bench_task(
&mut group,
&bench_index,
query_str,
TopDocs::with_limit(10).order_by((
SortByStaticFastValue::<u64>::for_field("score"),
SortByStaticFastValue::<u64>::for_field("score2"),
)),
"top10_by_2ff",
);
}
group.set_name(format!("single_field — {}", label));
group.register_with_input("+a_+b_count", &single_view, |benv: &BenchIndex| {
black_box(benv.count_query("+a +b"))
});
group.register_with_input("+a_+b_+c_count", &single_view, |benv: &BenchIndex| {
black_box(benv.count_query("+a +b +c"))
});
group.register_with_input("+a_+b_top10", &single_view, |benv: &BenchIndex| {
black_box(benv.topk_len("+a +b", 10))
});
group.register_with_input("+a_+b_+c_top10", &single_view, |benv: &BenchIndex| {
black_box(benv.topk_len("+a +b +c", 10))
});
// OR queries
group.register_with_input("a_OR_b_count", &single_view, |benv: &BenchIndex| {
black_box(benv.count_query("a OR b"))
});
group.register_with_input("a_OR_b_OR_c_count", &single_view, |benv: &BenchIndex| {
black_box(benv.count_query("a OR b OR c"))
});
group.register_with_input("a_OR_b_top10", &single_view, |benv: &BenchIndex| {
black_box(benv.topk_len("a OR b", 10))
});
group.register_with_input("a_OR_b_OR_c_top10", &single_view, |benv: &BenchIndex| {
black_box(benv.topk_len("a OR b OR c", 10))
});
group.run();
}
// Multi-field group: default fields are [title, body]
{
let mut group = runner.new_group();
group.set_name(format!("multi_field — {}", label));
group.register_with_input("+a_+b_count", &multi_view, |benv: &BenchIndex| {
black_box(benv.count_query("+a +b"))
});
group.register_with_input("+a_+b_+c_count", &multi_view, |benv: &BenchIndex| {
black_box(benv.count_query("+a +b +c"))
});
group.register_with_input("+a_+b_top10", &multi_view, |benv: &BenchIndex| {
black_box(benv.topk_len("+a +b", 10))
});
group.register_with_input("+a_+b_+c_top10", &multi_view, |benv: &BenchIndex| {
black_box(benv.topk_len("+a +b +c", 10))
});
// OR queries
group.register_with_input("a_OR_b_count", &multi_view, |benv: &BenchIndex| {
black_box(benv.count_query("a OR b"))
});
group.register_with_input("a_OR_b_OR_c_count", &multi_view, |benv: &BenchIndex| {
black_box(benv.count_query("a OR b OR c"))
});
group.register_with_input("a_OR_b_top10", &multi_view, |benv: &BenchIndex| {
black_box(benv.topk_len("a OR b", 10))
});
group.register_with_input("a_OR_b_OR_c_top10", &multi_view, |benv: &BenchIndex| {
black_box(benv.topk_len("a OR b OR c", 10))
});
group.run();
}
}
}
fn add_bench_task<C: Collector + 'static>(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query_str: &str,
collector: C,
collector_name: &str,
) {
let task_name = format!("{}_{}", query_str.replace(" ", "_"), collector_name);
let query = bench_index.query_parser.parse_query(query_str).unwrap();
let search_task = SearchTask {
searcher: bench_index.searcher.clone(),
collector,
query,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
struct SearchTask<C: Collector> {
searcher: Searcher,
collector: C,
query: Box<dyn Query>,
}
impl<C: Collector> SearchTask<C> {
#[inline(never)]
pub fn run(&self) -> usize {
self.searcher.search(&self.query, &self.collector).unwrap();
1
}
}

288
benches/bool_queries_with_range.rs
View File

@@ -1,288 +0,0 @@
use binggan::{black_box, BenchGroup, BenchRunner};
use rand::prelude::*;
use rand::rngs::StdRng;
use rand::SeedableRng;
use tantivy::collector::{Collector, Count, DocSetCollector, TopDocs};
use tantivy::query::{Query, QueryParser};
use tantivy::schema::{Schema, FAST, INDEXED, TEXT};
use tantivy::{doc, Index, Order, ReloadPolicy, Searcher};
#[derive(Clone)]
struct BenchIndex {
#[allow(dead_code)]
index: Index,
searcher: Searcher,
query_parser: QueryParser,
}
fn build_shared_indices(num_docs: usize, p_title_a: f32, distribution: &str) -> BenchIndex {
// Unified schema
let mut schema_builder = Schema::builder();
let f_title = schema_builder.add_text_field("title", TEXT);
let f_num_rand = schema_builder.add_u64_field("num_rand", INDEXED);
let f_num_asc = schema_builder.add_u64_field("num_asc", INDEXED);
let f_num_rand_fast = schema_builder.add_u64_field("num_rand_fast", INDEXED | FAST);
let f_num_asc_fast = schema_builder.add_u64_field("num_asc_fast", INDEXED | FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema.clone());
// Populate index with stable RNG for reproducibility.
let mut rng = StdRng::from_seed([7u8; 32]);
{
let mut writer = index.writer_with_num_threads(1, 4_000_000_000).unwrap();
match distribution {
"dense" => {
for doc_id in 0..num_docs {
// Always add title to avoid empty documents
let title_token = if rng.gen_bool(p_title_a as f64) {
"a"
} else {
"b"
};
let num_rand = rng.gen_range(0u64..1000u64);
let num_asc = (doc_id / 10000) as u64;
writer
.add_document(doc!(
f_title=>title_token,
f_num_rand=>num_rand,
f_num_asc=>num_asc,
f_num_rand_fast=>num_rand,
f_num_asc_fast=>num_asc,
))
.unwrap();
}
}
"sparse" => {
for doc_id in 0..num_docs {
// Always add title to avoid empty documents
let title_token = if rng.gen_bool(p_title_a as f64) {
"a"
} else {
"b"
};
let num_rand = rng.gen_range(0u64..10000000u64);
let num_asc = doc_id as u64;
writer
.add_document(doc!(
f_title=>title_token,
f_num_rand=>num_rand,
f_num_asc=>num_asc,
f_num_rand_fast=>num_rand,
f_num_asc_fast=>num_asc,
))
.unwrap();
}
}
_ => {
panic!("Unsupported distribution type");
}
}
writer.commit().unwrap();
}
// Prepare reader/searcher once.
let reader = index
.reader_builder()
.reload_policy(ReloadPolicy::Manual)
.try_into()
.unwrap();
let searcher = reader.searcher();
// Build query parser for title field
let qp_title = QueryParser::for_index(&index, vec![f_title]);
BenchIndex {
index,
searcher,
query_parser: qp_title,
}
}
fn main() {
// Prepare corpora with varying scenarios
let scenarios = vec![
(
"dense and 99% a".to_string(),
10_000_000,
0.99,
"dense",
0,
9,
),
(
"dense and 99% a".to_string(),
10_000_000,
0.99,
"dense",
990,
999,
),
(
"sparse and 99% a".to_string(),
10_000_000,
0.99,
"sparse",
0,
9,
),
(
"sparse and 99% a".to_string(),
10_000_000,
0.99,
"sparse",
9_999_990,
9_999_999,
),
];
let mut runner = BenchRunner::new();
for (scenario_id, n, p_title_a, num_rand_distribution, range_low, range_high) in scenarios {
// Build index for this scenario
let bench_index = build_shared_indices(n, p_title_a, num_rand_distribution);
// Create benchmark group
let mut group = runner.new_group();
// Now set the name (this moves scenario_id)
group.set_name(scenario_id);
// Define all four field types
let field_names = ["num_rand", "num_asc", "num_rand_fast", "num_asc_fast"];
// Define the three terms we want to test with
let terms = ["a", "b", "z"];
// Generate all combinations of terms and field names
let mut queries = Vec::new();
for &term in &terms {
for &field_name in &field_names {
let query_str = format!(
"{} AND {}:[{} TO {}]",
term, field_name, range_low, range_high
);
queries.push((query_str, field_name.to_string()));
}
}
let query_str = format!(
"{}:[{} TO {}] AND {}:[{} TO {}]",
"num_rand_fast", range_low, range_high, "num_asc_fast", range_low, range_high
);
queries.push((query_str, "num_asc_fast".to_string()));
// Run all benchmark tasks for each query and its corresponding field name
for (query_str, field_name) in queries {
run_benchmark_tasks(&mut group, &bench_index, &query_str, &field_name);
}
group.run();
}
}
/// Run all benchmark tasks for a given query string and field name
fn run_benchmark_tasks(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query_str: &str,
field_name: &str,
) {
// Test count
add_bench_task(bench_group, bench_index, query_str, Count, "count");
// Test all results
add_bench_task(
bench_group,
bench_index,
query_str,
DocSetCollector,
"all results",
);
// Test top 100 by the field (if it's a FAST field)
if field_name.ends_with("_fast") {
// Ascending order
{
let collector_name = format!("top100_by_{}_asc", field_name);
let field_name_owned = field_name.to_string();
add_bench_task(
bench_group,
bench_index,
query_str,
TopDocs::with_limit(100).order_by_fast_field::<u64>(field_name_owned, Order::Asc),
&collector_name,
);
}
// Descending order
{
let collector_name = format!("top100_by_{}_desc", field_name);
let field_name_owned = field_name.to_string();
add_bench_task(
bench_group,
bench_index,
query_str,
TopDocs::with_limit(100).order_by_fast_field::<u64>(field_name_owned, Order::Desc),
&collector_name,
);
}
}
}
fn add_bench_task<C: Collector + 'static>(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query_str: &str,
collector: C,
collector_name: &str,
) {
let task_name = format!("{}_{}", query_str.replace(" ", "_"), collector_name);
let query = bench_index.query_parser.parse_query(query_str).unwrap();
let search_task = SearchTask {
searcher: bench_index.searcher.clone(),
collector,
query,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
struct SearchTask<C: Collector> {
searcher: Searcher,
collector: C,
query: Box<dyn Query>,
}
impl<C: Collector> SearchTask<C> {
#[inline(never)]
pub fn run(&self) -> usize {
let result = self.searcher.search(&self.query, &self.collector).unwrap();
if let Some(count) = (&result as &dyn std::any::Any).downcast_ref::<usize>() {
*count
} else if let Some(top_docs) = (&result as &dyn std::any::Any)
.downcast_ref::<Vec<(Option<u64>, tantivy::DocAddress)>>()
{
top_docs.len()
} else if let Some(top_docs) =
(&result as &dyn std::any::Any).downcast_ref::<Vec<(u64, tantivy::DocAddress)>>()
{
top_docs.len()
} else if let Some(doc_set) = (&result as &dyn std::any::Any)
.downcast_ref::<std::collections::HashSet<tantivy::DocAddress>>()
{
doc_set.len()
} else {
eprintln!(
"Unknown collector result type: {:?}",
std::any::type_name::<C::Fruit>()
);
0
}
}
}

365
benches/range_queries.rs
View File

@@ -1,365 +0,0 @@
use std::ops::Bound;
use binggan::{black_box, BenchGroup, BenchRunner};
use rand::prelude::*;
use rand::rngs::StdRng;
use rand::SeedableRng;
use tantivy::collector::{Count, DocSetCollector, TopDocs};
use tantivy::query::RangeQuery;
use tantivy::schema::{Schema, FAST, INDEXED};
use tantivy::{doc, Index, Order, ReloadPolicy, Searcher, Term};
#[derive(Clone)]
struct BenchIndex {
#[allow(dead_code)]
index: Index,
searcher: Searcher,
}
fn build_shared_indices(num_docs: usize, distribution: &str) -> BenchIndex {
// Schema with fast fields only
let mut schema_builder = Schema::builder();
let f_num_rand_fast = schema_builder.add_u64_field("num_rand_fast", INDEXED | FAST);
let f_num_asc_fast = schema_builder.add_u64_field("num_asc_fast", INDEXED | FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema.clone());
// Populate index with stable RNG for reproducibility.
let mut rng = StdRng::from_seed([7u8; 32]);
{
let mut writer = index.writer_with_num_threads(1, 4_000_000_000).unwrap();
match distribution {
"dense" => {
for doc_id in 0..num_docs {
let num_rand = rng.gen_range(0u64..1000u64);
let num_asc = (doc_id / 10000) as u64;
writer
.add_document(doc!(
f_num_rand_fast=>num_rand,
f_num_asc_fast=>num_asc,
))
.unwrap();
}
}
"sparse" => {
for doc_id in 0..num_docs {
let num_rand = rng.gen_range(0u64..10000000u64);
let num_asc = doc_id as u64;
writer
.add_document(doc!(
f_num_rand_fast=>num_rand,
f_num_asc_fast=>num_asc,
))
.unwrap();
}
}
_ => {
panic!("Unsupported distribution type");
}
}
writer.commit().unwrap();
}
// Prepare reader/searcher once.
let reader = index
.reader_builder()
.reload_policy(ReloadPolicy::Manual)
.try_into()
.unwrap();
let searcher = reader.searcher();
BenchIndex { index, searcher }
}
fn main() {
// Prepare corpora with varying scenarios
let scenarios = vec![
// Dense distribution - random values in small range (0-999)
(
"dense_values_search_low_value_range".to_string(),
10_000_000,
"dense",
0,
9,
),
(
"dense_values_search_high_value_range".to_string(),
10_000_000,
"dense",
990,
999,
),
(
"dense_values_search_out_of_range".to_string(),
10_000_000,
"dense",
1000,
1002,
),
(
"sparse_values_search_low_value_range".to_string(),
10_000_000,
"sparse",
0,
9,
),
(
"sparse_values_search_high_value_range".to_string(),
10_000_000,
"sparse",
9_999_990,
9_999_999,
),
(
"sparse_values_search_out_of_range".to_string(),
10_000_000,
"sparse",
10_000_000,
10_000_002,
),
];
let mut runner = BenchRunner::new();
for (scenario_id, n, num_rand_distribution, range_low, range_high) in scenarios {
// Build index for this scenario
let bench_index = build_shared_indices(n, num_rand_distribution);
// Create benchmark group
let mut group = runner.new_group();
// Now set the name (this moves scenario_id)
group.set_name(scenario_id);
// Define fast field types
let field_names = ["num_rand_fast", "num_asc_fast"];
// Generate range queries for fast fields
for &field_name in &field_names {
// Create the range query
let field = bench_index.searcher.schema().get_field(field_name).unwrap();
let lower_term = Term::from_field_u64(field, range_low);
let upper_term = Term::from_field_u64(field, range_high);
let query = RangeQuery::new(Bound::Included(lower_term), Bound::Included(upper_term));
run_benchmark_tasks(
&mut group,
&bench_index,
query,
field_name,
range_low,
range_high,
);
}
group.run();
}
}
/// Run all benchmark tasks for a given range query and field name
fn run_benchmark_tasks(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query: RangeQuery,
field_name: &str,
range_low: u64,
range_high: u64,
) {
// Test count
add_bench_task_count(
bench_group,
bench_index,
query.clone(),
"count",
field_name,
range_low,
range_high,
);
// Test top 100 by the field (ascending order)
{
let collector_name = format!("top100_by_{}_asc", field_name);
let field_name_owned = field_name.to_string();
add_bench_task_top100_asc(
bench_group,
bench_index,
query.clone(),
&collector_name,
field_name,
range_low,
range_high,
field_name_owned,
);
}
// Test top 100 by the field (descending order)
{
let collector_name = format!("top100_by_{}_desc", field_name);
let field_name_owned = field_name.to_string();
add_bench_task_top100_desc(
bench_group,
bench_index,
query,
&collector_name,
field_name,
range_low,
range_high,
field_name_owned,
);
}
}
fn add_bench_task_count(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query: RangeQuery,
collector_name: &str,
field_name: &str,
range_low: u64,
range_high: u64,
) {
let task_name = format!(
"range_{}_[{} TO {}]_{}",
field_name, range_low, range_high, collector_name
);
let search_task = CountSearchTask {
searcher: bench_index.searcher.clone(),
query,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
fn add_bench_task_docset(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query: RangeQuery,
collector_name: &str,
field_name: &str,
range_low: u64,
range_high: u64,
) {
let task_name = format!(
"range_{}_[{} TO {}]_{}",
field_name, range_low, range_high, collector_name
);
let search_task = DocSetSearchTask {
searcher: bench_index.searcher.clone(),
query,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
fn add_bench_task_top100_asc(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query: RangeQuery,
collector_name: &str,
field_name: &str,
range_low: u64,
range_high: u64,
field_name_owned: String,
) {
let task_name = format!(
"range_{}_[{} TO {}]_{}",
field_name, range_low, range_high, collector_name
);
let search_task = Top100AscSearchTask {
searcher: bench_index.searcher.clone(),
query,
field_name: field_name_owned,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
fn add_bench_task_top100_desc(
bench_group: &mut BenchGroup,
bench_index: &BenchIndex,
query: RangeQuery,
collector_name: &str,
field_name: &str,
range_low: u64,
range_high: u64,
field_name_owned: String,
) {
let task_name = format!(
"range_{}_[{} TO {}]_{}",
field_name, range_low, range_high, collector_name
);
let search_task = Top100DescSearchTask {
searcher: bench_index.searcher.clone(),
query,
field_name: field_name_owned,
};
bench_group.register(task_name, move |_| black_box(search_task.run()));
}
struct CountSearchTask {
searcher: Searcher,
query: RangeQuery,
}
impl CountSearchTask {
#[inline(never)]
pub fn run(&self) -> usize {
self.searcher.search(&self.query, &Count).unwrap()
}
}
struct DocSetSearchTask {
searcher: Searcher,
query: RangeQuery,
}
impl DocSetSearchTask {
#[inline(never)]
pub fn run(&self) -> usize {
let result = self.searcher.search(&self.query, &DocSetCollector).unwrap();
result.len()
}
}
struct Top100AscSearchTask {
searcher: Searcher,
query: RangeQuery,
field_name: String,
}
impl Top100AscSearchTask {
#[inline(never)]
pub fn run(&self) -> usize {
let collector =
TopDocs::with_limit(100).order_by_fast_field::<u64>(&self.field_name, Order::Asc);
let result = self.searcher.search(&self.query, &collector).unwrap();
for (_score, doc_address) in &result {
let _doc: tantivy::TantivyDocument = self.searcher.doc(*doc_address).unwrap();
}
result.len()
}
}
struct Top100DescSearchTask {
searcher: Searcher,
query: RangeQuery,
field_name: String,
}
impl Top100DescSearchTask {
#[inline(never)]
pub fn run(&self) -> usize {
let collector =
TopDocs::with_limit(100).order_by_fast_field::<u64>(&self.field_name, Order::Desc);
let result = self.searcher.search(&self.query, &collector).unwrap();
for (_score, doc_address) in &result {
let _doc: tantivy::TantivyDocument = self.searcher.doc(*doc_address).unwrap();
}
result.len()
}
}

260
benches/range_query.rs
View File

@@ -1,260 +0,0 @@
use std::fmt::Display;
use std::net::Ipv6Addr;
use std::ops::RangeInclusive;
use binggan::plugins::PeakMemAllocPlugin;
use binggan::{black_box, BenchRunner, OutputValue, PeakMemAlloc, INSTRUMENTED_SYSTEM};
use columnar::MonotonicallyMappableToU128;
use rand::rngs::StdRng;
use rand::{Rng, SeedableRng};
use tantivy::collector::{Count, TopDocs};
use tantivy::query::QueryParser;
use tantivy::schema::*;
use tantivy::{doc, Index};
#[global_allocator]
pub static GLOBAL: &PeakMemAlloc<std::alloc::System> = &INSTRUMENTED_SYSTEM;
fn main() {
bench_range_query();
}
fn bench_range_query() {
let index = get_index_0_to_100();
let mut runner = BenchRunner::new();
runner.add_plugin(PeakMemAllocPlugin::new(GLOBAL));
runner.set_name("range_query on u64");
let field_name_and_descr: Vec<_> = vec![
("id", "Single Valued Range Field"),
("ids", "Multi Valued Range Field"),
];
let range_num_hits = vec![
("90_percent", get_90_percent()),
("10_percent", get_10_percent()),
("1_percent", get_1_percent()),
];
test_range(&mut runner, &index, &field_name_and_descr, range_num_hits);
runner.set_name("range_query on ip");
let field_name_and_descr: Vec<_> = vec![
("ip", "Single Valued Range Field"),
("ips", "Multi Valued Range Field"),
];
let range_num_hits = vec![
("90_percent", get_90_percent_ip()),
("10_percent", get_10_percent_ip()),
("1_percent", get_1_percent_ip()),
];
test_range(&mut runner, &index, &field_name_and_descr, range_num_hits);
}
fn test_range<T: Display>(
runner: &mut BenchRunner,
index: &Index,
field_name_and_descr: &[(&str, &str)],
range_num_hits: Vec<(&str, RangeInclusive<T>)>,
) {
for (field, suffix) in field_name_and_descr {
let term_num_hits = vec![
("", ""),
("1_percent", "veryfew"),
("10_percent", "few"),
("90_percent", "most"),
];
let mut group = runner.new_group();
group.set_name(suffix);
// all intersect combinations
for (range_name, range) in &range_num_hits {
for (term_name, term) in &term_num_hits {
let index = &index;
let test_name = if term_name.is_empty() {
format!("id_range_hit_{}", range_name)
} else {
format!(
"id_range_hit_{}_intersect_with_term_{}",
range_name, term_name
)
};
group.register(test_name, move |_| {
let query = if term_name.is_empty() {
"".to_string()
} else {
format!("AND id_name:{}", term)
};
black_box(execute_query(field, range, &query, index));
});
}
}
group.run();
}
}
fn get_index_0_to_100() -> Index {
let mut rng = StdRng::from_seed([1u8; 32]);
let num_vals = 100_000;
let docs: Vec<_> = (0..num_vals)
.map(|_i| {
let id_name = if rng.gen_bool(0.01) {
"veryfew".to_string() // 1%
} else if rng.gen_bool(0.1) {
"few".to_string() // 9%
} else {
"most".to_string() // 90%
};
Doc {
id_name,
id: rng.gen_range(0..100),
// Multiply by 1000, so that we create most buckets in the compact space
// The benches depend on this range to select n-percent of elements with the
// methods below.
ip: Ipv6Addr::from_u128(rng.gen_range(0..100) * 1000),
}
})
.collect();
create_index_from_docs(&docs)
}
#[derive(Clone, Debug)]
pub struct Doc {
pub id_name: String,
pub id: u64,
pub ip: Ipv6Addr,
}
pub fn create_index_from_docs(docs: &[Doc]) -> Index {
let mut schema_builder = Schema::builder();
let id_u64_field = schema_builder.add_u64_field("id", INDEXED | STORED | FAST);
let ids_u64_field =
schema_builder.add_u64_field("ids", NumericOptions::default().set_fast().set_indexed());
let id_f64_field = schema_builder.add_f64_field("id_f64", INDEXED | STORED | FAST);
let ids_f64_field = schema_builder.add_f64_field(
"ids_f64",
NumericOptions::default().set_fast().set_indexed(),
);
let id_i64_field = schema_builder.add_i64_field("id_i64", INDEXED | STORED | FAST);
let ids_i64_field = schema_builder.add_i64_field(
"ids_i64",
NumericOptions::default().set_fast().set_indexed(),
);
let text_field = schema_builder.add_text_field("id_name", STRING | STORED);
let text_field2 = schema_builder.add_text_field("id_name_fast", STRING | STORED | FAST);
let ip_field = schema_builder.add_ip_addr_field("ip", FAST);
let ips_field = schema_builder.add_ip_addr_field("ips", FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
{
let mut index_writer = index.writer_with_num_threads(1, 50_000_000).unwrap();
for doc in docs.iter() {
index_writer
.add_document(doc!(
ids_i64_field => doc.id as i64,
ids_i64_field => doc.id as i64,
ids_f64_field => doc.id as f64,
ids_f64_field => doc.id as f64,
ids_u64_field => doc.id,
ids_u64_field => doc.id,
id_u64_field => doc.id,
id_f64_field => doc.id as f64,
id_i64_field => doc.id as i64,
text_field => doc.id_name.to_string(),
text_field2 => doc.id_name.to_string(),
ips_field => doc.ip,
ips_field => doc.ip,
ip_field => doc.ip,
))
.unwrap();
}
index_writer.commit().unwrap();
}
index
}
fn get_90_percent() -> RangeInclusive<u64> {
0..=90
}
fn get_10_percent() -> RangeInclusive<u64> {
0..=10
}
fn get_1_percent() -> RangeInclusive<u64> {
10..=10
}
fn get_90_percent_ip() -> RangeInclusive<Ipv6Addr> {
let start = Ipv6Addr::from_u128(0);
let end = Ipv6Addr::from_u128(90 * 1000);
start..=end
}
fn get_10_percent_ip() -> RangeInclusive<Ipv6Addr> {
let start = Ipv6Addr::from_u128(0);
let end = Ipv6Addr::from_u128(10 * 1000);
start..=end
}
fn get_1_percent_ip() -> RangeInclusive<Ipv6Addr> {
let start = Ipv6Addr::from_u128(10 * 1000);
let end = Ipv6Addr::from_u128(10 * 1000);
start..=end
}
struct NumHits {
count: usize,
}
impl OutputValue for NumHits {
fn column_title() -> &'static str {
"NumHits"
}
fn format(&self) -> Option<String> {
Some(self.count.to_string())
}
}
fn execute_query<T: Display>(
field: &str,
id_range: &RangeInclusive<T>,
suffix: &str,
index: &Index,
) -> NumHits {
let gen_query_inclusive = |from: &T, to: &T| {
format!(
"{}:[{} TO {}] {}",
field,
&from.to_string(),
&to.to_string(),
suffix
)
};
let query = gen_query_inclusive(id_range.start(), id_range.end());
execute_query_(&query, index)
}
fn execute_query_(query: &str, index: &Index) -> NumHits {
let query_from_text = |text: &str| {
QueryParser::for_index(index, vec![])
.parse_query(text)
.unwrap()
};
let query = query_from_text(query);
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let num_hits = searcher
.search(&query, &(TopDocs::with_limit(10).order_by_score(), Count))
.unwrap()
.1;
NumHits { count: num_hits }
}

4
bitpacker/src/bitpacker.rs
View File

@@ -258,7 +258,7 @@ mod test {
bitpacker.write(val, num_bits, &mut data).unwrap();
}
bitpacker.close(&mut data).unwrap();
assert_eq!(data.len(), ((num_bits as usize) * len).div_ceil(8));
assert_eq!(data.len(), ((num_bits as usize) * len + 7) / 8);
let bitunpacker = BitUnpacker::new(num_bits);
(bitunpacker, vals, data)
}
@@ -304,7 +304,7 @@ mod test {
bitpacker.write(val, num_bits, &mut buffer).unwrap();
}
bitpacker.flush(&mut buffer).unwrap();
assert_eq!(buffer.len(), (vals.len() * num_bits as usize).div_ceil(8));
assert_eq!(buffer.len(), (vals.len() * num_bits as usize + 7) / 8);
let bitunpacker = BitUnpacker::new(num_bits);
let max_val = if num_bits == 64 {
u64::MAX

27
bitpacker/src/filter_vec/avx2.rs
View File

@@ -19,7 +19,7 @@ fn u32_to_i32(val: u32) -> i32 {
#[inline]
unsafe fn u32_to_i32_avx2(vals_u32x8s: DataType) -> DataType {
const HIGHEST_BIT_MASK: DataType = from_u32x8([HIGHEST_BIT; NUM_LANES]);
unsafe { op_xor(vals_u32x8s, HIGHEST_BIT_MASK) }
op_xor(vals_u32x8s, HIGHEST_BIT_MASK)
}
pub fn filter_vec_in_place(range: RangeInclusive<u32>, offset: u32, output: &mut Vec<u32>) {
@@ -66,19 +66,17 @@ unsafe fn filter_vec_avx2_aux(
]);
const SHIFT: __m256i = from_u32x8([NUM_LANES as u32; NUM_LANES]);
for _ in 0..num_words {
unsafe {
let word = load_unaligned(input);
let word = u32_to_i32_avx2(word);
let keeper_bitset = compute_filter_bitset(word, range_simd.clone());
let added_len = keeper_bitset.count_ones();
let filtered_doc_ids = compact(ids, keeper_bitset);
store_unaligned(output_tail as *mut __m256i, filtered_doc_ids);
output_tail = output_tail.offset(added_len as isize);
ids = op_add(ids, SHIFT);
input = input.offset(1);
}
let word = load_unaligned(input);
let word = u32_to_i32_avx2(word);
let keeper_bitset = compute_filter_bitset(word, range_simd.clone());
let added_len = keeper_bitset.count_ones();
let filtered_doc_ids = compact(ids, keeper_bitset);
store_unaligned(output_tail as *mut __m256i, filtered_doc_ids);
output_tail = output_tail.offset(added_len as isize);
ids = op_add(ids, SHIFT);
input = input.offset(1);
}
unsafe { output_tail.offset_from(output) as usize }
output_tail.offset_from(output) as usize
}
#[inline]
@@ -94,7 +92,8 @@ unsafe fn compute_filter_bitset(val: __m256i, range: std::ops::RangeInclusive<__
let too_low = op_greater(*range.start(), val);
let too_high = op_greater(val, *range.end());
let inside = op_or(too_low, too_high);
255 - std::arch::x86_64::_mm256_movemask_ps(_mm256_castsi256_ps(inside)) as u8
255 - std::arch::x86_64::_mm256_movemask_ps(std::mem::transmute::<DataType, __m256>(inside))
as u8
}
union U8x32 {

2
columnar/README.md
View File

@@ -73,7 +73,7 @@ The crate introduces the following concepts.
`Columnar` is an equivalent of a dataframe.
It maps `column_key` to `Column`.
A `Column<T>` associates a `RowId` (u32) to any
A `Column<T>` asssociates a `RowId` (u32) to any
number of values.
This is made possible by wrapping a `ColumnIndex` and a `ColumnValue` object.

7
columnar/benches/bench_first_vals.rs
View File

@@ -89,6 +89,13 @@ fn main() {
black_box(sum);
});
group.register("first_block_fetch", |column| {
let mut block: Vec<Option<u64>> = vec![None; 64];
let fetch_docids = (0..64).collect::<Vec<_>>();
column.first_vals(&fetch_docids, &mut block);
black_box(block[0]);
});
group.register("first_block_single_calls", |column| {
let mut block: Vec<Option<u64>> = vec![None; 64];
let fetch_docids = (0..64).collect::<Vec<_>>();

11
columnar/src/column/mod.rs
View File

@@ -131,8 +131,6 @@ impl<T: PartialOrd + Copy + Debug + Send + Sync + 'static> Column<T> {
self.index.docids_to_rowids(doc_ids, doc_ids_out, row_ids)
}
/// Get an iterator over the values for the provided docid.
#[inline]
pub fn values_for_doc(&self, doc_id: DocId) -> impl Iterator<Item = T> + '_ {
self.index
.value_row_ids(doc_id)
@@ -160,6 +158,15 @@ impl<T: PartialOrd + Copy + Debug + Send + Sync + 'static> Column<T> {
.select_batch_in_place(selected_docid_range.start, doc_ids);
}
/// Fills the output vector with the (possibly multiple values that are associated_with
/// `row_id`.
///
/// This method clears the `output` vector.
pub fn fill_vals(&self, row_id: RowId, output: &mut Vec<T>) {
output.clear();
output.extend(self.values_for_doc(row_id));
}
pub fn first_or_default_col(self, default_value: T) -> Arc<dyn ColumnValues<T>> {
Arc::new(FirstValueWithDefault {
column: self,

2
columnar/src/column_values/monotonic_mapping_u128.rs
View File

@@ -1,7 +1,7 @@
use std::fmt::Debug;
use std::net::Ipv6Addr;
/// Monotonic maps a value to u128 value space
/// 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 + Debug + Send + Sync {
/// Converts a value to u128.

6
columnar/src/column_values/u64_based/bitpacked.rs
View File

@@ -41,6 +41,12 @@ fn transform_range_before_linear_transformation(
if range.is_empty() {
return None;
}
if stats.min_value > *range.end() {
return None;
}
if stats.max_value < *range.start() {
return None;
}
let shifted_range =
range.start().saturating_sub(stats.min_value)..=range.end().saturating_sub(stats.min_value);
let start_before_gcd_multiplication: u64 = div_ceil(*shifted_range.start(), stats.gcd);

4
columnar/src/column_values/u64_based/line.rs
View File

@@ -8,7 +8,7 @@ use crate::column_values::ColumnValues;
const MID_POINT: u64 = (1u64 << 32) - 1u64;
/// `Line` describes a line function `y: ax + b` using integer
/// arithmetic.
/// arithmetics.
///
/// The slope is in fact a decimal split into a 32 bit integer value,
/// and a 32-bit decimal value.
@@ -94,7 +94,7 @@ impl Line {
// `(i, ys[])`.
//
// The best intercept therefore has the form
// `y[i] - line.eval(i)` (using wrapping arithmetic).
// `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.
//

2
columnar/src/column_values/u64_based/mod.rs
View File

@@ -52,7 +52,7 @@ pub trait ColumnCodecEstimator<T = u64>: 'static {
) -> io::Result<()>;
}
/// A column codec describes a column serialization format.
/// A column codec describes a colunm serialization format.
pub trait ColumnCodec<T: PartialOrd = u64> {
/// Specialized `ColumnValues` type.
type ColumnValues: ColumnValues<T> + 'static;

84
columnar/src/dynamic_column.rs
View File

@@ -3,8 +3,7 @@ use std::sync::Arc;
use std::{fmt, io};
use common::file_slice::FileSlice;
use common::{ByteCount, DateTime, OwnedBytes};
use serde::{Deserialize, Serialize};
use common::{ByteCount, DateTime, HasLen, OwnedBytes};
use crate::column::{BytesColumn, Column, StrColumn};
use crate::column_values::{StrictlyMonotonicFn, monotonic_map_column};
@@ -318,89 +317,10 @@ impl DynamicColumnHandle {
}
pub fn num_bytes(&self) -> ByteCount {
self.file_slice.num_bytes()
}
/// Legacy helper returning the column space usage.
pub fn column_and_dictionary_num_bytes(&self) -> io::Result<ColumnSpaceUsage> {
self.space_usage()
}
/// Return the space usage of the column, optionally broken down by dictionary and column
/// values.
///
/// For dictionary encoded columns (strings and bytes), this splits the total footprint into
/// the dictionary and the remaining column data (including index and values).
/// For all other column types, the dictionary size is `None` and the column size
/// equals the total bytes.
pub fn space_usage(&self) -> io::Result<ColumnSpaceUsage> {
let total_num_bytes = self.num_bytes();
let dynamic_column = self.open()?;
let dictionary_num_bytes = match &dynamic_column {
DynamicColumn::Bytes(bytes_column) => bytes_column.dictionary().num_bytes(),
DynamicColumn::Str(str_column) => str_column.dictionary().num_bytes(),
_ => {
return Ok(ColumnSpaceUsage::new(self.num_bytes(), None));
}
};
assert!(dictionary_num_bytes <= total_num_bytes);
let column_num_bytes =
ByteCount::from(total_num_bytes.get_bytes() - dictionary_num_bytes.get_bytes());
Ok(ColumnSpaceUsage::new(
column_num_bytes,
Some(dictionary_num_bytes),
))
self.file_slice.len().into()
}
pub fn column_type(&self) -> ColumnType {
self.column_type
}
}
/// Represents space usage of a column.
///
/// `column_num_bytes` tracks the column payload (index, values and footer).
/// For dictionary encoded columns, `dictionary_num_bytes` captures the dictionary footprint.
/// [`ColumnSpaceUsage::total_num_bytes`] returns the sum of both parts.
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct ColumnSpaceUsage {
column_num_bytes: ByteCount,
dictionary_num_bytes: Option<ByteCount>,
}
impl ColumnSpaceUsage {
pub(crate) fn new(
column_num_bytes: ByteCount,
dictionary_num_bytes: Option<ByteCount>,
) -> Self {
ColumnSpaceUsage {
column_num_bytes,
dictionary_num_bytes,
}
}
pub fn column_num_bytes(&self) -> ByteCount {
self.column_num_bytes
}
pub fn dictionary_num_bytes(&self) -> Option<ByteCount> {
self.dictionary_num_bytes
}
pub fn total_num_bytes(&self) -> ByteCount {
self.column_num_bytes + self.dictionary_num_bytes.unwrap_or_default()
}
/// Merge two space usage values by summing their components.
pub fn merge(&self, other: &ColumnSpaceUsage) -> ColumnSpaceUsage {
let dictionary_num_bytes = match (self.dictionary_num_bytes, other.dictionary_num_bytes) {
(Some(lhs), Some(rhs)) => Some(lhs + rhs),
(Some(val), None) | (None, Some(val)) => Some(val),
(None, None) => None,
};
ColumnSpaceUsage {
column_num_bytes: self.column_num_bytes + other.column_num_bytes,
dictionary_num_bytes,
}
}
}

2
columnar/src/lib.rs
View File

@@ -48,7 +48,7 @@ pub use columnar::{
use sstable::VoidSSTable;
pub use value::{NumericalType, NumericalValue};
pub use self::dynamic_column::{ColumnSpaceUsage, DynamicColumn, DynamicColumnHandle};
pub use self::dynamic_column::{DynamicColumn, DynamicColumnHandle};
pub type RowId = u32;
pub type DocId = u32;

4
common/src/vint.rs
View File

@@ -28,7 +28,6 @@ impl BinarySerializable for VIntU128 {
writer.write_all(&buffer)
}
#[allow(clippy::unbuffered_bytes)]
fn deserialize<R: Read>(reader: &mut R) -> io::Result<Self> {
#[allow(clippy::unbuffered_bytes)]
let mut bytes = reader.bytes();
@@ -54,7 +53,7 @@ impl BinarySerializable for VIntU128 {
}
}
/// Wrapper over a `u64` that serializes as a variable int.
/// Wrapper over a `u64` that serializes as a variable int.
#[derive(Clone, Copy, Debug, Eq, PartialEq)]
pub struct VInt(pub u64);
@@ -197,7 +196,6 @@ impl BinarySerializable for VInt {
writer.write_all(&buffer[0..num_bytes])
}
#[allow(clippy::unbuffered_bytes)]
fn deserialize<R: Read>(reader: &mut R) -> io::Result<Self> {
#[allow(clippy::unbuffered_bytes)]
let mut bytes = reader.bytes();

BIN
doc/assets/images/searchbenchmark.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 653 KiB

4
examples/basic_search.rs
View File

@@ -208,7 +208,7 @@ fn main() -> tantivy::Result<()> {
// is the role of the `TopDocs` collector.
// We can now perform our query.
let top_docs = searcher.search(&query, &TopDocs::with_limit(10).order_by_score())?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
// The actual documents still need to be
// retrieved from Tantivy's store.
@@ -226,7 +226,7 @@ fn main() -> tantivy::Result<()> {
let query = query_parser.parse_query("title:sea^20 body:whale^70")?;
let (_score, doc_address) = searcher
.search(&query, &TopDocs::with_limit(1).order_by_score())?
.search(&query, &TopDocs::with_limit(1))?
.into_iter()
.next()
.unwrap();

2
examples/custom_tokenizer.rs
View File

@@ -100,7 +100,7 @@ fn main() -> tantivy::Result<()> {
// here we want to get a hit on the 'ken' in Frankenstein
let query = query_parser.parse_query("ken")?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10).order_by_score())?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
for (_, doc_address) in top_docs {
let retrieved_doc: TantivyDocument = searcher.doc(doc_address)?;

4
examples/date_time_field.rs
View File

@@ -50,14 +50,14 @@ fn main() -> tantivy::Result<()> {
{
// Simple exact search on the date
let query = query_parser.parse_query("occurred_at:\"2022-06-22T12:53:50.53Z\"")?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(5).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(5))?;
assert_eq!(count_docs.len(), 1);
}
{
// Range query on the date field
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).order_by_score())?;
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::<TantivyDocument>(doc_address)?;

2
examples/deleting_updating_documents.rs
View File

@@ -28,7 +28,7 @@ fn extract_doc_given_isbn(
// The second argument is here to tell we don't care about decoding positions,
// or term frequencies.
let term_query = TermQuery::new(isbn_term.clone(), IndexRecordOption::Basic);
let top_docs = searcher.search(&term_query, &TopDocs::with_limit(1).order_by_score())?;
let top_docs = searcher.search(&term_query, &TopDocs::with_limit(1))?;
if let Some((_score, doc_address)) = top_docs.first() {
let doc = searcher.doc(*doc_address)?;

212
examples/filter_aggregation.rs
View File

@@ -1,212 +0,0 @@
// # Filter Aggregation Example
//
// This example demonstrates filter aggregations - creating buckets of documents
// matching specific queries, with nested aggregations computed on each bucket.
//
// Filter aggregations are useful for computing metrics on different subsets of
// your data in a single query, like "average price overall + average price for
// electronics + count of in-stock items".
use serde_json::json;
use tantivy::aggregation::agg_req::Aggregations;
use tantivy::aggregation::AggregationCollector;
use tantivy::query::AllQuery;
use tantivy::schema::{Schema, FAST, INDEXED, TEXT};
use tantivy::{doc, Index};
fn main() -> tantivy::Result<()> {
// Create a simple product schema
let mut schema_builder = Schema::builder();
schema_builder.add_text_field("category", TEXT | FAST);
schema_builder.add_text_field("brand", TEXT | FAST);
schema_builder.add_u64_field("price", FAST);
schema_builder.add_f64_field("rating", FAST);
schema_builder.add_bool_field("in_stock", FAST | INDEXED);
let schema = schema_builder.build();
// Create index and add sample products
let index = Index::create_in_ram(schema.clone());
let mut writer = index.writer(50_000_000)?;
writer.add_document(doc!(
schema.get_field("category")? => "electronics",
schema.get_field("brand")? => "apple",
schema.get_field("price")? => 999u64,
schema.get_field("rating")? => 4.5f64,
schema.get_field("in_stock")? => true
))?;
writer.add_document(doc!(
schema.get_field("category")? => "electronics",
schema.get_field("brand")? => "samsung",
schema.get_field("price")? => 799u64,
schema.get_field("rating")? => 4.2f64,
schema.get_field("in_stock")? => true
))?;
writer.add_document(doc!(
schema.get_field("category")? => "clothing",
schema.get_field("brand")? => "nike",
schema.get_field("price")? => 120u64,
schema.get_field("rating")? => 4.1f64,
schema.get_field("in_stock")? => false
))?;
writer.add_document(doc!(
schema.get_field("category")? => "books",
schema.get_field("brand")? => "penguin",
schema.get_field("price")? => 25u64,
schema.get_field("rating")? => 4.8f64,
schema.get_field("in_stock")? => true
))?;
writer.commit()?;
let reader = index.reader()?;
let searcher = reader.searcher();
// Example 1: Basic filter with metric aggregation
println!("=== Example 1: Electronics average price ===");
let agg_req = json!({
"electronics": {
"filter": "category:electronics",
"aggs": {
"avg_price": { "avg": { "field": "price" } }
}
}
});
let agg: Aggregations = serde_json::from_value(agg_req)?;
let collector = AggregationCollector::from_aggs(agg, Default::default());
let result = searcher.search(&AllQuery, &collector)?;
let expected = json!({
"electronics": {
"doc_count": 2,
"avg_price": { "value": 899.0 }
}
});
assert_eq!(serde_json::to_value(&result)?, expected);
println!("{}\n", serde_json::to_string_pretty(&result)?);
// Example 2: Multiple independent filters
println!("=== Example 2: Multiple filters in one query ===");
let agg_req = json!({
"electronics": {
"filter": "category:electronics",
"aggs": { "avg_price": { "avg": { "field": "price" } } }
},
"in_stock": {
"filter": "in_stock:true",
"aggs": { "count": { "value_count": { "field": "brand" } } }
},
"high_rated": {
"filter": "rating:[4.5 TO *]",
"aggs": { "count": { "value_count": { "field": "brand" } } }
}
});
let agg: Aggregations = serde_json::from_value(agg_req)?;
let collector = AggregationCollector::from_aggs(agg, Default::default());
let result = searcher.search(&AllQuery, &collector)?;
let expected = json!({
"electronics": {
"doc_count": 2,
"avg_price": { "value": 899.0 }
},
"in_stock": {
"doc_count": 3,
"count": { "value": 3.0 }
},
"high_rated": {
"doc_count": 2,
"count": { "value": 2.0 }
}
});
assert_eq!(serde_json::to_value(&result)?, expected);
println!("{}\n", serde_json::to_string_pretty(&result)?);
// Example 3: Nested filters - progressive refinement
println!("=== Example 3: Nested filters ===");
let agg_req = json!({
"in_stock": {
"filter": "in_stock:true",
"aggs": {
"electronics": {
"filter": "category:electronics",
"aggs": {
"expensive": {
"filter": "price:[800 TO *]",
"aggs": {
"avg_rating": { "avg": { "field": "rating" } }
}
}
}
}
}
}
});
let agg: Aggregations = serde_json::from_value(agg_req)?;
let collector = AggregationCollector::from_aggs(agg, Default::default());
let result = searcher.search(&AllQuery, &collector)?;
let expected = json!({
"in_stock": {
"doc_count": 3, // apple, samsung, penguin
"electronics": {
"doc_count": 2, // apple, samsung
"expensive": {
"doc_count": 1, // only apple (999)
"avg_rating": { "value": 4.5 }
}
}
}
});
assert_eq!(serde_json::to_value(&result)?, expected);
println!("{}\n", serde_json::to_string_pretty(&result)?);
// Example 4: Filter with sub-aggregation (terms)
println!("=== Example 4: Filter with terms sub-aggregation ===");
let agg_req = json!({
"electronics": {
"filter": "category:electronics",
"aggs": {
"by_brand": {
"terms": { "field": "brand" },
"aggs": {
"avg_price": { "avg": { "field": "price" } }
}
}
}
}
});
let agg: Aggregations = serde_json::from_value(agg_req)?;
let collector = AggregationCollector::from_aggs(agg, Default::default());
let result = searcher.search(&AllQuery, &collector)?;
let expected = json!({
"electronics": {
"doc_count": 2,
"by_brand": {
"buckets": [
{
"key": "samsung",
"doc_count": 1,
"avg_price": { "value": 799.0 }
},
{
"key": "apple",
"doc_count": 1,
"avg_price": { "value": 999.0 }
}
],
"sum_other_doc_count": 0,
"doc_count_error_upper_bound": 0
}
}
});
assert_eq!(serde_json::to_value(&result)?, expected);
println!("{}", serde_json::to_string_pretty(&result)?);
Ok(())
}

3
examples/fuzzy_search.rs
View File

@@ -85,6 +85,7 @@ fn main() -> tantivy::Result<()> {
index_writer.add_document(doc!(
title => "The Diary of a Young Girl",
))?;
index_writer.commit()?;
// ### Committing
//
@@ -145,7 +146,7 @@ fn main() -> tantivy::Result<()> {
let query = FuzzyTermQuery::new(term, 2, true);
let (top_docs, count) = searcher
.search(&query, &(TopDocs::with_limit(5).order_by_score(), Count))
.search(&query, &(TopDocs::with_limit(5), Count))
.unwrap();
assert_eq!(count, 3);
assert_eq!(top_docs.len(), 3);

8
examples/ip_field.rs
View File

@@ -69,25 +69,25 @@ fn main() -> tantivy::Result<()> {
{
// Inclusive range queries
let query = query_parser.parse_query("ip:[192.168.0.80 TO 192.168.0.100]")?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(5).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(5))?;
assert_eq!(count_docs.len(), 1);
}
{
// Exclusive range queries
let query = query_parser.parse_query("ip:{192.168.0.80 TO 192.168.1.100]")?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
assert_eq!(count_docs.len(), 0);
}
{
// Find docs with IP addresses smaller equal 192.168.1.100
let query = query_parser.parse_query("ip:[* TO 192.168.1.100]")?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
assert_eq!(count_docs.len(), 2);
}
{
// Find docs with IP addresses smaller than 192.168.1.100
let query = query_parser.parse_query("ip:[* TO 192.168.1.100}")?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
assert_eq!(count_docs.len(), 2);
}

14
examples/json_field.rs
View File

@@ -59,12 +59,12 @@ fn main() -> tantivy::Result<()> {
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).order_by_score())?;
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).order_by_score())?;
let count_docs = searcher.search(&*query, &TopDocs::with_limit(2))?;
assert_eq!(count_docs.len(), 2);
}
{
@@ -74,33 +74,33 @@ fn main() -> tantivy::Result<()> {
}
{
let query = query_parser.parse_query("click AND cart.product_id:133")?;
let hits = searcher.search(&*query, &TopDocs::with_limit(2).order_by_score())?;
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).order_by_score())?;
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).order_by_score())?;
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).order_by_score())?;
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).order_by_score())?;
let hits = searcher.search(&*query, &TopDocs::with_limit(2))?;
assert_eq!(hits.len(), 1);
}
Ok(())

2
examples/phrase_prefix_search.rs
View File

@@ -63,7 +63,7 @@ fn main() -> Result<()> {
// but not "in the Gulf Stream".
let query = query_parser.parse_query("\"in the su\"*")?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10).order_by_score())?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
let mut titles = top_docs
.into_iter()
.map(|(_score, doc_address)| {

6
examples/pre_tokenized_text.rs
View File

@@ -107,8 +107,7 @@ fn main() -> tantivy::Result<()> {
IndexRecordOption::Basic,
);
let (top_docs, count) =
searcher.search(&query, &(TopDocs::with_limit(2).order_by_score(), Count))?;
let (top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count))?;
assert_eq!(count, 2);
@@ -129,8 +128,7 @@ fn main() -> tantivy::Result<()> {
IndexRecordOption::Basic,
);
let (_top_docs, count) =
searcher.search(&query, &(TopDocs::with_limit(2).order_by_score(), Count))?;
let (_top_docs, count) = searcher.search(&query, &(TopDocs::with_limit(2), Count))?;
assert_eq!(count, 0);

2
examples/snippet.rs
View File

@@ -50,7 +50,7 @@ fn main() -> tantivy::Result<()> {
let query_parser = QueryParser::for_index(&index, vec![title, body]);
let query = query_parser.parse_query("sycamore spring")?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10).order_by_score())?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
let snippet_generator = SnippetGenerator::create(&searcher, &*query, body)?;

2
examples/stop_words.rs
View File

@@ -102,7 +102,7 @@ fn main() -> tantivy::Result<()> {
// stop words are applied on the query as well.
// The following will be equivalent to `title:frankenstein`
let query = query_parser.parse_query("title:\"the Frankenstein\"")?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10).order_by_score())?;
let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
for (score, doc_address) in top_docs {
let retrieved_doc: TantivyDocument = searcher.doc(doc_address)?;

2
examples/warmer.rs
View File

@@ -164,7 +164,7 @@ fn main() -> tantivy::Result<()> {
move |doc_id: DocId| Reverse(price[doc_id as usize])
};
let most_expensive_first = TopDocs::with_limit(10).order_by(score_by_price);
let most_expensive_first = TopDocs::with_limit(10).custom_score(score_by_price);
let hits = searcher.search(&query, &most_expensive_first)?;
assert_eq!(

39
query-grammar/src/query_grammar.rs
View File

@@ -758,17 +758,7 @@ fn negate(expr: UserInputAst) -> UserInputAst {
fn leaf(inp: &str) -> IResult<&str, UserInputAst> {
alt((
delimited(char('('), ast, char(')')),
map(
terminated(
char('*'),
peek(alt((
value((), multispace1),
value((), char(')')),
value((), eof),
))),
),
|_| UserInputAst::from(UserInputLeaf::All),
),
map(char('*'), |_| UserInputAst::from(UserInputLeaf::All)),
map(preceded(tuple((tag("NOT"), multispace1)), leaf), negate),
literal,
))(inp)
@@ -789,17 +779,7 @@ fn leaf_infallible(inp: &str) -> JResult<&str, Option<UserInputAst>> {
),
),
(
value(
(),
terminated(
char('*'),
peek(alt((
value((), multispace1),
value((), char(')')),
value((), eof),
))),
),
),
value((), char('*')),
map(nothing, |_| {
(Some(UserInputAst::from(UserInputLeaf::All)), Vec::new())
}),
@@ -1691,21 +1671,6 @@ mod test {
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\"]");
// Phrase prefixed with *
test_parse_query_to_ast_helper("foo:(*A)", "\"foo\":*A");
test_parse_query_to_ast_helper("*A", "*A");
test_parse_query_to_ast_helper("(*A)", "*A");
test_parse_query_to_ast_helper("foo:(A OR B)", "(?\"foo\":A ?\"foo\":B)");
test_parse_query_to_ast_helper("foo:(A* OR B*)", "(?\"foo\":A* ?\"foo\":B*)");
test_parse_query_to_ast_helper("foo:(*A OR *B)", "(?\"foo\":*A ?\"foo\":*B)");
}
#[test]
fn test_parse_query_all() {
test_parse_query_to_ast_helper("*", "*");
test_parse_query_to_ast_helper("(*)", "*");
test_parse_query_to_ast_helper("(* )", "*");
}
#[test]

7
src/aggregation/README.md
View File

@@ -20,16 +20,17 @@ Contains all metric aggregations, like average aggregation. Metric aggregations
#### agg_req
agg_req contains the users aggregation request. Deserialization from json is compatible with elasticsearch aggregation requests.
#### agg_data
agg_data contains the users aggregation request enriched with fast field accessors etc, which are
#### agg_req_with_accessor
agg_req_with_accessor contains the users aggregation request enriched with fast field accessors etc, which are
used during collection.
#### segment_agg_result
segment_agg_result contains the aggregation result tree, which is used for collection of a segment.
agg_data is passed during collection.
The tree from agg_req_with_accessor is passed during collection.
#### intermediate_agg_result
intermediate_agg_result contains the aggregation tree for merging with other trees.
#### agg_result
agg_result contains the final aggregation tree.

105
src/aggregation/accessor_helpers.rs
View File

@@ -1,105 +0,0 @@
//! This will enhance the request tree with access to the fastfield and metadata.
use std::io;
use columnar::{Column, ColumnType};
use crate::aggregation::{f64_to_fastfield_u64, Key};
use crate::index::SegmentReader;
/// Get the missing value as internal u64 representation
///
/// For terms we use u64::MAX as sentinel value
/// For numerical data we convert the value into the representation
/// we would get from the fast field, when we open it as u64_lenient_for_type.
///
/// That way we can use it the same way as if it would come from the fastfield.
pub(crate) fn get_missing_val_as_u64_lenient(
column_type: ColumnType,
column_max_value: u64,
missing: &Key,
field_name: &str,
) -> crate::Result<Option<u64>> {
let missing_val = match missing {
Key::Str(_) if column_type == ColumnType::Str => Some(column_max_value + 1),
// Allow fallback to number on text fields
Key::F64(_) if column_type == ColumnType::Str => Some(column_max_value + 1),
Key::U64(_) if column_type == ColumnType::Str => Some(column_max_value + 1),
Key::I64(_) if column_type == ColumnType::Str => Some(column_max_value + 1),
Key::F64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val, &column_type)
}
// NOTE: We may loose precision of the passed missing value by casting i64 and u64 to f64.
Key::I64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val as f64, &column_type)
}
Key::U64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val as f64, &column_type)
}
_ => {
return Err(crate::TantivyError::InvalidArgument(format!(
"Missing value {missing:?} for field {field_name} is not supported for column \
type {column_type:?}"
)));
}
};
Ok(missing_val)
}
pub(crate) fn get_numeric_or_date_column_types() -> &'static [ColumnType] {
&[
ColumnType::F64,
ColumnType::U64,
ColumnType::I64,
ColumnType::DateTime,
]
}
/// Get fast field reader or empty as default.
pub(crate) fn get_ff_reader(
reader: &SegmentReader,
field_name: &str,
allowed_column_types: Option<&[ColumnType]>,
) -> crate::Result<(columnar::Column<u64>, ColumnType)> {
let ff_fields = reader.fast_fields();
let ff_field_with_type = ff_fields
.u64_lenient_for_type(allowed_column_types, field_name)?
.unwrap_or_else(|| {
(
Column::build_empty_column(reader.num_docs()),
ColumnType::U64,
)
});
Ok(ff_field_with_type)
}
pub(crate) fn get_dynamic_columns(
reader: &SegmentReader,
field_name: &str,
) -> crate::Result<Vec<columnar::DynamicColumn>> {
let ff_fields = reader.fast_fields().dynamic_column_handles(field_name)?;
let cols = ff_fields
.iter()
.map(|h| h.open())
.collect::<io::Result<_>>()?;
assert!(!ff_fields.is_empty(), "field {field_name} not found");
Ok(cols)
}
/// Get all fast field reader or empty as default.
///
/// Is guaranteed to return at least one column.
pub(crate) fn get_all_ff_reader_or_empty(
reader: &SegmentReader,
field_name: &str,
allowed_column_types: Option<&[ColumnType]>,
fallback_type: ColumnType,
) -> crate::Result<Vec<(columnar::Column<u64>, ColumnType)>> {
let ff_fields = reader.fast_fields();
let mut ff_field_with_type =
ff_fields.u64_lenient_for_type_all(allowed_column_types, field_name)?;
if ff_field_with_type.is_empty() {
ff_field_with_type.push((Column::build_empty_column(reader.num_docs()), fallback_type));
}
Ok(ff_field_with_type)
}

1095
src/aggregation/agg_data.rs
View File

File diff suppressed because it is too large Load Diff

3
src/aggregation/agg_limits.rs
View File

@@ -35,7 +35,6 @@ pub struct AggregationLimitsGuard {
/// Allocated memory with this guard.
allocated_with_the_guard: u64,
}
impl Clone for AggregationLimitsGuard {
fn clone(&self) -> Self {
Self {
@@ -71,7 +70,7 @@ impl AggregationLimitsGuard {
/// *memory_limit*
/// memory_limit is defined in bytes.
/// Aggregation fails when the estimated memory consumption of the aggregation is higher than
/// memory_limit.
/// memory_limit.
/// memory_limit will default to `DEFAULT_MEMORY_LIMIT` (500MB)
///
/// *bucket_limit*

19
src/aggregation/agg_req.rs
View File

@@ -26,14 +26,12 @@
//! let _agg_req: Aggregations = serde_json::from_str(elasticsearch_compatible_json_req).unwrap();
//! ```
use std::collections::HashSet;
use std::collections::{HashMap, HashSet};
use rustc_hash::FxHashMap;
use serde::{Deserialize, Serialize};
use super::bucket::{
DateHistogramAggregationReq, FilterAggregation, HistogramAggregation, RangeAggregation,
TermsAggregation,
DateHistogramAggregationReq, HistogramAggregation, RangeAggregation, TermsAggregation,
};
use super::metric::{
AverageAggregation, CardinalityAggregationReq, CountAggregation, ExtendedStatsAggregation,
@@ -45,7 +43,7 @@ use super::metric::{
/// defined names. It is also used in buckets aggregations to define sub-aggregations.
///
/// The key is the user defined name of the aggregation.
pub type Aggregations = FxHashMap<String, Aggregation>;
pub type Aggregations = HashMap<String, Aggregation>;
/// Aggregation request.
///
@@ -131,9 +129,6 @@ pub enum AggregationVariants {
/// Put data into buckets of terms.
#[serde(rename = "terms")]
Terms(TermsAggregation),
/// Filter documents into a single bucket.
#[serde(rename = "filter")]
Filter(FilterAggregation),
// Metric aggregation types
/// Computes the average of the extracted values.
@@ -179,7 +174,6 @@ impl AggregationVariants {
AggregationVariants::Range(range) => vec![range.field.as_str()],
AggregationVariants::Histogram(histogram) => vec![histogram.field.as_str()],
AggregationVariants::DateHistogram(histogram) => vec![histogram.field.as_str()],
AggregationVariants::Filter(filter) => filter.get_fast_field_names(),
AggregationVariants::Average(avg) => vec![avg.field_name()],
AggregationVariants::Count(count) => vec![count.field_name()],
AggregationVariants::Max(max) => vec![max.field_name()],
@@ -214,6 +208,13 @@ impl AggregationVariants {
_ => None,
}
}
pub(crate) fn as_top_hits(&self) -> Option<&TopHitsAggregationReq> {
match &self {
AggregationVariants::TopHits(top_hits) => Some(top_hits),
_ => None,
}
}
pub(crate) fn as_percentile(&self) -> Option<&PercentilesAggregationReq> {
match &self {
AggregationVariants::Percentiles(percentile_req) => Some(percentile_req),

471
src/aggregation/agg_req_with_accessor.rs Normal file
View File

@@ -0,0 +1,471 @@
//! This will enhance the request tree with access to the fastfield and metadata.
use std::collections::HashMap;
use std::io;
use columnar::{Column, ColumnBlockAccessor, ColumnType, DynamicColumn, StrColumn};
use super::agg_req::{Aggregation, AggregationVariants, Aggregations};
use super::bucket::{
DateHistogramAggregationReq, HistogramAggregation, RangeAggregation, TermsAggregation,
};
use super::metric::{
AverageAggregation, CardinalityAggregationReq, CountAggregation, ExtendedStatsAggregation,
MaxAggregation, MinAggregation, StatsAggregation, SumAggregation,
};
use super::segment_agg_result::AggregationLimitsGuard;
use super::VecWithNames;
use crate::aggregation::{f64_to_fastfield_u64, Key};
use crate::index::SegmentReader;
use crate::SegmentOrdinal;
#[derive(Default)]
pub(crate) struct AggregationsWithAccessor {
pub aggs: VecWithNames<AggregationWithAccessor>,
}
impl AggregationsWithAccessor {
fn from_data(aggs: VecWithNames<AggregationWithAccessor>) -> Self {
Self { aggs }
}
pub fn is_empty(&self) -> bool {
self.aggs.is_empty()
}
}
pub struct AggregationWithAccessor {
pub(crate) segment_ordinal: SegmentOrdinal,
/// In general there can be buckets without fast field access, e.g. buckets that are created
/// based on search terms. That is not that case currently, but eventually this needs to be
/// Option or moved.
pub(crate) accessor: Column<u64>,
/// Load insert u64 for missing use case
pub(crate) missing_value_for_accessor: Option<u64>,
pub(crate) str_dict_column: Option<StrColumn>,
pub(crate) field_type: ColumnType,
pub(crate) sub_aggregation: AggregationsWithAccessor,
pub(crate) limits: AggregationLimitsGuard,
pub(crate) column_block_accessor: ColumnBlockAccessor<u64>,
/// Used for missing term aggregation, which checks all columns for existence.
/// And also for `top_hits` aggregation, which may sort on multiple fields.
/// By convention the missing aggregation is chosen, when this property is set
/// (instead bein set in `agg`).
/// If this needs to used by other aggregations, we need to refactor this.
// NOTE: we can make all other aggregations use this instead of the `accessor` and `field_type`
// (making them obsolete) But will it have a performance impact?
pub(crate) accessors: Vec<(Column<u64>, ColumnType)>,
/// Map field names to all associated column accessors.
/// This field is used for `docvalue_fields`, which is currently only supported for `top_hits`.
pub(crate) value_accessors: HashMap<String, Vec<DynamicColumn>>,
pub(crate) agg: Aggregation,
}
impl AggregationWithAccessor {
/// May return multiple accessors if the aggregation is e.g. on mixed field types.
fn try_from_agg(
agg: &Aggregation,
sub_aggregation: &Aggregations,
reader: &SegmentReader,
segment_ordinal: SegmentOrdinal,
limits: AggregationLimitsGuard,
) -> crate::Result<Vec<AggregationWithAccessor>> {
let mut agg = agg.clone();
let add_agg_with_accessor = |agg: &Aggregation,
accessor: Column<u64>,
column_type: ColumnType,
aggs: &mut Vec<AggregationWithAccessor>|
-> crate::Result<()> {
let res = AggregationWithAccessor {
segment_ordinal,
accessor,
accessors: Default::default(),
value_accessors: Default::default(),
field_type: column_type,
sub_aggregation: get_aggs_with_segment_accessor_and_validate(
sub_aggregation,
reader,
segment_ordinal,
&limits,
)?,
agg: agg.clone(),
limits: limits.clone(),
missing_value_for_accessor: None,
str_dict_column: None,
column_block_accessor: Default::default(),
};
aggs.push(res);
Ok(())
};
let add_agg_with_accessors = |agg: &Aggregation,
accessors: Vec<(Column<u64>, ColumnType)>,
aggs: &mut Vec<AggregationWithAccessor>,
value_accessors: HashMap<String, Vec<DynamicColumn>>|
-> crate::Result<()> {
let (accessor, field_type) = accessors.first().expect("at least one accessor");
let limits = limits.clone();
let res = AggregationWithAccessor {
segment_ordinal,
// TODO: We should do away with the `accessor` field altogether
accessor: accessor.clone(),
value_accessors,
field_type: *field_type,
accessors,
sub_aggregation: get_aggs_with_segment_accessor_and_validate(
sub_aggregation,
reader,
segment_ordinal,
&limits,
)?,
agg: agg.clone(),
limits,
missing_value_for_accessor: None,
str_dict_column: None,
column_block_accessor: Default::default(),
};
aggs.push(res);
Ok(())
};
let mut res: Vec<AggregationWithAccessor> = Vec::new();
use AggregationVariants::*;
match agg.agg {
Range(RangeAggregation {
field: ref field_name,
..
}) => {
let (accessor, column_type) =
get_ff_reader(reader, field_name, Some(get_numeric_or_date_column_types()))?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
Histogram(HistogramAggregation {
field: ref field_name,
..
}) => {
let (accessor, column_type) =
get_ff_reader(reader, field_name, Some(get_numeric_or_date_column_types()))?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
DateHistogram(DateHistogramAggregationReq {
field: ref field_name,
..
}) => {
let (accessor, column_type) =
// Only DateTime is supported for DateHistogram
get_ff_reader(reader, field_name, Some(&[ColumnType::DateTime]))?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
Terms(TermsAggregation {
field: ref field_name,
ref missing,
..
})
| Cardinality(CardinalityAggregationReq {
field: ref field_name,
ref missing,
..
}) => {
let str_dict_column = reader.fast_fields().str(field_name)?;
let allowed_column_types = [
ColumnType::I64,
ColumnType::U64,
ColumnType::F64,
ColumnType::Str,
ColumnType::DateTime,
ColumnType::Bool,
ColumnType::IpAddr,
// ColumnType::Bytes Unsupported
];
// In case the column is empty we want the shim column to match the missing type
let fallback_type = missing
.as_ref()
.map(|missing| match missing {
Key::Str(_) => ColumnType::Str,
Key::F64(_) => ColumnType::F64,
Key::I64(_) => ColumnType::I64,
Key::U64(_) => ColumnType::U64,
})
.unwrap_or(ColumnType::U64);
let column_and_types = get_all_ff_reader_or_empty(
reader,
field_name,
Some(&allowed_column_types),
fallback_type,
)?;
let missing_and_more_than_one_col = column_and_types.len() > 1 && missing.is_some();
let text_on_non_text_col = column_and_types.len() == 1
&& column_and_types[0].1.numerical_type().is_some()
&& missing
.as_ref()
.map(|m| matches!(m, Key::Str(_)))
.unwrap_or(false);
// Actually we could convert the text to a number and have the fast path, if it is
// provided in Rfc3339 format. But this use case is probably common
// enough to justify the effort.
let text_on_date_col = column_and_types.len() == 1
&& column_and_types[0].1 == ColumnType::DateTime
&& missing
.as_ref()
.map(|m| matches!(m, Key::Str(_)))
.unwrap_or(false);
let use_special_missing_agg =
missing_and_more_than_one_col || text_on_non_text_col || text_on_date_col;
if use_special_missing_agg {
let column_and_types =
get_all_ff_reader_or_empty(reader, field_name, None, fallback_type)?;
let accessors = column_and_types
.iter()
.map(|c_t| (c_t.0.clone(), c_t.1))
.collect();
add_agg_with_accessors(&agg, accessors, &mut res, Default::default())?;
}
for (accessor, column_type) in column_and_types {
let missing_value_term_agg = if use_special_missing_agg {
None
} else {
missing.clone()
};
let missing_value_for_accessor =
if let Some(missing) = missing_value_term_agg.as_ref() {
get_missing_val_as_u64_lenient(
column_type,
missing,
agg.agg.get_fast_field_names()[0],
)?
} else {
None
};
let limits = limits.clone();
let agg = AggregationWithAccessor {
segment_ordinal,
missing_value_for_accessor,
accessor,
accessors: Default::default(),
value_accessors: Default::default(),
field_type: column_type,
sub_aggregation: get_aggs_with_segment_accessor_and_validate(
sub_aggregation,
reader,
segment_ordinal,
&limits,
)?,
agg: agg.clone(),
str_dict_column: str_dict_column.clone(),
limits,
column_block_accessor: Default::default(),
};
res.push(agg);
}
}
Average(AverageAggregation {
field: ref field_name,
..
})
| Max(MaxAggregation {
field: ref field_name,
..
})
| Min(MinAggregation {
field: ref field_name,
..
})
| Stats(StatsAggregation {
field: ref field_name,
..
})
| ExtendedStats(ExtendedStatsAggregation {
field: ref field_name,
..
})
| Sum(SumAggregation {
field: ref field_name,
..
}) => {
let (accessor, column_type) =
get_ff_reader(reader, field_name, Some(get_numeric_or_date_column_types()))?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
Count(CountAggregation {
field: ref field_name,
..
}) => {
let allowed_column_types = [
ColumnType::I64,
ColumnType::U64,
ColumnType::F64,
ColumnType::Str,
ColumnType::DateTime,
ColumnType::Bool,
ColumnType::IpAddr,
// ColumnType::Bytes Unsupported
];
let (accessor, column_type) =
get_ff_reader(reader, field_name, Some(&allowed_column_types))?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
Percentiles(ref percentiles) => {
let (accessor, column_type) = get_ff_reader(
reader,
percentiles.field_name(),
Some(get_numeric_or_date_column_types()),
)?;
add_agg_with_accessor(&agg, accessor, column_type, &mut res)?;
}
TopHits(ref mut top_hits) => {
top_hits.validate_and_resolve_field_names(reader.fast_fields().columnar())?;
let accessors: Vec<(Column<u64>, ColumnType)> = top_hits
.field_names()
.iter()
.map(|field| {
get_ff_reader(reader, field, Some(get_numeric_or_date_column_types()))
})
.collect::<crate::Result<_>>()?;
let value_accessors = top_hits
.value_field_names()
.iter()
.map(|field_name| {
Ok((
field_name.to_string(),
get_dynamic_columns(reader, field_name)?,
))
})
.collect::<crate::Result<_>>()?;
add_agg_with_accessors(&agg, accessors, &mut res, value_accessors)?;
}
};
Ok(res)
}
}
/// Get the missing value as internal u64 representation
///
/// For terms we use u64::MAX as sentinel value
/// For numerical data we convert the value into the representation
/// we would get from the fast field, when we open it as u64_lenient_for_type.
///
/// That way we can use it the same way as if it would come from the fastfield.
fn get_missing_val_as_u64_lenient(
column_type: ColumnType,
missing: &Key,
field_name: &str,
) -> crate::Result<Option<u64>> {
let missing_val = match missing {
Key::Str(_) if column_type == ColumnType::Str => Some(u64::MAX),
// Allow fallback to number on text fields
Key::F64(_) if column_type == ColumnType::Str => Some(u64::MAX),
Key::U64(_) if column_type == ColumnType::Str => Some(u64::MAX),
Key::I64(_) if column_type == ColumnType::Str => Some(u64::MAX),
Key::F64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val, &column_type)
}
// NOTE: We may loose precision of the passed missing value by casting i64 and u64 to f64.
Key::I64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val as f64, &column_type)
}
Key::U64(val) if column_type.numerical_type().is_some() => {
f64_to_fastfield_u64(*val as f64, &column_type)
}
_ => {
return Err(crate::TantivyError::InvalidArgument(format!(
"Missing value {missing:?} for field {field_name} is not supported for column \
type {column_type:?}"
)));
}
};
Ok(missing_val)
}
fn get_numeric_or_date_column_types() -> &'static [ColumnType] {
&[
ColumnType::F64,
ColumnType::U64,
ColumnType::I64,
ColumnType::DateTime,
]
}
pub(crate) fn get_aggs_with_segment_accessor_and_validate(
aggs: &Aggregations,
reader: &SegmentReader,
segment_ordinal: SegmentOrdinal,
limits: &AggregationLimitsGuard,
) -> crate::Result<AggregationsWithAccessor> {
let mut aggss = Vec::new();
for (key, agg) in aggs.iter() {
let aggs = AggregationWithAccessor::try_from_agg(
agg,
agg.sub_aggregation(),
reader,
segment_ordinal,
limits.clone(),
)?;
for agg in aggs {
aggss.push((key.to_string(), agg));
}
}
Ok(AggregationsWithAccessor::from_data(
VecWithNames::from_entries(aggss),
))
}
/// Get fast field reader or empty as default.
fn get_ff_reader(
reader: &SegmentReader,
field_name: &str,
allowed_column_types: Option<&[ColumnType]>,
) -> crate::Result<(columnar::Column<u64>, ColumnType)> {
let ff_fields = reader.fast_fields();
let ff_field_with_type = ff_fields
.u64_lenient_for_type(allowed_column_types, field_name)?
.unwrap_or_else(|| {
(
Column::build_empty_column(reader.num_docs()),
ColumnType::U64,
)
});
Ok(ff_field_with_type)
}
fn get_dynamic_columns(
reader: &SegmentReader,
field_name: &str,
) -> crate::Result<Vec<columnar::DynamicColumn>> {
let ff_fields = reader.fast_fields().dynamic_column_handles(field_name)?;
let cols = ff_fields
.iter()
.map(|h| h.open())
.collect::<io::Result<_>>()?;
assert!(!ff_fields.is_empty(), "field {field_name} not found");
Ok(cols)
}
/// Get all fast field reader or empty as default.
///
/// Is guaranteed to return at least one column.
fn get_all_ff_reader_or_empty(
reader: &SegmentReader,
field_name: &str,
allowed_column_types: Option<&[ColumnType]>,
fallback_type: ColumnType,
) -> crate::Result<Vec<(columnar::Column<u64>, ColumnType)>> {
let ff_fields = reader.fast_fields();
let mut ff_field_with_type =
ff_fields.u64_lenient_for_type_all(allowed_column_types, field_name)?;
if ff_field_with_type.is_empty() {
ff_field_with_type.push((Column::build_empty_column(reader.num_docs()), fallback_type));
}
Ok(ff_field_with_type)
}

31
src/aggregation/agg_result.rs
View File

@@ -16,7 +16,7 @@ use super::{AggregationError, Key};
use crate::TantivyError;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
/// The final aggregation result.
/// The final aggegation result.
pub struct AggregationResults(pub FxHashMap<String, AggregationResult>);
impl AggregationResults {
@@ -156,8 +156,6 @@ pub enum BucketResult {
/// The upper bound error for the doc count of each term.
doc_count_error_upper_bound: Option<u64>,
},
/// This is the filter result - a single bucket with sub-aggregations
Filter(FilterBucketResult),
}
impl BucketResult {
@@ -174,11 +172,6 @@ impl BucketResult {
sum_other_doc_count: _,
doc_count_error_upper_bound: _,
} => buckets.iter().map(|bucket| bucket.get_bucket_count()).sum(),
BucketResult::Filter(filter_result) => {
// Filter doesn't add to bucket count - it's not a user-facing bucket
// Only count sub-aggregation buckets
filter_result.sub_aggregations.get_bucket_count()
}
}
}
}
@@ -315,25 +308,3 @@ impl RangeBucketEntry {
1 + self.sub_aggregation.get_bucket_count()
}
}
/// This is the filter bucket result, which contains the document count and sub-aggregations.
///
/// # JSON Format
/// ```json
/// {
/// "electronics_only": {
/// "doc_count": 2,
/// "avg_price": {
/// "value": 150.0
/// }
/// }
/// }
/// ```
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct FilterBucketResult {
/// Number of documents in the filter bucket
pub doc_count: u64,
/// Sub-aggregation results
#[serde(flatten)]
pub sub_aggregations: AggregationResults,
}

7
src/aggregation/agg_tests.rs
View File

@@ -5,6 +5,7 @@ use crate::aggregation::agg_result::AggregationResults;
use crate::aggregation::buf_collector::DOC_BLOCK_SIZE;
use crate::aggregation::collector::AggregationCollector;
use crate::aggregation::intermediate_agg_result::IntermediateAggregationResults;
use crate::aggregation::segment_agg_result::AggregationLimitsGuard;
use crate::aggregation::tests::{get_test_index_2_segments, get_test_index_from_values_and_terms};
use crate::aggregation::DistributedAggregationCollector;
use crate::query::{AllQuery, TermQuery};
@@ -127,8 +128,10 @@ fn test_aggregation_flushing(
.unwrap();
let agg_res: AggregationResults = if use_distributed_collector {
let collector =
DistributedAggregationCollector::from_aggs(agg_req.clone(), Default::default());
let collector = DistributedAggregationCollector::from_aggs(
agg_req.clone(),
AggregationLimitsGuard::default(),
);
let searcher = reader.searcher();
let intermediate_agg_result = searcher.search(&AllQuery, &collector).unwrap();

1754
src/aggregation/bucket/filter.rs
View File

File diff suppressed because it is too large Load Diff

157
src/aggregation/bucket/histogram/histogram.rs
View File

@@ -1,54 +1,25 @@
use std::cmp::Ordering;
use columnar::{Column, ColumnBlockAccessor, ColumnType};
use rustc_hash::FxHashMap;
use serde::{Deserialize, Serialize};
use tantivy_bitpacker::minmax;
use crate::aggregation::agg_data::{
build_segment_agg_collectors, AggRefNode, AggregationsSegmentCtx,
};
use crate::aggregation::agg_limits::MemoryConsumption;
use crate::aggregation::agg_req::Aggregations;
use crate::aggregation::agg_req_with_accessor::{
AggregationWithAccessor, AggregationsWithAccessor,
};
use crate::aggregation::agg_result::BucketEntry;
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateBucketResult,
IntermediateHistogramBucketEntry,
};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::segment_agg_result::{
build_segment_agg_collector, SegmentAggregationCollector,
};
use crate::aggregation::*;
use crate::TantivyError;
/// Contains all information required by the SegmentHistogramCollector to perform the
/// histogram or date_histogram aggregation on a segment.
pub struct HistogramAggReqData {
/// The column accessor to access the fast field values.
pub accessor: Column<u64>,
/// The field type of the fast field.
pub field_type: ColumnType,
/// The column block accessor to access the fast field values.
pub column_block_accessor: ColumnBlockAccessor<u64>,
/// The name of the aggregation.
pub name: String,
/// The sub aggregation blueprint, used to create sub aggregations for each bucket.
/// Will be filled during initialization of the collector.
pub sub_aggregation_blueprint: Option<Box<dyn SegmentAggregationCollector>>,
/// The histogram aggregation request.
pub req: HistogramAggregation,
/// True if this is a date_histogram aggregation.
pub is_date_histogram: bool,
/// The bounds to limit the buckets to.
pub bounds: HistogramBounds,
/// The offset used to calculate the bucket position.
pub offset: f64,
}
impl HistogramAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
/// Histogram is a bucket aggregation, where buckets are created dynamically for given `interval`.
/// Each document value is rounded down to its bucket.
///
@@ -263,12 +234,12 @@ impl SegmentHistogramBucketEntry {
pub(crate) fn into_intermediate_bucket_entry(
self,
sub_aggregation: Option<Box<dyn SegmentAggregationCollector>>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
) -> crate::Result<IntermediateHistogramBucketEntry> {
let mut sub_aggregation_res = IntermediateAggregationResults::default();
if let Some(sub_aggregation) = sub_aggregation {
sub_aggregation
.add_intermediate_aggregation_result(agg_data, &mut sub_aggregation_res)?;
.add_intermediate_aggregation_result(agg_with_accessor, &mut sub_aggregation_res)?;
}
Ok(IntermediateHistogramBucketEntry {
key: self.key,
@@ -285,20 +256,24 @@ pub struct SegmentHistogramCollector {
/// The buckets containing the aggregation data.
buckets: FxHashMap<i64, SegmentHistogramBucketEntry>,
sub_aggregations: FxHashMap<i64, Box<dyn SegmentAggregationCollector>>,
sub_aggregation_blueprint: Option<Box<dyn SegmentAggregationCollector>>,
column_type: ColumnType,
interval: f64,
offset: f64,
bounds: HistogramBounds,
accessor_idx: usize,
}
impl SegmentAggregationCollector for SegmentHistogramCollector {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let name = agg_data
.get_histogram_req_data(self.accessor_idx)
.name
.clone();
let bucket = self.into_intermediate_bucket_result(agg_data)?;
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let agg_with_accessor = &agg_with_accessor.aggs.values[self.accessor_idx];
let bucket = self.into_intermediate_bucket_result(agg_with_accessor)?;
results.push(name, IntermediateAggregationResult::Bucket(bucket))?;
Ok(())
@@ -308,52 +283,56 @@ impl SegmentAggregationCollector for SegmentHistogramCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.collect_block(&[doc], agg_data)
self.collect_block(&[doc], agg_with_accessor)
}
#[inline]
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let mut req = agg_data.take_histogram_req_data(self.accessor_idx);
let bucket_agg_accessor = &mut agg_with_accessor.aggs.values[self.accessor_idx];
let mem_pre = self.get_memory_consumption();
let bounds = req.bounds;
let interval = req.req.interval;
let offset = req.offset;
let bounds = self.bounds;
let interval = self.interval;
let offset = self.offset;
let get_bucket_pos = |val| get_bucket_pos_f64(val, interval, offset) as i64;
req.column_block_accessor.fetch_block(docs, &req.accessor);
for (doc, val) in req
bucket_agg_accessor
.column_block_accessor
.iter_docid_vals(docs, &req.accessor)
.fetch_block(docs, &bucket_agg_accessor.accessor);
for (doc, val) in bucket_agg_accessor
.column_block_accessor
.iter_docid_vals(docs, &bucket_agg_accessor.accessor)
{
let val = f64_from_fastfield_u64(val, &req.field_type);
let val = self.f64_from_fastfield_u64(val);
let bucket_pos = get_bucket_pos(val);
if bounds.contains(val) {
let bucket = self.buckets.entry(bucket_pos).or_insert_with(|| {
let key = get_bucket_key_from_pos(bucket_pos as f64, interval, offset);
SegmentHistogramBucketEntry { key, doc_count: 0 }
});
bucket.doc_count += 1;
if let Some(sub_aggregation_blueprint) = req.sub_aggregation_blueprint.as_ref() {
if let Some(sub_aggregation_blueprint) = self.sub_aggregation_blueprint.as_mut() {
self.sub_aggregations
.entry(bucket_pos)
.or_insert_with(|| sub_aggregation_blueprint.clone())
.collect(doc, agg_data)?;
.collect(doc, &mut bucket_agg_accessor.sub_aggregation)?;
}
}
}
agg_data.put_back_histogram_req_data(self.accessor_idx, req);
let mem_delta = self.get_memory_consumption() - mem_pre;
if mem_delta > 0 {
agg_data
.context
bucket_agg_accessor
.limits
.add_memory_consumed(mem_delta as u64)?;
}
@@ -361,9 +340,12 @@ impl SegmentAggregationCollector for SegmentHistogramCollector {
Ok(())
}
fn flush(&mut self, agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
fn flush(&mut self, agg_with_accessor: &mut AggregationsWithAccessor) -> crate::Result<()> {
let sub_aggregation_accessor =
&mut agg_with_accessor.aggs.values[self.accessor_idx].sub_aggregation;
for sub_aggregation in self.sub_aggregations.values_mut() {
sub_aggregation.flush(agg_data)?;
sub_aggregation.flush(sub_aggregation_accessor)?;
}
Ok(())
@@ -380,58 +362,65 @@ impl SegmentHistogramCollector {
/// Converts the collector result into a intermediate bucket result.
pub fn into_intermediate_bucket_result(
self,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationWithAccessor,
) -> crate::Result<IntermediateBucketResult> {
let mut buckets = Vec::with_capacity(self.buckets.len());
for (bucket_pos, bucket) in self.buckets {
let bucket_res = bucket.into_intermediate_bucket_entry(
self.sub_aggregations.get(&bucket_pos).cloned(),
agg_data,
&agg_with_accessor.sub_aggregation,
);
buckets.push(bucket_res?);
}
buckets.sort_unstable_by(|b1, b2| b1.key.total_cmp(&b2.key));
let is_date_agg = agg_data
.get_histogram_req_data(self.accessor_idx)
.field_type
== ColumnType::DateTime;
Ok(IntermediateBucketResult::Histogram {
buckets,
is_date_agg,
is_date_agg: self.column_type == ColumnType::DateTime,
})
}
pub(crate) fn from_req_and_validate(
agg_data: &mut AggregationsSegmentCtx,
node: &AggRefNode,
mut req: HistogramAggregation,
sub_aggregation: &mut AggregationsWithAccessor,
field_type: ColumnType,
accessor_idx: usize,
) -> crate::Result<Self> {
let blueprint = if !node.children.is_empty() {
Some(build_segment_agg_collectors(agg_data, &node.children)?)
} else {
None
};
let req_data = agg_data.get_histogram_req_data_mut(node.idx_in_req_data);
req_data.req.validate()?;
if req_data.field_type == ColumnType::DateTime && !req_data.is_date_histogram {
req_data.req.normalize_date_time();
req.validate()?;
if field_type == ColumnType::DateTime {
req.normalize_date_time();
}
req_data.bounds = req_data.req.hard_bounds.unwrap_or(HistogramBounds {
let sub_aggregation_blueprint = if sub_aggregation.is_empty() {
None
} else {
let sub_aggregation = build_segment_agg_collector(sub_aggregation)?;
Some(sub_aggregation)
};
let bounds = req.hard_bounds.unwrap_or(HistogramBounds {
min: f64::MIN,
max: f64::MAX,
});
req_data.offset = req_data.req.offset.unwrap_or(0.0);
req_data.sub_aggregation_blueprint = blueprint;
Ok(Self {
buckets: Default::default(),
column_type: field_type,
interval: req.interval,
offset: req.offset.unwrap_or(0.0),
bounds,
sub_aggregations: Default::default(),
accessor_idx: node.idx_in_req_data,
sub_aggregation_blueprint,
accessor_idx,
})
}
#[inline]
fn f64_from_fastfield_u64(&self, val: u64) -> f64 {
f64_from_fastfield_u64(val, &self.column_type)
}
}
#[inline]

2
src/aggregation/bucket/mod.rs
View File

@@ -22,7 +22,6 @@
//! - [Range](RangeAggregation)
//! - [Terms](TermsAggregation)
mod filter;
mod histogram;
mod range;
mod term_agg;
@@ -31,7 +30,6 @@ mod term_missing_agg;
use std::collections::HashMap;
use std::fmt;
pub use filter::*;
pub use histogram::*;
pub use range::*;
use serde::{de, Deserialize, Deserializer, Serialize, Serializer};

155
src/aggregation/bucket/range.rs
View File

@@ -1,43 +1,20 @@
use std::fmt::Debug;
use std::ops::Range;
use columnar::{Column, ColumnBlockAccessor, ColumnType};
use rustc_hash::FxHashMap;
use serde::{Deserialize, Serialize};
use crate::aggregation::agg_data::{
build_segment_agg_collectors, AggRefNode, AggregationsSegmentCtx,
};
use crate::aggregation::agg_req_with_accessor::AggregationsWithAccessor;
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateBucketResult,
IntermediateRangeBucketEntry, IntermediateRangeBucketResult,
};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::segment_agg_result::{
build_segment_agg_collector, SegmentAggregationCollector,
};
use crate::aggregation::*;
use crate::TantivyError;
/// Contains all information required by the SegmentRangeCollector to perform the
/// range aggregation on a segment.
pub struct RangeAggReqData {
/// The column accessor to access the fast field values.
pub accessor: Column<u64>,
/// The type of the fast field.
pub field_type: ColumnType,
/// The column block accessor to access the fast field values.
pub column_block_accessor: ColumnBlockAccessor<u64>,
/// The range aggregation request.
pub req: RangeAggregation,
/// The name of the aggregation.
pub name: String,
}
impl RangeAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
/// Provide user-defined buckets to aggregate on.
///
/// Two special buckets will automatically be created to cover the whole range of values.
@@ -184,12 +161,12 @@ impl Debug for SegmentRangeBucketEntry {
impl SegmentRangeBucketEntry {
pub(crate) fn into_intermediate_bucket_entry(
self,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
) -> crate::Result<IntermediateRangeBucketEntry> {
let mut sub_aggregation_res = IntermediateAggregationResults::default();
if let Some(sub_aggregation) = self.sub_aggregation {
sub_aggregation
.add_intermediate_aggregation_result(agg_data, &mut sub_aggregation_res)?
.add_intermediate_aggregation_result(agg_with_accessor, &mut sub_aggregation_res)?
} else {
Default::default()
};
@@ -207,14 +184,12 @@ impl SegmentRangeBucketEntry {
impl SegmentAggregationCollector for SegmentRangeCollector {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let field_type = self.column_type;
let name = agg_data
.get_range_req_data(self.accessor_idx)
.name
.to_string();
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let sub_agg = &agg_with_accessor.aggs.values[self.accessor_idx].sub_aggregation;
let buckets: FxHashMap<SerializedKey, IntermediateRangeBucketEntry> = self
.buckets
@@ -224,7 +199,7 @@ impl SegmentAggregationCollector for SegmentRangeCollector {
range_to_string(&range_bucket.range, &field_type)?,
range_bucket
.bucket
.into_intermediate_bucket_entry(agg_data)?,
.into_intermediate_bucket_entry(sub_agg)?,
))
})
.collect::<crate::Result<_>>()?;
@@ -243,70 +218,66 @@ impl SegmentAggregationCollector for SegmentRangeCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.collect_block(&[doc], agg_data)
self.collect_block(&[doc], agg_with_accessor)
}
#[inline]
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
// Take request data to avoid borrow conflicts during sub-aggregation
let mut req = agg_data.take_range_req_data(self.accessor_idx);
let bucket_agg_accessor = &mut agg_with_accessor.aggs.values[self.accessor_idx];
req.column_block_accessor.fetch_block(docs, &req.accessor);
for (doc, val) in req
bucket_agg_accessor
.column_block_accessor
.iter_docid_vals(docs, &req.accessor)
.fetch_block(docs, &bucket_agg_accessor.accessor);
for (doc, val) in bucket_agg_accessor
.column_block_accessor
.iter_docid_vals(docs, &bucket_agg_accessor.accessor)
{
let bucket_pos = self.get_bucket_pos(val);
let bucket = &mut self.buckets[bucket_pos];
bucket.bucket.doc_count += 1;
if let Some(sub_agg) = bucket.bucket.sub_aggregation.as_mut() {
sub_agg.collect(doc, agg_data)?;
if let Some(sub_aggregation) = &mut bucket.bucket.sub_aggregation {
sub_aggregation.collect(doc, &mut bucket_agg_accessor.sub_aggregation)?;
}
}
agg_data.put_back_range_req_data(self.accessor_idx, req);
Ok(())
}
fn flush(&mut self, agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
fn flush(&mut self, agg_with_accessor: &mut AggregationsWithAccessor) -> crate::Result<()> {
let sub_aggregation_accessor =
&mut agg_with_accessor.aggs.values[self.accessor_idx].sub_aggregation;
for bucket in self.buckets.iter_mut() {
if let Some(sub_agg) = bucket.bucket.sub_aggregation.as_mut() {
sub_agg.flush(agg_data)?;
sub_agg.flush(sub_aggregation_accessor)?;
}
}
Ok(())
}
}
impl SegmentRangeCollector {
pub(crate) fn from_req_and_validate(
req_data: &mut AggregationsSegmentCtx,
node: &AggRefNode,
req: &RangeAggregation,
sub_aggregation: &mut AggregationsWithAccessor,
limits: &mut AggregationLimitsGuard,
field_type: ColumnType,
accessor_idx: usize,
) -> crate::Result<Self> {
let accessor_idx = node.idx_in_req_data;
let (field_type, ranges) = {
let req_view = req_data.get_range_req_data(node.idx_in_req_data);
(req_view.field_type, req_view.req.ranges.clone())
};
// The range input on the request is f64.
// We need to convert to u64 ranges, because we read the values as u64.
// The mapping from the conversion is monotonic so ordering is preserved.
let sub_agg_prototype = if !node.children.is_empty() {
Some(build_segment_agg_collectors(req_data, &node.children)?)
} else {
None
};
let buckets: Vec<_> = extend_validate_ranges(&ranges, &field_type)?
let buckets: Vec<_> = extend_validate_ranges(&req.ranges, &field_type)?
.iter()
.map(|range| {
let key = range
@@ -324,7 +295,11 @@ impl SegmentRangeCollector {
} else {
Some(f64_from_fastfield_u64(range.range.start, &field_type))
};
let sub_aggregation = sub_agg_prototype.clone();
let sub_aggregation = if sub_aggregation.is_empty() {
None
} else {
Some(build_segment_agg_collector(sub_aggregation)?)
};
Ok(SegmentRangeAndBucketEntry {
range: range.range.clone(),
@@ -339,7 +314,7 @@ impl SegmentRangeCollector {
})
.collect::<crate::Result<_>>()?;
req_data.context.limits.add_memory_consumed(
limits.add_memory_consumed(
buckets.len() as u64 * std::mem::size_of::<SegmentRangeAndBucketEntry>() as u64,
)?;
@@ -492,45 +467,15 @@ mod tests {
ranges,
..Default::default()
};
// Build buckets directly as in from_req_and_validate without AggregationsData
let buckets: Vec<_> = extend_validate_ranges(&req.ranges, &field_type)
.expect("unexpected error in extend_validate_ranges")
.iter()
.map(|range| {
let key = range
.key
.clone()
.map(|key| Ok(Key::Str(key)))
.unwrap_or_else(|| range_to_key(&range.range, &field_type))
.expect("unexpected error in range_to_key");
let to = if range.range.end == u64::MAX {
None
} else {
Some(f64_from_fastfield_u64(range.range.end, &field_type))
};
let from = if range.range.start == u64::MIN {
None
} else {
Some(f64_from_fastfield_u64(range.range.start, &field_type))
};
SegmentRangeAndBucketEntry {
range: range.range.clone(),
bucket: SegmentRangeBucketEntry {
doc_count: 0,
sub_aggregation: None,
key,
from,
to,
},
}
})
.collect();
SegmentRangeCollector {
buckets,
column_type: field_type,
accessor_idx: 0,
}
SegmentRangeCollector::from_req_and_validate(
&req,
&mut Default::default(),
&mut AggregationLimitsGuard::default(),
field_type,
0,
)
.expect("unexpected error")
}
#[test]

1086
src/aggregation/bucket/term_agg.rs
View File

File diff suppressed because it is too large Load Diff

76
src/aggregation/bucket/term_missing_agg.rs
View File

@@ -1,39 +1,13 @@
use columnar::{Column, ColumnType};
use rustc_hash::FxHashMap;
use crate::aggregation::agg_data::{
build_segment_agg_collectors, AggRefNode, AggregationsSegmentCtx,
};
use crate::aggregation::bucket::term_agg::TermsAggregation;
use crate::aggregation::agg_req_with_accessor::AggregationsWithAccessor;
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateBucketResult,
IntermediateKey, IntermediateTermBucketEntry, IntermediateTermBucketResult,
};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
/// Special aggregation to handle missing values for term aggregations.
/// This missing aggregation will check multiple columns for existence.
///
/// This is needed when:
/// - The field is multi-valued and we therefore have multiple columns
/// - The field is not text and missing is provided as string (we cannot use the numeric missing
/// value optimization)
#[derive(Default)]
pub struct MissingTermAggReqData {
/// The accessors to check for existence of a value.
pub accessors: Vec<(Column<u64>, ColumnType)>,
/// The name of the aggregation.
pub name: String,
/// The original terms aggregation request.
pub req: TermsAggregation,
}
impl MissingTermAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
use crate::aggregation::segment_agg_result::{
build_segment_agg_collector, SegmentAggregationCollector,
};
/// The specialized missing term aggregation.
#[derive(Default, Debug, Clone)]
@@ -44,13 +18,12 @@ pub struct TermMissingAgg {
}
impl TermMissingAgg {
pub(crate) fn new(
req_data: &mut AggregationsSegmentCtx,
node: &AggRefNode,
accessor_idx: usize,
sub_aggregations: &mut AggregationsWithAccessor,
) -> crate::Result<Self> {
let has_sub_aggregations = !node.children.is_empty();
let accessor_idx = node.idx_in_req_data;
let has_sub_aggregations = !sub_aggregations.is_empty();
let sub_agg = if has_sub_aggregations {
let sub_aggregation = build_segment_agg_collectors(req_data, &node.children)?;
let sub_aggregation = build_segment_agg_collector(sub_aggregations)?;
Some(sub_aggregation)
} else {
None
@@ -67,11 +40,16 @@ impl TermMissingAgg {
impl SegmentAggregationCollector for TermMissingAgg {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let req_data = agg_data.get_missing_term_req_data(self.accessor_idx);
let term_agg = &req_data.req;
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let agg_with_accessor = &agg_with_accessor.aggs.values[self.accessor_idx];
let term_agg = agg_with_accessor
.agg
.agg
.as_term()
.expect("TermMissingAgg collector must be term agg req");
let missing = term_agg
.missing
.as_ref()
@@ -86,7 +64,10 @@ impl SegmentAggregationCollector for TermMissingAgg {
};
if let Some(sub_agg) = self.sub_agg {
let mut res = IntermediateAggregationResults::default();
sub_agg.add_intermediate_aggregation_result(agg_data, &mut res)?;
sub_agg.add_intermediate_aggregation_result(
&agg_with_accessor.sub_aggregation,
&mut res,
)?;
missing_entry.sub_aggregation = res;
}
entries.insert(missing.into(), missing_entry);
@@ -99,10 +80,7 @@ impl SegmentAggregationCollector for TermMissingAgg {
},
};
results.push(
req_data.name.to_string(),
IntermediateAggregationResult::Bucket(bucket),
)?;
results.push(name, IntermediateAggregationResult::Bucket(bucket))?;
Ok(())
}
@@ -110,17 +88,17 @@ impl SegmentAggregationCollector for TermMissingAgg {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_missing_term_req_data(self.accessor_idx);
let has_value = req_data
let agg = &mut agg_with_accessor.aggs.values[self.accessor_idx];
let has_value = agg
.accessors
.iter()
.any(|(acc, _)| acc.index.has_value(doc));
if !has_value {
self.missing_count += 1;
if let Some(sub_agg) = self.sub_agg.as_mut() {
sub_agg.collect(doc, agg_data)?;
sub_agg.collect(doc, &mut agg.sub_aggregation)?;
}
}
Ok(())
@@ -129,10 +107,10 @@ impl SegmentAggregationCollector for TermMissingAgg {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
for doc in docs {
self.collect(*doc, agg_data)?;
self.collect(*doc, agg_with_accessor)?;
}
Ok(())
}

28
src/aggregation/buf_collector.rs
View File

@@ -1,14 +1,9 @@
use super::agg_req_with_accessor::AggregationsWithAccessor;
use super::intermediate_agg_result::IntermediateAggregationResults;
use super::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::DocId;
#[cfg(test)]
pub(crate) const DOC_BLOCK_SIZE: usize = 64;
#[cfg(not(test))]
pub(crate) const DOC_BLOCK_SIZE: usize = 256;
pub(crate) type DocBlock = [DocId; DOC_BLOCK_SIZE];
/// BufAggregationCollector buffers documents before calling collect_block().
@@ -20,7 +15,7 @@ pub(crate) struct BufAggregationCollector {
}
impl std::fmt::Debug for BufAggregationCollector {
fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.debug_struct("SegmentAggregationResultsCollector")
.field("staged_docs", &&self.staged_docs[..self.num_staged_docs])
.field("num_staged_docs", &self.num_staged_docs)
@@ -42,23 +37,23 @@ impl SegmentAggregationCollector for BufAggregationCollector {
#[inline]
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
Box::new(self.collector).add_intermediate_aggregation_result(agg_data, results)
Box::new(self.collector).add_intermediate_aggregation_result(agg_with_accessor, results)
}
#[inline]
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.staged_docs[self.num_staged_docs] = doc;
self.num_staged_docs += 1;
if self.num_staged_docs == self.staged_docs.len() {
self.collector
.collect_block(&self.staged_docs[..self.num_staged_docs], agg_data)?;
.collect_block(&self.staged_docs[..self.num_staged_docs], agg_with_accessor)?;
self.num_staged_docs = 0;
}
Ok(())
@@ -68,19 +63,20 @@ impl SegmentAggregationCollector for BufAggregationCollector {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.collector.collect_block(docs, agg_data)?;
self.collector.collect_block(docs, agg_with_accessor)?;
Ok(())
}
#[inline]
fn flush(&mut self, agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
fn flush(&mut self, agg_with_accessor: &mut AggregationsWithAccessor) -> crate::Result<()> {
self.collector
.collect_block(&self.staged_docs[..self.num_staged_docs], agg_data)?;
.collect_block(&self.staged_docs[..self.num_staged_docs], agg_with_accessor)?;
self.num_staged_docs = 0;
self.collector.flush(agg_data)?;
self.collector.flush(agg_with_accessor)?;
Ok(())
}

39
src/aggregation/collector.rs
View File

@@ -1,12 +1,12 @@
use super::agg_req::Aggregations;
use super::agg_req_with_accessor::AggregationsWithAccessor;
use super::agg_result::AggregationResults;
use super::buf_collector::BufAggregationCollector;
use super::intermediate_agg_result::IntermediateAggregationResults;
use super::segment_agg_result::SegmentAggregationCollector;
use super::AggContextParams;
use crate::aggregation::agg_data::{
build_aggregations_data_from_req, build_segment_agg_collectors_root, AggregationsSegmentCtx,
use super::segment_agg_result::{
build_segment_agg_collector, AggregationLimitsGuard, SegmentAggregationCollector,
};
use crate::aggregation::agg_req_with_accessor::get_aggs_with_segment_accessor_and_validate;
use crate::collector::{Collector, SegmentCollector};
use crate::index::SegmentReader;
use crate::{DocId, SegmentOrdinal, TantivyError};
@@ -22,7 +22,7 @@ pub const DEFAULT_MEMORY_LIMIT: u64 = 500_000_000;
/// The collector collects all aggregations by the underlying aggregation request.
pub struct AggregationCollector {
agg: Aggregations,
context: AggContextParams,
limits: AggregationLimitsGuard,
}
impl AggregationCollector {
@@ -30,8 +30,8 @@ impl AggregationCollector {
///
/// Aggregation fails when the limits in `AggregationLimits` is exceeded. (memory limit and
/// bucket limit)
pub fn from_aggs(agg: Aggregations, context: AggContextParams) -> Self {
Self { agg, context }
pub fn from_aggs(agg: Aggregations, limits: AggregationLimitsGuard) -> Self {
Self { agg, limits }
}
}
@@ -45,7 +45,7 @@ impl AggregationCollector {
/// into the final `AggregationResults` via the `into_final_result()` method.
pub struct DistributedAggregationCollector {
agg: Aggregations,
context: AggContextParams,
limits: AggregationLimitsGuard,
}
impl DistributedAggregationCollector {
@@ -53,8 +53,8 @@ impl DistributedAggregationCollector {
///
/// Aggregation fails when the limits in `AggregationLimits` is exceeded. (memory limit and
/// bucket limit)
pub fn from_aggs(agg: Aggregations, context: AggContextParams) -> Self {
Self { agg, context }
pub fn from_aggs(agg: Aggregations, limits: AggregationLimitsGuard) -> Self {
Self { agg, limits }
}
}
@@ -72,7 +72,7 @@ impl Collector for DistributedAggregationCollector {
&self.agg,
reader,
segment_local_id,
&self.context,
&self.limits,
)
}
@@ -102,7 +102,7 @@ impl Collector for AggregationCollector {
&self.agg,
reader,
segment_local_id,
&self.context,
&self.limits,
)
}
@@ -115,7 +115,7 @@ impl Collector for AggregationCollector {
segment_fruits: Vec<<Self::Child as SegmentCollector>::Fruit>,
) -> crate::Result<Self::Fruit> {
let res = merge_fruits(segment_fruits)?;
res.into_final_result(self.agg.clone(), self.context.limits.clone())
res.into_final_result(self.agg.clone(), self.limits.clone())
}
}
@@ -135,7 +135,7 @@ fn merge_fruits(
/// `AggregationSegmentCollector` does the aggregation collection on a segment.
pub struct AggregationSegmentCollector {
aggs_with_accessor: AggregationsSegmentCtx,
aggs_with_accessor: AggregationsWithAccessor,
agg_collector: BufAggregationCollector,
error: Option<TantivyError>,
}
@@ -147,15 +147,14 @@ impl AggregationSegmentCollector {
agg: &Aggregations,
reader: &SegmentReader,
segment_ordinal: SegmentOrdinal,
context: &AggContextParams,
limits: &AggregationLimitsGuard,
) -> crate::Result<Self> {
let mut agg_data =
build_aggregations_data_from_req(agg, reader, segment_ordinal, context.clone())?;
let mut aggs_with_accessor =
get_aggs_with_segment_accessor_and_validate(agg, reader, segment_ordinal, limits)?;
let result =
BufAggregationCollector::new(build_segment_agg_collectors_root(&mut agg_data)?);
BufAggregationCollector::new(build_segment_agg_collector(&mut aggs_with_accessor)?);
Ok(AggregationSegmentCollector {
aggs_with_accessor: agg_data,
aggs_with_accessor,
agg_collector: result,
error: None,
})

61
src/aggregation/intermediate_agg_result.rs
View File

@@ -24,9 +24,7 @@ use super::metric::{
};
use super::segment_agg_result::AggregationLimitsGuard;
use super::{format_date, AggregationError, Key, SerializedKey};
use crate::aggregation::agg_result::{
AggregationResults, BucketEntries, BucketEntry, FilterBucketResult,
};
use crate::aggregation::agg_result::{AggregationResults, BucketEntries, BucketEntry};
use crate::aggregation::bucket::TermsAggregationInternal;
use crate::aggregation::metric::CardinalityCollector;
use crate::TantivyError;
@@ -181,17 +179,12 @@ impl IntermediateAggregationResults {
}
/// Merge another intermediate aggregation result into this result.
pub fn merge_fruits(&mut self, mut other: IntermediateAggregationResults) -> crate::Result<()> {
for (key, left) in self.aggs_res.iter_mut() {
if let Some(key) = other.aggs_res.remove(key) {
left.merge_fruits(key)?;
}
}
// Move remainder of other aggs_res into self.
// Note: Currently we don't expect this to happen, as we create empty intermediate results
// via [IntermediateAggregationResults::empty_from_req].
for (key, value) in other.aggs_res {
self.aggs_res.insert(key, value);
///
/// The order of the values need to be the same on both results. This is ensured when the same
/// (key values) are present on the underlying `VecWithNames` struct.
pub fn merge_fruits(&mut self, other: IntermediateAggregationResults) -> crate::Result<()> {
for (left, right) in self.aggs_res.values_mut().zip(other.aggs_res.into_values()) {
left.merge_fruits(right)?;
}
Ok(())
}
@@ -248,16 +241,11 @@ pub(crate) fn empty_from_req(req: &Aggregation) -> IntermediateAggregationResult
Cardinality(_) => IntermediateAggregationResult::Metric(
IntermediateMetricResult::Cardinality(CardinalityCollector::default()),
),
Filter(_) => IntermediateAggregationResult::Bucket(IntermediateBucketResult::Filter {
doc_count: 0,
sub_aggregations: IntermediateAggregationResults::default(),
}),
}
}
/// An aggregation is either a bucket or a metric.
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
#[allow(clippy::large_enum_variant)]
pub enum IntermediateAggregationResult {
/// Bucket variant
Bucket(IntermediateBucketResult),
@@ -438,13 +426,6 @@ pub enum IntermediateBucketResult {
/// The term buckets
buckets: IntermediateTermBucketResult,
},
/// Filter aggregation - a single bucket with sub-aggregations
Filter {
/// Document count in the filter bucket
doc_count: u64,
/// Sub-aggregation results
sub_aggregations: IntermediateAggregationResults,
},
}
impl IntermediateBucketResult {
@@ -528,18 +509,6 @@ impl IntermediateBucketResult {
req.sub_aggregation(),
limits,
),
IntermediateBucketResult::Filter {
doc_count,
sub_aggregations,
} => {
// Convert sub-aggregation results to final format
let final_sub_aggregations = sub_aggregations
.into_final_result(req.sub_aggregation().clone(), limits.clone())?;
Ok(BucketResult::Filter(FilterBucketResult {
doc_count,
sub_aggregations: final_sub_aggregations,
}))
}
}
}
@@ -593,19 +562,6 @@ impl IntermediateBucketResult {
*buckets_left = buckets?;
}
(
IntermediateBucketResult::Filter {
doc_count: doc_count_left,
sub_aggregations: sub_aggs_left,
},
IntermediateBucketResult::Filter {
doc_count: doc_count_right,
sub_aggregations: sub_aggs_right,
},
) => {
*doc_count_left += doc_count_right;
sub_aggs_left.merge_fruits(sub_aggs_right)?;
}
(IntermediateBucketResult::Range(_), _) => {
panic!("try merge on different types")
}
@@ -615,9 +571,6 @@ impl IntermediateBucketResult {
(IntermediateBucketResult::Terms { .. }, _) => {
panic!("try merge on different types")
}
(IntermediateBucketResult::Filter { .. }, _) => {
panic!("try merge on different types")
}
}
Ok(())
}

91
src/aggregation/metric/cardinality.rs
View File

@@ -2,13 +2,15 @@ use std::collections::hash_map::DefaultHasher;
use std::hash::{BuildHasher, Hasher};
use columnar::column_values::CompactSpaceU64Accessor;
use columnar::{Column, ColumnBlockAccessor, ColumnType, Dictionary, StrColumn};
use columnar::Dictionary;
use common::f64_to_u64;
use hyperloglogplus::{HyperLogLog, HyperLogLogPlus};
use rustc_hash::FxHashSet;
use serde::{Deserialize, Serialize};
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::aggregation::agg_req_with_accessor::{
AggregationWithAccessor, AggregationsWithAccessor,
};
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateMetricResult,
};
@@ -95,32 +97,6 @@ pub struct CardinalityAggregationReq {
pub missing: Option<Key>,
}
/// Contains all information required by the SegmentCardinalityCollector to perform the
/// cardinality aggregation on a segment.
pub struct CardinalityAggReqData {
/// The column accessor to access the fast field values.
pub accessor: Column<u64>,
/// The column_type of the field.
pub column_type: ColumnType,
/// The string dictionary column if the field is of type string.
pub str_dict_column: Option<StrColumn>,
/// The missing value normalized to the internal u64 representation of the field type.
pub missing_value_for_accessor: Option<u64>,
/// The column block accessor to access the fast field values.
pub(crate) column_block_accessor: ColumnBlockAccessor<u64>,
/// The name of the aggregation.
pub name: String,
/// The aggregation request.
pub req: CardinalityAggregationReq,
}
impl CardinalityAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
impl CardinalityAggregationReq {
/// Creates a new [`CardinalityAggregationReq`] instance from a field name.
pub fn from_field_name(field_name: String) -> Self {
@@ -139,44 +115,47 @@ impl CardinalityAggregationReq {
pub(crate) struct SegmentCardinalityCollector {
cardinality: CardinalityCollector,
entries: FxHashSet<u64>,
column_type: ColumnType,
accessor_idx: usize,
missing: Option<Key>,
}
impl SegmentCardinalityCollector {
pub fn from_req(column_type: ColumnType, accessor_idx: usize) -> Self {
pub fn from_req(column_type: ColumnType, accessor_idx: usize, missing: &Option<Key>) -> Self {
Self {
cardinality: CardinalityCollector::new(column_type as u8),
entries: Default::default(),
column_type,
accessor_idx,
missing: missing.clone(),
}
}
fn fetch_block_with_field(
&mut self,
docs: &[crate::DocId],
agg_data: &mut CardinalityAggReqData,
agg_accessor: &mut AggregationWithAccessor,
) {
if let Some(missing) = agg_data.missing_value_for_accessor {
agg_data.column_block_accessor.fetch_block_with_missing(
if let Some(missing) = agg_accessor.missing_value_for_accessor {
agg_accessor.column_block_accessor.fetch_block_with_missing(
docs,
&agg_data.accessor,
&agg_accessor.accessor,
missing,
);
} else {
agg_data
agg_accessor
.column_block_accessor
.fetch_block(docs, &agg_data.accessor);
.fetch_block(docs, &agg_accessor.accessor);
}
}
fn into_intermediate_metric_result(
mut self,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationWithAccessor,
) -> crate::Result<IntermediateMetricResult> {
let req_data = &agg_data.get_cardinality_req_data(self.accessor_idx);
if req_data.column_type == ColumnType::Str {
if self.column_type == ColumnType::Str {
let fallback_dict = Dictionary::empty();
let dict = req_data
let dict = agg_with_accessor
.str_dict_column
.as_ref()
.map(|el| el.dictionary())
@@ -201,10 +180,10 @@ impl SegmentCardinalityCollector {
})?;
if has_missing {
// Replace missing with the actual value provided
let missing_key =
req_data.req.missing.as_ref().expect(
"Found sentinel value u64::MAX for term_ord but `missing` is not set",
);
let missing_key = self
.missing
.as_ref()
.expect("Found sentinel value u64::MAX for term_ord but `missing` is not set");
match missing_key {
Key::Str(missing) => {
self.cardinality.sketch.insert_any(&missing);
@@ -230,13 +209,13 @@ impl SegmentCardinalityCollector {
impl SegmentAggregationCollector for SegmentCardinalityCollector {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let req_data = &agg_data.get_cardinality_req_data(self.accessor_idx);
let name = req_data.name.to_string();
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let agg_with_accessor = &agg_with_accessor.aggs.values[self.accessor_idx];
let intermediate_result = self.into_intermediate_metric_result(agg_data)?;
let intermediate_result = self.into_intermediate_metric_result(agg_with_accessor)?;
results.push(
name,
IntermediateAggregationResult::Metric(intermediate_result),
@@ -248,26 +227,26 @@ impl SegmentAggregationCollector for SegmentCardinalityCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.collect_block(&[doc], agg_data)
self.collect_block(&[doc], agg_with_accessor)
}
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_cardinality_req_data_mut(self.accessor_idx);
self.fetch_block_with_field(docs, req_data);
let bucket_agg_accessor = &mut agg_with_accessor.aggs.values[self.accessor_idx];
self.fetch_block_with_field(docs, bucket_agg_accessor);
let col_block_accessor = &req_data.column_block_accessor;
if req_data.column_type == ColumnType::Str {
let col_block_accessor = &bucket_agg_accessor.column_block_accessor;
if self.column_type == ColumnType::Str {
for term_ord in col_block_accessor.iter_vals() {
self.entries.insert(term_ord);
}
} else if req_data.column_type == ColumnType::IpAddr {
let compact_space_accessor = req_data
} else if self.column_type == ColumnType::IpAddr {
let compact_space_accessor = bucket_agg_accessor
.accessor
.values
.clone()

37
src/aggregation/metric/extended_stats.rs
View File

@@ -4,11 +4,12 @@ use std::mem;
use serde::{Deserialize, Serialize};
use super::*;
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::aggregation::agg_req_with_accessor::{
AggregationWithAccessor, AggregationsWithAccessor,
};
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateMetricResult,
};
use crate::aggregation::metric::MetricAggReqData;
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::*;
use crate::{DocId, TantivyError};
@@ -62,7 +63,7 @@ impl ExtendedStatsAggregation {
/// Extended stats contains a collection of statistics
/// they extends stats adding variance, standard deviation
/// and bound information
/// and bound informations
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct ExtendedStats {
/// The number of documents.
@@ -347,20 +348,20 @@ impl SegmentExtendedStatsCollector {
pub(crate) fn collect_block_with_field(
&mut self,
docs: &[DocId],
req_data: &mut MetricAggReqData,
agg_accessor: &mut AggregationWithAccessor,
) {
if let Some(missing) = self.missing.as_ref() {
req_data.column_block_accessor.fetch_block_with_missing(
agg_accessor.column_block_accessor.fetch_block_with_missing(
docs,
&req_data.accessor,
&agg_accessor.accessor,
*missing,
);
} else {
req_data
agg_accessor
.column_block_accessor
.fetch_block(docs, &req_data.accessor);
.fetch_block(docs, &agg_accessor.accessor);
}
for val in req_data.column_block_accessor.iter_vals() {
for val in agg_accessor.column_block_accessor.iter_vals() {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.extended_stats.collect(val1);
}
@@ -371,10 +372,10 @@ impl SegmentAggregationCollector for SegmentExtendedStatsCollector {
#[inline]
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let name = agg_data.get_metric_req_data(self.accessor_idx).name.clone();
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
results.push(
name,
IntermediateAggregationResult::Metric(IntermediateMetricResult::ExtendedStats(
@@ -389,12 +390,12 @@ impl SegmentAggregationCollector for SegmentExtendedStatsCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data(self.accessor_idx);
let field = &agg_with_accessor.aggs.values[self.accessor_idx].accessor;
if let Some(missing) = self.missing {
let mut has_val = false;
for val in req_data.accessor.values_for_doc(doc) {
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.extended_stats.collect(val1);
has_val = true;
@@ -404,7 +405,7 @@ impl SegmentAggregationCollector for SegmentExtendedStatsCollector {
.collect(f64_from_fastfield_u64(missing, &self.field_type));
}
} else {
for val in req_data.accessor.values_for_doc(doc) {
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.extended_stats.collect(val1);
}
@@ -417,10 +418,10 @@ impl SegmentAggregationCollector for SegmentExtendedStatsCollector {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data_mut(self.accessor_idx);
self.collect_block_with_field(docs, req_data);
let field = &mut agg_with_accessor.aggs.values[self.accessor_idx];
self.collect_block_with_field(docs, field);
Ok(())
}
}

30
src/aggregation/metric/mod.rs
View File

@@ -31,7 +31,6 @@ use std::collections::HashMap;
pub use average::*;
pub use cardinality::*;
use columnar::{Column, ColumnBlockAccessor, ColumnType};
pub use count::*;
pub use extended_stats::*;
pub use max::*;
@@ -45,35 +44,6 @@ pub use top_hits::*;
use crate::schema::OwnedValue;
/// Contains all information required by metric aggregations like avg, min, max, sum, stats,
/// extended_stats, count, percentiles.
#[repr(C)]
pub struct MetricAggReqData {
/// True if the field is of number or date type.
pub is_number_or_date_type: bool,
/// The type of the field.
pub field_type: ColumnType,
/// The missing value normalized to the internal u64 representation of the field type.
pub missing_u64: Option<u64>,
/// The column block accessor to access the fast field values.
pub column_block_accessor: ColumnBlockAccessor<u64>,
/// The column accessor to access the fast field values.
pub accessor: Column<u64>,
/// Used when converting to intermediate result
pub collecting_for: StatsType,
/// The missing value
pub missing: Option<f64>,
/// The name of the aggregation.
pub name: String,
}
impl MetricAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
/// Single-metric aggregations use this common result structure.
///
/// Main reason to wrap it in value is to match elasticsearch output structure.

65
src/aggregation/metric/percentiles.rs
View File

@@ -3,11 +3,12 @@ use std::fmt::Debug;
use serde::{Deserialize, Serialize};
use super::*;
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::aggregation::agg_req_with_accessor::{
AggregationWithAccessor, AggregationsWithAccessor,
};
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateMetricResult,
};
use crate::aggregation::metric::MetricAggReqData;
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::*;
use crate::{DocId, TantivyError};
@@ -111,8 +112,7 @@ impl PercentilesAggregationReq {
&self.field
}
/// Validates the request parameters.
pub fn validate(&self) -> crate::Result<()> {
fn validate(&self) -> crate::Result<()> {
if let Some(percents) = self.percents.as_ref() {
let all_in_range = percents
.iter()
@@ -133,8 +133,10 @@ impl PercentilesAggregationReq {
#[derive(Clone, Debug, PartialEq)]
pub(crate) struct SegmentPercentilesCollector {
field_type: ColumnType,
pub(crate) percentiles: PercentilesCollector,
pub(crate) accessor_idx: usize,
missing: Option<u64>,
}
#[derive(Clone, Serialize, Deserialize)]
@@ -229,32 +231,43 @@ impl PercentilesCollector {
}
impl SegmentPercentilesCollector {
pub fn from_req_and_validate(accessor_idx: usize) -> crate::Result<Self> {
pub fn from_req_and_validate(
req: &PercentilesAggregationReq,
field_type: ColumnType,
accessor_idx: usize,
) -> crate::Result<Self> {
req.validate()?;
let missing = req
.missing
.and_then(|val| f64_to_fastfield_u64(val, &field_type));
Ok(Self {
field_type,
percentiles: PercentilesCollector::new(),
accessor_idx,
missing,
})
}
#[inline]
pub(crate) fn collect_block_with_field(
&mut self,
docs: &[DocId],
req_data: &mut MetricAggReqData,
agg_accessor: &mut AggregationWithAccessor,
) {
if let Some(missing) = req_data.missing_u64.as_ref() {
req_data.column_block_accessor.fetch_block_with_missing(
if let Some(missing) = self.missing.as_ref() {
agg_accessor.column_block_accessor.fetch_block_with_missing(
docs,
&req_data.accessor,
&agg_accessor.accessor,
*missing,
);
} else {
req_data
agg_accessor
.column_block_accessor
.fetch_block(docs, &req_data.accessor);
.fetch_block(docs, &agg_accessor.accessor);
}
for val in req_data.column_block_accessor.iter_vals() {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
for val in agg_accessor.column_block_accessor.iter_vals() {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.percentiles.collect(val1);
}
}
@@ -264,10 +277,10 @@ impl SegmentAggregationCollector for SegmentPercentilesCollector {
#[inline]
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let name = agg_data.get_metric_req_data(self.accessor_idx).name.clone();
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let intermediate_metric_result = IntermediateMetricResult::Percentiles(self.percentiles);
results.push(
@@ -282,24 +295,24 @@ impl SegmentAggregationCollector for SegmentPercentilesCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data(self.accessor_idx);
let field = &agg_with_accessor.aggs.values[self.accessor_idx].accessor;
if let Some(missing) = req_data.missing_u64 {
if let Some(missing) = self.missing {
let mut has_val = false;
for val in req_data.accessor.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.percentiles.collect(val1);
has_val = true;
}
if !has_val {
self.percentiles
.collect(f64_from_fastfield_u64(missing, &req_data.field_type));
.collect(f64_from_fastfield_u64(missing, &self.field_type));
}
} else {
for val in req_data.accessor.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.percentiles.collect(val1);
}
}
@@ -311,10 +324,10 @@ impl SegmentAggregationCollector for SegmentPercentilesCollector {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data_mut(self.accessor_idx);
self.collect_block_with_field(docs, req_data);
let field = &mut agg_with_accessor.aggs.values[self.accessor_idx];
self.collect_block_with_field(docs, field);
Ok(())
}
}

117
src/aggregation/metric/stats.rs
View File

@@ -3,11 +3,12 @@ use std::fmt::Debug;
use serde::{Deserialize, Serialize};
use super::*;
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::aggregation::agg_req_with_accessor::{
AggregationWithAccessor, AggregationsWithAccessor,
};
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateAggregationResults, IntermediateMetricResult,
};
use crate::aggregation::metric::MetricAggReqData;
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::*;
use crate::{DocId, TantivyError};
@@ -165,65 +166,74 @@ impl IntermediateStats {
}
}
/// The type of stats aggregation to perform.
/// Note that not all stats types are supported in the stats aggregation.
#[derive(Clone, Copy, Debug)]
pub enum StatsType {
/// The average of the values.
#[derive(Clone, Debug, PartialEq)]
pub(crate) enum SegmentStatsType {
Average,
/// The count of the values.
Count,
/// The maximum value.
Max,
/// The minimum value.
Min,
/// The stats (count, sum, min, max, avg) of the values.
Stats,
/// The extended stats (count, sum, min, max, avg, sum_of_squares, variance, std_deviation,
ExtendedStats(Option<f64>), // sigma
/// The sum of the values.
Sum,
/// The percentiles of the values.
Percentiles,
}
#[derive(Clone, Debug)]
#[derive(Clone, Debug, PartialEq)]
pub(crate) struct SegmentStatsCollector {
missing: Option<u64>,
field_type: ColumnType,
pub(crate) collecting_for: SegmentStatsType,
pub(crate) stats: IntermediateStats,
pub(crate) accessor_idx: usize,
val_cache: Vec<u64>,
}
impl SegmentStatsCollector {
pub fn from_req(accessor_idx: usize) -> Self {
pub fn from_req(
field_type: ColumnType,
collecting_for: SegmentStatsType,
accessor_idx: usize,
missing: Option<f64>,
) -> Self {
let missing = missing.and_then(|val| f64_to_fastfield_u64(val, &field_type));
Self {
field_type,
collecting_for,
stats: IntermediateStats::default(),
accessor_idx,
missing,
val_cache: Default::default(),
}
}
#[inline]
pub(crate) fn collect_block_with_field(
&mut self,
docs: &[DocId],
req_data: &mut MetricAggReqData,
agg_accessor: &mut AggregationWithAccessor,
) {
if let Some(missing) = req_data.missing_u64.as_ref() {
req_data.column_block_accessor.fetch_block_with_missing(
if let Some(missing) = self.missing.as_ref() {
agg_accessor.column_block_accessor.fetch_block_with_missing(
docs,
&req_data.accessor,
&agg_accessor.accessor,
*missing,
);
} else {
req_data
agg_accessor
.column_block_accessor
.fetch_block(docs, &req_data.accessor);
.fetch_block(docs, &agg_accessor.accessor);
}
if req_data.is_number_or_date_type {
for val in req_data.column_block_accessor.iter_vals() {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
if [
ColumnType::I64,
ColumnType::U64,
ColumnType::F64,
ColumnType::DateTime,
]
.contains(&self.field_type)
{
for val in agg_accessor.column_block_accessor.iter_vals() {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.stats.collect(val1);
}
} else {
for _val in req_data.column_block_accessor.iter_vals() {
for _val in agg_accessor.column_block_accessor.iter_vals() {
// we ignore the value and simply record that we got something
self.stats.collect(0.0);
}
@@ -235,28 +245,27 @@ impl SegmentAggregationCollector for SegmentStatsCollector {
#[inline]
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
let req = agg_data.get_metric_req_data(self.accessor_idx);
let name = req.name.clone();
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let intermediate_metric_result = match req.collecting_for {
StatsType::Average => {
let intermediate_metric_result = match self.collecting_for {
SegmentStatsType::Average => {
IntermediateMetricResult::Average(IntermediateAverage::from_collector(*self))
}
StatsType::Count => {
SegmentStatsType::Count => {
IntermediateMetricResult::Count(IntermediateCount::from_collector(*self))
}
StatsType::Max => IntermediateMetricResult::Max(IntermediateMax::from_collector(*self)),
StatsType::Min => IntermediateMetricResult::Min(IntermediateMin::from_collector(*self)),
StatsType::Stats => IntermediateMetricResult::Stats(self.stats),
StatsType::Sum => IntermediateMetricResult::Sum(IntermediateSum::from_collector(*self)),
_ => {
return Err(TantivyError::InvalidArgument(format!(
"Unsupported stats type for stats aggregation: {:?}",
req.collecting_for
)))
SegmentStatsType::Max => {
IntermediateMetricResult::Max(IntermediateMax::from_collector(*self))
}
SegmentStatsType::Min => {
IntermediateMetricResult::Min(IntermediateMin::from_collector(*self))
}
SegmentStatsType::Stats => IntermediateMetricResult::Stats(self.stats),
SegmentStatsType::Sum => {
IntermediateMetricResult::Sum(IntermediateSum::from_collector(*self))
}
};
@@ -272,23 +281,23 @@ impl SegmentAggregationCollector for SegmentStatsCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data(self.accessor_idx);
if let Some(missing) = req_data.missing_u64 {
let field = &agg_with_accessor.aggs.values[self.accessor_idx].accessor;
if let Some(missing) = self.missing {
let mut has_val = false;
for val in req_data.accessor.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.stats.collect(val1);
has_val = true;
}
if !has_val {
self.stats
.collect(f64_from_fastfield_u64(missing, &req_data.field_type));
.collect(f64_from_fastfield_u64(missing, &self.field_type));
}
} else {
for val in req_data.accessor.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &req_data.field_type);
for val in field.values_for_doc(doc) {
let val1 = f64_from_fastfield_u64(val, &self.field_type);
self.stats.collect(val1);
}
}
@@ -300,10 +309,10 @@ impl SegmentAggregationCollector for SegmentStatsCollector {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_metric_req_data_mut(self.accessor_idx);
self.collect_block_with_field(docs, req_data);
let field = &mut agg_with_accessor.aggs.values[self.accessor_idx];
self.collect_block_with_field(docs, field);
Ok(())
}
}

94
src/aggregation/metric/top_hits.rs
View File

@@ -9,41 +9,15 @@ use serde::ser::SerializeMap;
use serde::{Deserialize, Deserializer, Serialize, Serializer};
use super::{TopHitsMetricResult, TopHitsVecEntry};
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use crate::aggregation::bucket::Order;
use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateMetricResult,
};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::AggregationError;
use crate::collector::sort_key::ReverseComparator;
use crate::collector::TopNComputer;
use crate::schema::OwnedValue;
use crate::{DocAddress, DocId, SegmentOrdinal};
// duplicate import removed; already imported above
/// Contains all information required by the TopHitsSegmentCollector to perform the
/// top_hits aggregation on a segment.
#[derive(Default)]
pub struct TopHitsAggReqData {
/// The accessors to access the fast field values.
pub accessors: Vec<(Column<u64>, ColumnType)>,
/// The accessors to access the fast field values for retrieving document fields.
pub value_accessors: HashMap<String, Vec<DynamicColumn>>,
/// The ordinal of the segment this request data is for.
pub segment_ordinal: SegmentOrdinal,
/// The name of the aggregation.
pub name: String,
/// The top_hits aggregation request.
pub req: TopHitsAggregationReq,
}
impl TopHitsAggReqData {
/// Estimate the memory consumption of this struct in bytes.
pub fn get_memory_consumption(&self) -> usize {
std::mem::size_of::<Self>()
}
}
/// # Top Hits
///
@@ -459,7 +433,7 @@ impl Eq for DocSortValuesAndFields {}
#[derive(Clone, Serialize, Deserialize, Debug)]
pub struct TopHitsTopNComputer {
req: TopHitsAggregationReq,
top_n: TopNComputer<DocSortValuesAndFields, DocAddress, ReverseComparator>,
top_n: TopNComputer<DocSortValuesAndFields, DocAddress, false>,
}
impl std::cmp::PartialEq for TopHitsTopNComputer {
@@ -483,7 +457,7 @@ impl TopHitsTopNComputer {
pub(crate) fn merge_fruits(&mut self, other_fruit: Self) -> crate::Result<()> {
for doc in other_fruit.top_n.into_vec() {
self.collect(doc.sort_key, doc.doc);
self.collect(doc.feature, doc.doc);
}
Ok(())
}
@@ -495,9 +469,9 @@ impl TopHitsTopNComputer {
.into_sorted_vec()
.into_iter()
.map(|doc| TopHitsVecEntry {
sort: doc.sort_key.sorts.iter().map(|f| f.value).collect(),
sort: doc.feature.sorts.iter().map(|f| f.value).collect(),
doc_value_fields: doc
.sort_key
.feature
.doc_value_fields
.into_iter()
.map(|(k, v)| (k, v.into()))
@@ -518,7 +492,7 @@ impl TopHitsTopNComputer {
pub(crate) struct TopHitsSegmentCollector {
segment_ordinal: SegmentOrdinal,
accessor_idx: usize,
top_n: TopNComputer<Vec<DocValueAndOrder>, DocAddress, ReverseComparator>,
top_n: TopNComputer<Vec<DocValueAndOrder>, DocAddress, false>,
}
impl TopHitsSegmentCollector {
@@ -545,7 +519,7 @@ impl TopHitsSegmentCollector {
let doc_value_fields = req.get_document_field_data(value_accessors, res.doc.doc_id);
top_hits_computer.collect(
DocSortValuesAndFields {
sorts: res.sort_key,
sorts: res.feature,
doc_value_fields,
},
res.doc,
@@ -592,18 +566,23 @@ impl TopHitsSegmentCollector {
impl SegmentAggregationCollector for TopHitsSegmentCollector {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &crate::aggregation::agg_req_with_accessor::AggregationsWithAccessor,
results: &mut crate::aggregation::intermediate_agg_result::IntermediateAggregationResults,
) -> crate::Result<()> {
let req_data = agg_data.get_top_hits_req_data(self.accessor_idx);
let name = agg_with_accessor.aggs.keys[self.accessor_idx].to_string();
let value_accessors = &req_data.value_accessors;
let value_accessors = &agg_with_accessor.aggs.values[self.accessor_idx].value_accessors;
let tophits_req = &agg_with_accessor.aggs.values[self.accessor_idx]
.agg
.agg
.as_top_hits()
.expect("aggregation request must be of type top hits");
let intermediate_result = IntermediateMetricResult::TopHits(
self.into_top_hits_collector(value_accessors, &req_data.req),
self.into_top_hits_collector(value_accessors, tophits_req),
);
results.push(
req_data.name.to_string(),
name,
IntermediateAggregationResult::Metric(intermediate_result),
)
}
@@ -612,22 +591,32 @@ impl SegmentAggregationCollector for TopHitsSegmentCollector {
fn collect(
&mut self,
doc_id: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut crate::aggregation::agg_req_with_accessor::AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_top_hits_req_data(self.accessor_idx);
self.collect_with(doc_id, &req_data.req, &req_data.accessors)?;
let tophits_req = &agg_with_accessor.aggs.values[self.accessor_idx]
.agg
.agg
.as_top_hits()
.expect("aggregation request must be of type top hits");
let accessors = &agg_with_accessor.aggs.values[self.accessor_idx].accessors;
self.collect_with(doc_id, tophits_req, accessors)?;
Ok(())
}
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut crate::aggregation::agg_req_with_accessor::AggregationsWithAccessor,
) -> crate::Result<()> {
let req_data = agg_data.get_top_hits_req_data(self.accessor_idx);
let tophits_req = &agg_with_accessor.aggs.values[self.accessor_idx]
.agg
.agg
.as_top_hits()
.expect("aggregation request must be of type top hits");
let accessors = &agg_with_accessor.aggs.values[self.accessor_idx].accessors;
// TODO: Consider getting fields with the column block accessor.
for doc in docs {
self.collect_with(*doc, &req_data.req, &req_data.accessors)?;
self.collect_with(*doc, tophits_req, accessors)?;
}
Ok(())
}
@@ -646,7 +635,6 @@ mod tests {
use crate::aggregation::bucket::tests::get_test_index_from_docs;
use crate::aggregation::tests::get_test_index_from_values;
use crate::aggregation::AggregationCollector;
use crate::collector::sort_key::ReverseComparator;
use crate::collector::ComparableDoc;
use crate::query::AllQuery;
use crate::schema::OwnedValue;
@@ -662,7 +650,7 @@ mod tests {
fn collector_with_capacity(capacity: usize) -> super::TopHitsTopNComputer {
super::TopHitsTopNComputer {
top_n: super::TopNComputer::new_with_comparator(capacity, ReverseComparator),
top_n: super::TopNComputer::new(capacity),
req: Default::default(),
}
}
@@ -776,12 +764,12 @@ mod tests {
#[test]
fn test_top_hits_collector_single_feature() -> crate::Result<()> {
let docs = vec![
ComparableDoc::<_, _> {
ComparableDoc::<_, _, false> {
doc: crate::DocAddress {
segment_ord: 0,
doc_id: 0,
},
sort_key: DocSortValuesAndFields {
feature: DocSortValuesAndFields {
sorts: vec![DocValueAndOrder {
value: Some(1),
order: Order::Asc,
@@ -794,7 +782,7 @@ mod tests {
segment_ord: 0,
doc_id: 2,
},
sort_key: DocSortValuesAndFields {
feature: DocSortValuesAndFields {
sorts: vec![DocValueAndOrder {
value: Some(3),
order: Order::Asc,
@@ -807,7 +795,7 @@ mod tests {
segment_ord: 0,
doc_id: 1,
},
sort_key: DocSortValuesAndFields {
feature: DocSortValuesAndFields {
sorts: vec![DocValueAndOrder {
value: Some(5),
order: Order::Asc,
@@ -819,7 +807,7 @@ mod tests {
let mut collector = collector_with_capacity(3);
for doc in docs.clone() {
collector.collect(doc.sort_key, doc.doc);
collector.collect(doc.feature, doc.doc);
}
let res = collector.into_final_result();
@@ -829,15 +817,15 @@ mod tests {
super::TopHitsMetricResult {
hits: vec![
super::TopHitsVecEntry {
sort: vec![docs[0].sort_key.sorts[0].value],
sort: vec![docs[0].feature.sorts[0].value],
doc_value_fields: Default::default(),
},
super::TopHitsVecEntry {
sort: vec![docs[1].sort_key.sorts[0].value],
sort: vec![docs[1].feature.sorts[0].value],
doc_value_fields: Default::default(),
},
super::TopHitsVecEntry {
sort: vec![docs[2].sort_key.sorts[0].value],
sort: vec![docs[2].feature.sorts[0].value],
doc_value_fields: Default::default(),
},
]

105
src/aggregation/mod.rs
View File

@@ -127,10 +127,9 @@
//! [`AggregationResults`](agg_result::AggregationResults) via the
//! [`into_final_result`](intermediate_agg_result::IntermediateAggregationResults::into_final_result) method.
mod accessor_helpers;
mod agg_data;
mod agg_limits;
pub mod agg_req;
mod agg_req_with_accessor;
pub mod agg_result;
pub mod bucket;
mod buf_collector;
@@ -141,6 +140,7 @@ pub mod intermediate_agg_result;
pub mod metric;
mod segment_agg_result;
use std::collections::HashMap;
use std::fmt::Display;
#[cfg(test)]
@@ -160,28 +160,6 @@ use itertools::Itertools;
use serde::de::{self, Visitor};
use serde::{Deserialize, Deserializer, Serialize};
use crate::tokenizer::TokenizerManager;
/// Context parameters for aggregation execution
///
/// This struct holds shared resources needed during aggregation execution:
/// - `limits`: Memory and bucket limits for the aggregation
/// - `tokenizers`: TokenizerManager for parsing query strings in filter aggregations
#[derive(Clone, Default)]
pub struct AggContextParams {
/// Aggregation limits (memory and bucket count)
pub limits: AggregationLimitsGuard,
/// Tokenizer manager for query string parsing
pub tokenizers: TokenizerManager,
}
impl AggContextParams {
/// Create new aggregation context parameters
pub fn new(limits: AggregationLimitsGuard, tokenizers: TokenizerManager) -> Self {
Self { limits, tokenizers }
}
}
fn parse_str_into_f64<E: de::Error>(value: &str) -> Result<f64, E> {
let parsed = value
.parse::<f64>()
@@ -279,6 +257,80 @@ where D: Deserializer<'de> {
deserializer.deserialize_any(StringOrFloatVisitor)
}
/// Represents an associative array `(key => values)` in a very efficient manner.
#[derive(PartialEq, Serialize, Deserialize)]
pub(crate) struct VecWithNames<T> {
pub(crate) values: Vec<T>,
keys: Vec<String>,
}
impl<T: Clone> Clone for VecWithNames<T> {
fn clone(&self) -> Self {
Self {
values: self.values.clone(),
keys: self.keys.clone(),
}
}
}
impl<T> Default for VecWithNames<T> {
fn default() -> Self {
Self {
values: Default::default(),
keys: Default::default(),
}
}
}
impl<T: std::fmt::Debug> std::fmt::Debug for VecWithNames<T> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.debug_map().entries(self.iter()).finish()
}
}
impl<T> From<HashMap<String, T>> for VecWithNames<T> {
fn from(map: HashMap<String, T>) -> Self {
VecWithNames::from_entries(map.into_iter().collect_vec())
}
}
impl<T> VecWithNames<T> {
fn from_entries(mut entries: Vec<(String, T)>) -> Self {
// Sort to ensure order of elements match across multiple instances
entries.sort_by(|left, right| left.0.cmp(&right.0));
let mut data = Vec::with_capacity(entries.len());
let mut data_names = Vec::with_capacity(entries.len());
for entry in entries {
data_names.push(entry.0);
data.push(entry.1);
}
VecWithNames {
values: data,
keys: data_names,
}
}
fn iter(&self) -> impl Iterator<Item = (&str, &T)> + '_ {
self.keys().zip(self.values.iter())
}
fn keys(&self) -> impl Iterator<Item = &str> + '_ {
self.keys.iter().map(|key| key.as_str())
}
fn values_mut(&mut self) -> impl Iterator<Item = &mut T> + '_ {
self.values.iter_mut()
}
fn is_empty(&self) -> bool {
self.keys.is_empty()
}
fn len(&self) -> usize {
self.keys.len()
}
fn get(&self, name: &str) -> Option<&T> {
self.keys()
.position(|key| key == name)
.map(|pos| &self.values[pos])
}
}
/// The serialized key is used in a `HashMap`.
pub type SerializedKey = String;
@@ -412,10 +464,7 @@ mod tests {
query: Option<(&str, &str)>,
limits: AggregationLimitsGuard,
) -> crate::Result<Value> {
let collector = AggregationCollector::from_aggs(
agg_req,
AggContextParams::new(limits, index.tokenizers().clone()),
);
let collector = AggregationCollector::from_aggs(agg_req, limits);
let reader = index.reader()?;
let searcher = reader.searcher();

170
src/aggregation/segment_agg_result.rs
View File

@@ -6,38 +6,48 @@
use std::fmt::Debug;
pub(crate) use super::agg_limits::AggregationLimitsGuard;
use super::agg_req::AggregationVariants;
use super::agg_req_with_accessor::{AggregationWithAccessor, AggregationsWithAccessor};
use super::bucket::{SegmentHistogramCollector, SegmentRangeCollector, SegmentTermCollector};
use super::intermediate_agg_result::IntermediateAggregationResults;
use crate::aggregation::agg_data::AggregationsSegmentCtx;
use super::metric::{
AverageAggregation, CountAggregation, ExtendedStatsAggregation, MaxAggregation, MinAggregation,
SegmentPercentilesCollector, SegmentStatsCollector, SegmentStatsType, StatsAggregation,
SumAggregation,
};
use crate::aggregation::bucket::TermMissingAgg;
use crate::aggregation::metric::{
CardinalityAggregationReq, SegmentCardinalityCollector, SegmentExtendedStatsCollector,
TopHitsSegmentCollector,
};
/// A SegmentAggregationCollector is used to collect aggregation results.
pub trait SegmentAggregationCollector: CollectorClone + Debug {
pub(crate) trait SegmentAggregationCollector: CollectorClone + Debug {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()>;
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()>;
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()>;
/// Finalize method. Some Aggregator collect blocks of docs before calling `collect_block`.
/// This method ensures those staged docs will be collected.
fn flush(&mut self, _agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
fn flush(&mut self, _agg_with_accessor: &mut AggregationsWithAccessor) -> crate::Result<()> {
Ok(())
}
}
/// A helper trait to enable cloning of Box<dyn SegmentAggregationCollector>
pub trait CollectorClone {
pub(crate) trait CollectorClone {
fn clone_box(&self) -> Box<dyn SegmentAggregationCollector>;
}
@@ -55,6 +65,119 @@ impl Clone for Box<dyn SegmentAggregationCollector> {
}
}
pub(crate) fn build_segment_agg_collector(
req: &mut AggregationsWithAccessor,
) -> crate::Result<Box<dyn SegmentAggregationCollector>> {
// Single collector special case
if req.aggs.len() == 1 {
let req = &mut req.aggs.values[0];
let accessor_idx = 0;
return build_single_agg_segment_collector(req, accessor_idx);
}
let agg = GenericSegmentAggregationResultsCollector::from_req_and_validate(req)?;
Ok(Box::new(agg))
}
pub(crate) fn build_single_agg_segment_collector(
req: &mut AggregationWithAccessor,
accessor_idx: usize,
) -> crate::Result<Box<dyn SegmentAggregationCollector>> {
use AggregationVariants::*;
match &req.agg.agg {
Terms(terms_req) => {
if req.accessors.is_empty() {
Ok(Box::new(SegmentTermCollector::from_req_and_validate(
terms_req,
&mut req.sub_aggregation,
req.field_type,
accessor_idx,
)?))
} else {
Ok(Box::new(TermMissingAgg::new(
accessor_idx,
&mut req.sub_aggregation,
)?))
}
}
Range(range_req) => Ok(Box::new(SegmentRangeCollector::from_req_and_validate(
range_req,
&mut req.sub_aggregation,
&mut req.limits,
req.field_type,
accessor_idx,
)?)),
Histogram(histogram) => Ok(Box::new(SegmentHistogramCollector::from_req_and_validate(
histogram.clone(),
&mut req.sub_aggregation,
req.field_type,
accessor_idx,
)?)),
DateHistogram(histogram) => Ok(Box::new(SegmentHistogramCollector::from_req_and_validate(
histogram.to_histogram_req()?,
&mut req.sub_aggregation,
req.field_type,
accessor_idx,
)?)),
Average(AverageAggregation { missing, .. }) => {
Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Average,
accessor_idx,
*missing,
)))
}
Count(CountAggregation { missing, .. }) => Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Count,
accessor_idx,
*missing,
))),
Max(MaxAggregation { missing, .. }) => Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Max,
accessor_idx,
*missing,
))),
Min(MinAggregation { missing, .. }) => Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Min,
accessor_idx,
*missing,
))),
Stats(StatsAggregation { missing, .. }) => Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Stats,
accessor_idx,
*missing,
))),
ExtendedStats(ExtendedStatsAggregation { missing, sigma, .. }) => Ok(Box::new(
SegmentExtendedStatsCollector::from_req(req.field_type, *sigma, accessor_idx, *missing),
)),
Sum(SumAggregation { missing, .. }) => Ok(Box::new(SegmentStatsCollector::from_req(
req.field_type,
SegmentStatsType::Sum,
accessor_idx,
*missing,
))),
Percentiles(percentiles_req) => Ok(Box::new(
SegmentPercentilesCollector::from_req_and_validate(
percentiles_req,
req.field_type,
accessor_idx,
)?,
)),
TopHits(top_hits_req) => Ok(Box::new(TopHitsSegmentCollector::from_req(
top_hits_req,
accessor_idx,
req.segment_ordinal,
))),
Cardinality(CardinalityAggregationReq { missing, .. }) => Ok(Box::new(
SegmentCardinalityCollector::from_req(req.field_type, accessor_idx, missing),
)),
}
}
#[derive(Clone, Default)]
/// The GenericSegmentAggregationResultsCollector is the generic version of the collector, which
/// can handle arbitrary complexity of sub-aggregations. Ideally we never have to pick this one
@@ -74,11 +197,11 @@ impl Debug for GenericSegmentAggregationResultsCollector {
impl SegmentAggregationCollector for GenericSegmentAggregationResultsCollector {
fn add_intermediate_aggregation_result(
self: Box<Self>,
agg_data: &AggregationsSegmentCtx,
agg_with_accessor: &AggregationsWithAccessor,
results: &mut IntermediateAggregationResults,
) -> crate::Result<()> {
for agg in self.aggs {
agg.add_intermediate_aggregation_result(agg_data, results)?;
agg.add_intermediate_aggregation_result(agg_with_accessor, results)?;
}
Ok(())
@@ -87,9 +210,9 @@ impl SegmentAggregationCollector for GenericSegmentAggregationResultsCollector {
fn collect(
&mut self,
doc: crate::DocId,
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
self.collect_block(&[doc], agg_data)?;
self.collect_block(&[doc], agg_with_accessor)?;
Ok(())
}
@@ -97,19 +220,32 @@ impl SegmentAggregationCollector for GenericSegmentAggregationResultsCollector {
fn collect_block(
&mut self,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
agg_with_accessor: &mut AggregationsWithAccessor,
) -> crate::Result<()> {
for collector in &mut self.aggs {
collector.collect_block(docs, agg_data)?;
collector.collect_block(docs, agg_with_accessor)?;
}
Ok(())
}
fn flush(&mut self, agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
fn flush(&mut self, agg_with_accessor: &mut AggregationsWithAccessor) -> crate::Result<()> {
for collector in &mut self.aggs {
collector.flush(agg_data)?;
collector.flush(agg_with_accessor)?;
}
Ok(())
}
}
impl GenericSegmentAggregationResultsCollector {
pub(crate) fn from_req_and_validate(req: &mut AggregationsWithAccessor) -> crate::Result<Self> {
let aggs = req
.aggs
.values_mut()
.enumerate()
.map(|(accessor_idx, req)| build_single_agg_segment_collector(req, accessor_idx))
.collect::<crate::Result<Vec<Box<dyn SegmentAggregationCollector>>>>()?;
Ok(GenericSegmentAggregationResultsCollector { aggs })
}
}

121
src/collector/custom_score_top_collector.rs Normal file
View File

@@ -0,0 +1,121 @@
use crate::collector::top_collector::{TopCollector, TopSegmentCollector};
use crate::collector::{Collector, SegmentCollector};
use crate::{DocAddress, DocId, Score, SegmentReader};
pub(crate) struct CustomScoreTopCollector<TCustomScorer, TScore = Score> {
custom_scorer: TCustomScorer,
collector: TopCollector<TScore>,
}
impl<TCustomScorer, TScore> CustomScoreTopCollector<TCustomScorer, TScore>
where TScore: Clone + PartialOrd
{
pub(crate) fn new(
custom_scorer: TCustomScorer,
collector: TopCollector<TScore>,
) -> CustomScoreTopCollector<TCustomScorer, TScore> {
CustomScoreTopCollector {
custom_scorer,
collector,
}
}
}
/// A custom segment scorer makes it possible to define any kind of score
/// for a given document belonging to a specific segment.
///
/// It is the segment local version of the [`CustomScorer`].
pub trait CustomSegmentScorer<TScore>: 'static {
/// Computes the score of a specific `doc`.
fn score(&mut self, doc: DocId) -> TScore;
}
/// `CustomScorer` makes it possible to define any kind of score.
///
/// The `CustomerScorer` itself does not make much of the computation itself.
/// Instead, it helps constructing `Self::Child` instances that will compute
/// the score at a segment scale.
pub trait CustomScorer<TScore>: Sync {
/// Type of the associated [`CustomSegmentScorer`].
type Child: CustomSegmentScorer<TScore>;
/// Builds a child scorer for a specific segment. The child scorer is associated with
/// a specific segment.
fn segment_scorer(&self, segment_reader: &SegmentReader) -> crate::Result<Self::Child>;
}
impl<TCustomScorer, TScore> Collector for CustomScoreTopCollector<TCustomScorer, TScore>
where
TCustomScorer: CustomScorer<TScore> + Send + Sync,
TScore: 'static + PartialOrd + Clone + Send + Sync,
{
type Fruit = Vec<(TScore, DocAddress)>;
type Child = CustomScoreTopSegmentCollector<TCustomScorer::Child, TScore>;
fn for_segment(
&self,
segment_local_id: u32,
segment_reader: &SegmentReader,
) -> crate::Result<Self::Child> {
let segment_collector = self.collector.for_segment(segment_local_id, segment_reader);
let segment_scorer = self.custom_scorer.segment_scorer(segment_reader)?;
Ok(CustomScoreTopSegmentCollector {
segment_collector,
segment_scorer,
})
}
fn requires_scoring(&self) -> bool {
false
}
fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> crate::Result<Self::Fruit> {
self.collector.merge_fruits(segment_fruits)
}
}
pub struct CustomScoreTopSegmentCollector<T, TScore>
where
TScore: 'static + PartialOrd + Clone + Send + Sync + Sized,
T: CustomSegmentScorer<TScore>,
{
segment_collector: TopSegmentCollector<TScore>,
segment_scorer: T,
}
impl<T, TScore> SegmentCollector for CustomScoreTopSegmentCollector<T, TScore>
where
TScore: 'static + PartialOrd + Clone + Send + Sync,
T: 'static + CustomSegmentScorer<TScore>,
{
type Fruit = Vec<(TScore, DocAddress)>;
fn collect(&mut self, doc: DocId, _score: Score) {
let score = self.segment_scorer.score(doc);
self.segment_collector.collect(doc, score);
}
fn harvest(self) -> Vec<(TScore, DocAddress)> {
self.segment_collector.harvest()
}
}
impl<F, TScore, T> CustomScorer<TScore> for F
where
F: 'static + Send + Sync + Fn(&SegmentReader) -> T,
T: CustomSegmentScorer<TScore>,
{
type Child = T;
fn segment_scorer(&self, segment_reader: &SegmentReader) -> crate::Result<Self::Child> {
Ok((self)(segment_reader))
}
}
impl<F, TScore> CustomSegmentScorer<TScore> for F
where F: 'static + FnMut(DocId) -> TScore
{
fn score(&mut self, doc: DocId) -> TScore {
(self)(doc)
}
}

48
src/collector/filter_collector_wrapper.rs
View File

@@ -12,7 +12,6 @@ use std::marker::PhantomData;
use columnar::{BytesColumn, Column, DynamicColumn, HasAssociatedColumnType};
use crate::collector::{Collector, SegmentCollector};
use crate::schema::Schema;
use crate::{DocId, Score, SegmentReader};
/// The `FilterCollector` filters docs using a fast field value and a predicate.
@@ -50,13 +49,13 @@ use crate::{DocId, Score, SegmentReader};
///
/// let query_parser = QueryParser::for_index(&index, vec![title]);
/// let query = query_parser.parse_query("diary")?;
/// let no_filter_collector = FilterCollector::new("price".to_string(), |value: u64| value > 20_120u64, TopDocs::with_limit(2).order_by_score());
/// let no_filter_collector = FilterCollector::new("price".to_string(), |value: u64| value > 20_120u64, TopDocs::with_limit(2));
/// let top_docs = searcher.search(&query, &no_filter_collector)?;
///
/// assert_eq!(top_docs.len(), 1);
/// assert_eq!(top_docs[0].1, DocAddress::new(0, 1));
///
/// let filter_all_collector: FilterCollector<_, _, u64> = FilterCollector::new("price".to_string(), |value| value < 5u64, TopDocs::with_limit(2).order_by_score());
/// let filter_all_collector: FilterCollector<_, _, u64> = FilterCollector::new("price".to_string(), |value| value < 5u64, TopDocs::with_limit(2));
/// let filtered_top_docs = searcher.search(&query, &filter_all_collector)?;
///
/// assert_eq!(filtered_top_docs.len(), 0);
@@ -105,11 +104,6 @@ where
type Child = FilterSegmentCollector<TCollector::Child, TPredicate, TPredicateValue>;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.collector.check_schema(schema)?;
Ok(())
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -126,7 +120,6 @@ where
segment_collector,
predicate: self.predicate.clone(),
t_predicate_value: PhantomData,
filtered_docs: Vec::with_capacity(crate::COLLECT_BLOCK_BUFFER_LEN),
})
}
@@ -147,7 +140,6 @@ pub struct FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue
segment_collector: TSegmentCollector,
predicate: TPredicate,
t_predicate_value: PhantomData<TPredicateValue>,
filtered_docs: Vec<DocId>,
}
impl<TSegmentCollector, TPredicate, TPredicateValue>
@@ -184,20 +176,6 @@ where
}
}
fn collect_block(&mut self, docs: &[DocId]) {
self.filtered_docs.clear();
for &doc in docs {
// TODO: `accept_document` could be further optimized to do batch lookups of column
// values for single-valued columns.
if self.accept_document(doc) {
self.filtered_docs.push(doc);
}
}
if !self.filtered_docs.is_empty() {
self.segment_collector.collect_block(&self.filtered_docs);
}
}
fn harvest(self) -> TSegmentCollector::Fruit {
self.segment_collector.harvest()
}
@@ -240,7 +218,7 @@ where
///
/// let query_parser = QueryParser::for_index(&index, vec![title]);
/// let query = query_parser.parse_query("diary")?;
/// let filter_collector = BytesFilterCollector::new("barcode".to_string(), |bytes: &[u8]| bytes.starts_with(b"01"), TopDocs::with_limit(2).order_by_score());
/// let filter_collector = BytesFilterCollector::new("barcode".to_string(), |bytes: &[u8]| bytes.starts_with(b"01"), TopDocs::with_limit(2));
/// let top_docs = searcher.search(&query, &filter_collector)?;
///
/// assert_eq!(top_docs.len(), 1);
@@ -280,10 +258,6 @@ where
type Child = BytesFilterSegmentCollector<TCollector::Child, TPredicate>;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.collector.check_schema(schema)
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -300,7 +274,6 @@ where
segment_collector,
predicate: self.predicate.clone(),
buffer: Vec::new(),
filtered_docs: Vec::with_capacity(crate::COLLECT_BLOCK_BUFFER_LEN),
})
}
@@ -323,7 +296,6 @@ where TPredicate: 'static
segment_collector: TSegmentCollector,
predicate: TPredicate,
buffer: Vec<u8>,
filtered_docs: Vec<DocId>,
}
impl<TSegmentCollector, TPredicate> BytesFilterSegmentCollector<TSegmentCollector, TPredicate>
@@ -362,20 +334,6 @@ where
}
}
fn collect_block(&mut self, docs: &[DocId]) {
self.filtered_docs.clear();
for &doc in docs {
// TODO: `accept_document` could be further optimized to do batch lookups of column
// values for single-valued columns.
if self.accept_document(doc) {
self.filtered_docs.push(doc);
}
}
if !self.filtered_docs.is_empty() {
self.segment_collector.collect_block(&self.filtered_docs);
}
}
fn harvest(self) -> TSegmentCollector::Fruit {
self.segment_collector.harvest()
}

141
src/collector/mod.rs
View File

@@ -57,7 +57,7 @@
//! # let query_parser = QueryParser::for_index(&index, vec![title]);
//! # let query = query_parser.parse_query("diary")?;
//! let (doc_count, top_docs): (usize, Vec<(Score, DocAddress)>) =
//! searcher.search(&query, &(Count, TopDocs::with_limit(2).order_by_score()))?;
//! searcher.search(&query, &(Count, TopDocs::with_limit(2)))?;
//! # Ok(())
//! # }
//! ```
@@ -83,15 +83,11 @@
use downcast_rs::impl_downcast;
use crate::schema::Schema;
use crate::{DocId, Score, SegmentOrdinal, SegmentReader};
mod count_collector;
pub use self::count_collector::Count;
/// Sort keys
pub mod sort_key;
mod histogram_collector;
pub use histogram_collector::HistogramCollector;
@@ -99,13 +95,16 @@ mod multi_collector;
pub use self::multi_collector::{FruitHandle, MultiCollector, MultiFruit};
mod top_collector;
pub use self::top_collector::ComparableDoc;
mod top_score_collector;
pub use self::top_collector::ComparableDoc;
pub use self::top_score_collector::{TopDocs, TopNComputer};
mod sort_key_top_collector;
pub use self::sort_key::{SegmentSortKeyComputer, SortKeyComputer};
mod custom_score_top_collector;
pub use self::custom_score_top_collector::{CustomScorer, CustomSegmentScorer};
mod tweak_score_top_collector;
pub use self::tweak_score_top_collector::{ScoreSegmentTweaker, ScoreTweaker};
mod facet_collector;
pub use self::facet_collector::{FacetCollector, FacetCounts};
use crate::query::Weight;
@@ -146,11 +145,6 @@ pub trait Collector: Sync + Send {
/// Type of the `SegmentCollector` associated with this collector.
type Child: SegmentCollector;
/// Returns an error if the schema is not compatible with the collector.
fn check_schema(&self, _schema: &Schema) -> crate::Result<()> {
Ok(())
}
/// `set_segment` is called before beginning to enumerate
/// on this segment.
fn for_segment(
@@ -176,50 +170,41 @@ pub trait Collector: Sync + Send {
segment_ord: u32,
reader: &SegmentReader,
) -> crate::Result<<Self::Child as SegmentCollector>::Fruit> {
let with_scoring = self.requires_scoring();
let mut segment_collector = self.for_segment(segment_ord, reader)?;
default_collect_segment_impl(&mut segment_collector, weight, reader, with_scoring)?;
match (reader.alive_bitset(), self.requires_scoring()) {
(Some(alive_bitset), true) => {
weight.for_each(reader, &mut |doc, score| {
if alive_bitset.is_alive(doc) {
segment_collector.collect(doc, score);
}
})?;
}
(Some(alive_bitset), false) => {
weight.for_each_no_score(reader, &mut |docs| {
for doc in docs.iter().cloned() {
if alive_bitset.is_alive(doc) {
segment_collector.collect(doc, 0.0);
}
}
})?;
}
(None, true) => {
weight.for_each(reader, &mut |doc, score| {
segment_collector.collect(doc, score);
})?;
}
(None, false) => {
weight.for_each_no_score(reader, &mut |docs| {
segment_collector.collect_block(docs);
})?;
}
}
Ok(segment_collector.harvest())
}
}
pub(crate) fn default_collect_segment_impl<TSegmentCollector: SegmentCollector>(
segment_collector: &mut TSegmentCollector,
weight: &dyn Weight,
reader: &SegmentReader,
with_scoring: bool,
) -> crate::Result<()> {
match (reader.alive_bitset(), with_scoring) {
(Some(alive_bitset), true) => {
weight.for_each(reader, &mut |doc, score| {
if alive_bitset.is_alive(doc) {
segment_collector.collect(doc, score);
}
})?;
}
(Some(alive_bitset), false) => {
weight.for_each_no_score(reader, &mut |docs| {
for doc in docs.iter().cloned() {
if alive_bitset.is_alive(doc) {
segment_collector.collect(doc, 0.0);
}
}
})?;
}
(None, true) => {
weight.for_each(reader, &mut |doc, score| {
segment_collector.collect(doc, score);
})?;
}
(None, false) => {
weight.for_each_no_score(reader, &mut |docs| {
segment_collector.collect_block(docs);
})?;
}
}
Ok(())
}
impl<TSegmentCollector: SegmentCollector> SegmentCollector for Option<TSegmentCollector> {
type Fruit = Option<TSegmentCollector::Fruit>;
@@ -229,12 +214,6 @@ impl<TSegmentCollector: SegmentCollector> SegmentCollector for Option<TSegmentCo
}
}
fn collect_block(&mut self, docs: &[DocId]) {
if let Some(segment_collector) = self {
segment_collector.collect_block(docs);
}
}
fn harvest(self) -> Self::Fruit {
self.map(|segment_collector| segment_collector.harvest())
}
@@ -245,13 +224,6 @@ impl<TCollector: Collector> Collector for Option<TCollector> {
type Child = Option<<TCollector as Collector>::Child>;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
if let Some(underlying_collector) = self {
underlying_collector.check_schema(schema)?;
}
Ok(())
}
fn for_segment(
&self,
segment_local_id: SegmentOrdinal,
@@ -327,12 +299,6 @@ where
type Fruit = (Left::Fruit, Right::Fruit);
type Child = (Left::Child, Right::Child);
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
Ok(())
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -376,11 +342,6 @@ where
self.1.collect(doc, score);
}
fn collect_block(&mut self, docs: &[DocId]) {
self.0.collect_block(docs);
self.1.collect_block(docs);
}
fn harvest(self) -> <Self as SegmentCollector>::Fruit {
(self.0.harvest(), self.1.harvest())
}
@@ -397,13 +358,6 @@ where
type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
type Child = (One::Child, Two::Child, Three::Child);
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
self.2.check_schema(schema)?;
Ok(())
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -453,12 +407,6 @@ where
self.2.collect(doc, score);
}
fn collect_block(&mut self, docs: &[DocId]) {
self.0.collect_block(docs);
self.1.collect_block(docs);
self.2.collect_block(docs);
}
fn harvest(self) -> <Self as SegmentCollector>::Fruit {
(self.0.harvest(), self.1.harvest(), self.2.harvest())
}
@@ -476,14 +424,6 @@ where
type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
type Child = (One::Child, Two::Child, Three::Child, Four::Child);
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
self.2.check_schema(schema)?;
self.3.check_schema(schema)?;
Ok(())
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -542,13 +482,6 @@ where
self.3.collect(doc, score);
}
fn collect_block(&mut self, docs: &[DocId]) {
self.0.collect_block(docs);
self.1.collect_block(docs);
self.2.collect_block(docs);
self.3.collect_block(docs);
}
fn harvest(self) -> <Self as SegmentCollector>::Fruit {
(
self.0.harvest(),

22
src/collector/multi_collector.rs
View File

@@ -3,7 +3,6 @@ use std::ops::Deref;
use super::{Collector, SegmentCollector};
use crate::collector::Fruit;
use crate::schema::Schema;
use crate::{DocId, Score, SegmentOrdinal, SegmentReader, TantivyError};
/// MultiFruit keeps Fruits from every nested Collector
@@ -17,10 +16,6 @@ impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
type Fruit = Box<dyn Fruit>;
type Child = Box<dyn BoxableSegmentCollector>;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)
}
fn for_segment(
&self,
segment_local_id: u32,
@@ -152,7 +147,7 @@ impl<TFruit: Fruit> FruitHandle<TFruit> {
/// let searcher = reader.searcher();
///
/// let mut collectors = MultiCollector::new();
/// let top_docs_handle = collectors.add_collector(TopDocs::with_limit(2).order_by_score());
/// let top_docs_handle = collectors.add_collector(TopDocs::with_limit(2));
/// let count_handle = collectors.add_collector(Count);
/// let query_parser = QueryParser::for_index(&index, vec![title]);
/// let query = query_parser.parse_query("diary").unwrap();
@@ -199,13 +194,6 @@ impl Collector for MultiCollector<'_> {
type Fruit = MultiFruit;
type Child = MultiCollectorChild;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
for collector in &self.collector_wrappers {
collector.check_schema(schema)?;
}
Ok(())
}
fn for_segment(
&self,
segment_local_id: SegmentOrdinal,
@@ -262,12 +250,6 @@ impl SegmentCollector for MultiCollectorChild {
}
}
fn collect_block(&mut self, docs: &[DocId]) {
for child in &mut self.children {
child.collect_block(docs);
}
}
fn harvest(self) -> MultiFruit {
MultiFruit {
sub_fruits: self
@@ -311,7 +293,7 @@ mod tests {
let query = TermQuery::new(term, IndexRecordOption::Basic);
let mut collectors = MultiCollector::new();
let topdocs_handler = collectors.add_collector(TopDocs::with_limit(2).order_by_score());
let topdocs_handler = collectors.add_collector(TopDocs::with_limit(2));
let count_handler = collectors.add_collector(Count);
let mut multifruits = searcher.search(&query, &collectors).unwrap();

454
src/collector/sort_key/mod.rs
View File

@@ -1,454 +0,0 @@
mod order;
mod sort_by_erased_type;
mod sort_by_score;
mod sort_by_static_fast_value;
mod sort_by_string;
mod sort_key_computer;
pub use order::*;
pub use sort_by_erased_type::SortByErasedType;
pub use sort_by_score::SortBySimilarityScore;
pub use sort_by_static_fast_value::SortByStaticFastValue;
pub use sort_by_string::SortByString;
pub use sort_key_computer::{SegmentSortKeyComputer, SortKeyComputer};
#[cfg(test)]
pub(crate) mod tests {
// By spec, regardless of whether ascending or descending order was requested, in presence of a
// tie, we sort by ascending doc id/doc address.
pub(crate) fn sort_hits<TSortKey: Ord, D: Ord>(
hits: &mut [ComparableDoc<TSortKey, D>],
order: Order,
) {
if order.is_asc() {
hits.sort_by(|l, r| l.sort_key.cmp(&r.sort_key).then(l.doc.cmp(&r.doc)));
} else {
hits.sort_by(|l, r| {
l.sort_key
.cmp(&r.sort_key)
.reverse() // This is descending
.then(l.doc.cmp(&r.doc))
});
}
}
use std::collections::HashMap;
use std::ops::Range;
use crate::collector::sort_key::{
SortByErasedType, SortBySimilarityScore, SortByStaticFastValue, SortByString,
};
use crate::collector::{ComparableDoc, DocSetCollector, TopDocs};
use crate::indexer::NoMergePolicy;
use crate::query::{AllQuery, QueryParser};
use crate::schema::{OwnedValue, Schema, FAST, TEXT};
use crate::{DocAddress, Document, Index, Order, Score, Searcher};
fn make_index() -> crate::Result<Index> {
let mut schema_builder = Schema::builder();
let id = schema_builder.add_u64_field("id", FAST);
let city = schema_builder.add_text_field("city", TEXT | FAST);
let catchphrase = schema_builder.add_text_field("catchphrase", TEXT);
let altitude = schema_builder.add_f64_field("altitude", FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
fn create_segment(index: &Index, docs: Vec<impl Document>) -> crate::Result<()> {
let mut index_writer = index.writer_for_tests()?;
index_writer.set_merge_policy(Box::new(NoMergePolicy));
for doc in docs {
index_writer.add_document(doc)?;
}
index_writer.commit()?;
Ok(())
}
create_segment(
&index,
vec![
doc!(
id => 0_u64,
city => "austin",
catchphrase => "Hills, Barbeque, Glow",
altitude => 149.0,
),
doc!(
id => 1_u64,
city => "greenville",
catchphrase => "Grow, Glow, Glow",
altitude => 27.0,
),
],
)?;
create_segment(
&index,
vec![doc!(
id => 2_u64,
city => "tokyo",
catchphrase => "Glow, Glow, Glow",
altitude => 40.0,
)],
)?;
create_segment(
&index,
vec![doc!(
id => 3_u64,
catchphrase => "No, No, No",
altitude => 0.0,
)],
)?;
Ok(index)
}
// NOTE: You cannot determine the SegmentIds that will be generated for Segments
// ahead of time, so DocAddresses must be mapped back to a unique id for each Searcher.
fn id_mapping(searcher: &Searcher) -> HashMap<DocAddress, u64> {
searcher
.search(&AllQuery, &DocSetCollector)
.unwrap()
.into_iter()
.map(|doc_address| {
let column = searcher.segment_readers()[doc_address.segment_ord as usize]
.fast_fields()
.u64("id")
.unwrap();
(doc_address, column.first(doc_address.doc_id).unwrap())
})
.collect()
}
#[test]
fn test_order_by_string() -> crate::Result<()> {
let index = make_index()?;
#[track_caller]
fn assert_query(
index: &Index,
order: Order,
doc_range: Range<usize>,
expected: Vec<(Option<String>, u64)>,
) -> crate::Result<()> {
let searcher = index.reader()?.searcher();
let ids = id_mapping(&searcher);
// Try as primitive.
let top_collector = TopDocs::for_doc_range(doc_range)
.order_by((SortByString::for_field("city"), order));
let actual = searcher
.search(&AllQuery, &top_collector)?
.into_iter()
.map(|(sort_key_opt, doc)| (sort_key_opt, ids[&doc]))
.collect::<Vec<_>>();
assert_eq!(actual, expected);
Ok(())
}
assert_query(
&index,
Order::Asc,
0..4,
vec![
(Some("austin".to_owned()), 0),
(Some("greenville".to_owned()), 1),
(Some("tokyo".to_owned()), 2),
(None, 3),
],
)?;
assert_query(
&index,
Order::Asc,
0..3,
vec![
(Some("austin".to_owned()), 0),
(Some("greenville".to_owned()), 1),
(Some("tokyo".to_owned()), 2),
],
)?;
assert_query(
&index,
Order::Asc,
0..2,
vec![
(Some("austin".to_owned()), 0),
(Some("greenville".to_owned()), 1),
],
)?;
assert_query(
&index,
Order::Asc,
0..1,
vec![(Some("austin".to_string()), 0)],
)?;
assert_query(
&index,
Order::Asc,
1..3,
vec![
(Some("greenville".to_owned()), 1),
(Some("tokyo".to_owned()), 2),
],
)?;
assert_query(
&index,
Order::Desc,
0..4,
vec![
(Some("tokyo".to_owned()), 2),
(Some("greenville".to_owned()), 1),
(Some("austin".to_owned()), 0),
(None, 3),
],
)?;
assert_query(
&index,
Order::Desc,
1..3,
vec![
(Some("greenville".to_owned()), 1),
(Some("austin".to_owned()), 0),
],
)?;
assert_query(
&index,
Order::Desc,
0..1,
vec![(Some("tokyo".to_owned()), 2)],
)?;
Ok(())
}
#[test]
fn test_order_by_f64() -> crate::Result<()> {
let index = make_index()?;
fn assert_query(
index: &Index,
order: Order,
expected: Vec<(Option<f64>, u64)>,
) -> crate::Result<()> {
let searcher = index.reader()?.searcher();
let ids = id_mapping(&searcher);
// Try as primitive.
let top_collector = TopDocs::with_limit(3)
.order_by((SortByStaticFastValue::<f64>::for_field("altitude"), order));
let actual = searcher
.search(&AllQuery, &top_collector)?
.into_iter()
.map(|(altitude_opt, doc)| (altitude_opt, ids[&doc]))
.collect::<Vec<_>>();
assert_eq!(actual, expected);
Ok(())
}
assert_query(
&index,
Order::Asc,
vec![(Some(0.0), 3), (Some(27.0), 1), (Some(40.0), 2)],
)?;
assert_query(
&index,
Order::Desc,
vec![(Some(149.0), 0), (Some(40.0), 2), (Some(27.0), 1)],
)?;
Ok(())
}
#[test]
fn test_order_by_score() -> crate::Result<()> {
let index = make_index()?;
fn query(index: &Index, order: Order) -> crate::Result<Vec<(Score, u64)>> {
let searcher = index.reader()?.searcher();
let ids = id_mapping(&searcher);
let top_collector = TopDocs::with_limit(4).order_by((SortBySimilarityScore, order));
let field = index.schema().get_field("catchphrase").unwrap();
let query_parser = QueryParser::for_index(index, vec![field]);
let text_query = query_parser.parse_query("glow")?;
Ok(searcher
.search(&text_query, &top_collector)?
.into_iter()
.map(|(score, doc)| (score, ids[&doc]))
.collect())
}
assert_eq!(
&query(&index, Order::Desc)?,
&[(0.5604893, 2), (0.4904281, 1), (0.35667497, 0),]
);
assert_eq!(
&query(&index, Order::Asc)?,
&[(0.35667497, 0), (0.4904281, 1), (0.5604893, 2),]
);
Ok(())
}
#[test]
fn test_order_by_score_then_string() -> crate::Result<()> {
let index = make_index()?;
type SortKey = (Score, Option<String>);
fn query(
index: &Index,
score_order: Order,
city_order: Order,
) -> crate::Result<Vec<(SortKey, u64)>> {
let searcher = index.reader()?.searcher();
let ids = id_mapping(&searcher);
let top_collector = TopDocs::with_limit(4).order_by((
(SortBySimilarityScore, score_order),
(SortByString::for_field("city"), city_order),
));
let results: Vec<((Score, Option<String>), DocAddress)> =
searcher.search(&AllQuery, &top_collector)?;
Ok(results.into_iter().map(|(f, doc)| (f, ids[&doc])).collect())
}
assert_eq!(
&query(&index, Order::Asc, Order::Asc)?,
&[
((1.0, Some("austin".to_owned())), 0),
((1.0, Some("greenville".to_owned())), 1),
((1.0, Some("tokyo".to_owned())), 2),
((1.0, None), 3),
]
);
assert_eq!(
&query(&index, Order::Asc, Order::Desc)?,
&[
((1.0, Some("tokyo".to_owned())), 2),
((1.0, Some("greenville".to_owned())), 1),
((1.0, Some("austin".to_owned())), 0),
((1.0, None), 3),
]
);
Ok(())
}
#[test]
fn test_order_by_score_then_owned_value() -> crate::Result<()> {
let index = make_index()?;
type SortKey = (Score, OwnedValue);
fn query(
index: &Index,
score_order: Order,
city_order: Order,
) -> crate::Result<Vec<(SortKey, u64)>> {
let searcher = index.reader()?.searcher();
let ids = id_mapping(&searcher);
let top_collector = TopDocs::with_limit(4).order_by::<(Score, OwnedValue)>((
(SortBySimilarityScore, score_order),
(SortByErasedType::for_field("city"), city_order),
));
let results: Vec<((Score, OwnedValue), DocAddress)> =
searcher.search(&AllQuery, &top_collector)?;
Ok(results.into_iter().map(|(f, doc)| (f, ids[&doc])).collect())
}
assert_eq!(
&query(&index, Order::Asc, Order::Asc)?,
&[
((1.0, OwnedValue::Str("austin".to_owned())), 0),
((1.0, OwnedValue::Str("greenville".to_owned())), 1),
((1.0, OwnedValue::Str("tokyo".to_owned())), 2),
((1.0, OwnedValue::Null), 3),
]
);
assert_eq!(
&query(&index, Order::Asc, Order::Desc)?,
&[
((1.0, OwnedValue::Str("tokyo".to_owned())), 2),
((1.0, OwnedValue::Str("greenville".to_owned())), 1),
((1.0, OwnedValue::Str("austin".to_owned())), 0),
((1.0, OwnedValue::Null), 3),
]
);
Ok(())
}
use proptest::prelude::*;
proptest! {
#[test]
fn test_order_by_string_prop(
order in prop_oneof!(Just(Order::Desc), Just(Order::Asc)),
limit in 1..64_usize,
offset in 0..64_usize,
segments_terms in
proptest::collection::vec(
proptest::collection::vec(0..32_u8, 1..32_usize),
0..8_usize,
)
) {
let mut schema_builder = Schema::builder();
let city = schema_builder.add_text_field("city", TEXT | FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut index_writer = index.writer_for_tests()?;
// A Vec<Vec<u8>>, where the outer Vec represents segments, and the inner Vec
// represents terms.
for segment_terms in segments_terms.into_iter() {
for term in segment_terms.into_iter() {
let term = format!("{term:0>3}");
index_writer.add_document(doc!(
city => term,
))?;
}
index_writer.commit()?;
}
let searcher = index.reader()?.searcher();
let top_n_results = searcher.search(&AllQuery, &TopDocs::with_limit(limit)
.and_offset(offset)
.order_by_string_fast_field("city", order))?;
let all_results = searcher.search(&AllQuery, &DocSetCollector)?.into_iter().map(|doc_address| {
// Get the term for this address.
let column = searcher.segment_readers()[doc_address.segment_ord as usize].fast_fields().str("city").unwrap().unwrap();
let value = column.term_ords(doc_address.doc_id).next().map(|term_ord| {
let mut city = Vec::new();
column.dictionary().ord_to_term(term_ord, &mut city).unwrap();
String::try_from(city).unwrap()
});
(value, doc_address)
});
// Using the TopDocs collector should always be equivalent to sorting, skipping the
// offset, and then taking the limit.
let sorted_docs: Vec<_> = {
let mut comparable_docs: Vec<ComparableDoc<_, _>> =
all_results.into_iter().map(|(sort_key, doc)| ComparableDoc { sort_key, doc}).collect();
sort_hits(&mut comparable_docs, order);
comparable_docs.into_iter().map(|cd| (cd.sort_key, cd.doc)).collect()
};
let expected_docs = sorted_docs.into_iter().skip(offset).take(limit).collect::<Vec<_>>();
prop_assert_eq!(
expected_docs,
top_n_results
);
}
}
}

567
src/collector/sort_key/order.rs
View File

@@ -1,567 +0,0 @@
use std::cmp::Ordering;
use columnar::MonotonicallyMappableToU64;
use serde::{Deserialize, Serialize};
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer};
use crate::schema::{OwnedValue, Schema};
use crate::{DocId, Order, Score};
fn compare_owned_value<const NULLS_FIRST: bool>(lhs: &OwnedValue, rhs: &OwnedValue) -> Ordering {
match (lhs, rhs) {
(OwnedValue::Null, OwnedValue::Null) => Ordering::Equal,
(OwnedValue::Null, _) => {
if NULLS_FIRST {
Ordering::Less
} else {
Ordering::Greater
}
}
(_, OwnedValue::Null) => {
if NULLS_FIRST {
Ordering::Greater
} else {
Ordering::Less
}
}
(OwnedValue::Str(a), OwnedValue::Str(b)) => a.cmp(b),
(OwnedValue::PreTokStr(a), OwnedValue::PreTokStr(b)) => a.cmp(b),
(OwnedValue::U64(a), OwnedValue::U64(b)) => a.cmp(b),
(OwnedValue::I64(a), OwnedValue::I64(b)) => a.cmp(b),
(OwnedValue::F64(a), OwnedValue::F64(b)) => a.to_u64().cmp(&b.to_u64()),
(OwnedValue::Bool(a), OwnedValue::Bool(b)) => a.cmp(b),
(OwnedValue::Date(a), OwnedValue::Date(b)) => a.cmp(b),
(OwnedValue::Facet(a), OwnedValue::Facet(b)) => a.cmp(b),
(OwnedValue::Bytes(a), OwnedValue::Bytes(b)) => a.cmp(b),
(OwnedValue::IpAddr(a), OwnedValue::IpAddr(b)) => a.cmp(b),
(OwnedValue::U64(a), OwnedValue::I64(b)) => {
if *b < 0 {
Ordering::Greater
} else {
a.cmp(&(*b as u64))
}
}
(OwnedValue::I64(a), OwnedValue::U64(b)) => {
if *a < 0 {
Ordering::Less
} else {
(*a as u64).cmp(b)
}
}
(OwnedValue::U64(a), OwnedValue::F64(b)) => (*a as f64).to_u64().cmp(&b.to_u64()),
(OwnedValue::F64(a), OwnedValue::U64(b)) => a.to_u64().cmp(&(*b as f64).to_u64()),
(OwnedValue::I64(a), OwnedValue::F64(b)) => (*a as f64).to_u64().cmp(&b.to_u64()),
(OwnedValue::F64(a), OwnedValue::I64(b)) => a.to_u64().cmp(&(*b as f64).to_u64()),
(a, b) => {
let ord = a.discriminant_value().cmp(&b.discriminant_value());
// If the discriminant is equal, it's because a new type was added, but hasn't been
// included in this `match` statement.
assert!(
ord != Ordering::Equal,
"Unimplemented comparison for type of {a:?}, {b:?}"
);
ord
}
}
}
/// Comparator trait defining the order in which documents should be ordered.
pub trait Comparator<T>: Send + Sync + std::fmt::Debug + Default {
/// Return the order between two values.
fn compare(&self, lhs: &T, rhs: &T) -> Ordering;
}
/// Compare values naturally (e.g. 1 < 2).
///
/// When used with `TopDocs`, which reverses the order, this results in a
/// "Descending" sort (Greatest values first).
///
/// `None` (or Null for `OwnedValue`) values are considered to be smaller than any other value,
/// and will therefore appear last in a descending sort (e.g. `[Some(20), Some(10), None]`).
#[derive(Debug, Copy, Clone, Default, Serialize, Deserialize)]
pub struct NaturalComparator;
impl<T: PartialOrd> Comparator<T> for NaturalComparator {
#[inline(always)]
fn compare(&self, lhs: &T, rhs: &T) -> Ordering {
lhs.partial_cmp(rhs).unwrap_or(Ordering::Equal)
}
}
/// A (partial) implementation of comparison for OwnedValue.
///
/// Intended for use within columns of homogenous types, and so will panic for OwnedValues with
/// mismatched types. The one exception is Null, for which we do define all comparisons.
impl Comparator<OwnedValue> for NaturalComparator {
#[inline(always)]
fn compare(&self, lhs: &OwnedValue, rhs: &OwnedValue) -> Ordering {
compare_owned_value::</* NULLS_FIRST= */ true>(lhs, rhs)
}
}
/// Compare values in reverse (e.g. 2 < 1).
///
/// When used with `TopDocs`, which reverses the order, this results in an
/// "Ascending" sort (Smallest values first).
///
/// `None` is considered smaller than `Some` in the underlying comparator, but because the
/// comparison is reversed, `None` is effectively treated as the lowest value in the resulting
/// Ascending sort (e.g. `[None, Some(10), Some(20)]`).
///
/// The ReverseComparator does not necessarily imply that the sort order is reversed compared
/// to the NaturalComparator. In presence of a tie on the sort key, documents will always be
/// sorted by ascending `DocId`/`DocAddress` in TopN results, regardless of the sort key's order.
#[derive(Debug, Copy, Clone, Default, Serialize, Deserialize)]
pub struct ReverseComparator;
impl<T> Comparator<T> for ReverseComparator
where NaturalComparator: Comparator<T>
{
#[inline(always)]
fn compare(&self, lhs: &T, rhs: &T) -> Ordering {
NaturalComparator.compare(rhs, lhs)
}
}
/// Compare values in reverse, but treating `None` as lower than `Some`.
///
/// When used with `TopDocs`, which reverses the order, this results in an
/// "Ascending" sort (Smallest values first), but with `None` values appearing last
/// (e.g. `[Some(10), Some(20), None]`).
///
/// This is usually what is wanted when sorting by a field in an ascending order.
/// For instance, in an e-commerce website, if sorting by price ascending,
/// the cheapest items would appear first, and items without a price would appear last.
#[derive(Debug, Copy, Clone, Default)]
pub struct ReverseNoneIsLowerComparator;
impl<T> Comparator<Option<T>> for ReverseNoneIsLowerComparator
where ReverseComparator: Comparator<T>
{
#[inline(always)]
fn compare(&self, lhs_opt: &Option<T>, rhs_opt: &Option<T>) -> Ordering {
match (lhs_opt, rhs_opt) {
(None, None) => Ordering::Equal,
(None, Some(_)) => Ordering::Less,
(Some(_), None) => Ordering::Greater,
(Some(lhs), Some(rhs)) => ReverseComparator.compare(lhs, rhs),
}
}
}
impl Comparator<u32> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &u32, rhs: &u32) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<u64> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &u64, rhs: &u64) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<f64> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &f64, rhs: &f64) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<f32> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &f32, rhs: &f32) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<i64> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &i64, rhs: &i64) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<String> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &String, rhs: &String) -> Ordering {
ReverseComparator.compare(lhs, rhs)
}
}
impl Comparator<OwnedValue> for ReverseNoneIsLowerComparator {
#[inline(always)]
fn compare(&self, lhs: &OwnedValue, rhs: &OwnedValue) -> Ordering {
compare_owned_value::</* NULLS_FIRST= */ false>(rhs, lhs)
}
}
/// Compare values naturally, but treating `None` as higher than `Some`.
///
/// When used with `TopDocs`, which reverses the order, this results in a
/// "Descending" sort (Greatest values first), but with `None` values appearing first
/// (e.g. `[None, Some(20), Some(10)]`).
#[derive(Debug, Copy, Clone, Default, Serialize, Deserialize)]
pub struct NaturalNoneIsHigherComparator;
impl<T> Comparator<Option<T>> for NaturalNoneIsHigherComparator
where NaturalComparator: Comparator<T>
{
#[inline(always)]
fn compare(&self, lhs_opt: &Option<T>, rhs_opt: &Option<T>) -> Ordering {
match (lhs_opt, rhs_opt) {
(None, None) => Ordering::Equal,
(None, Some(_)) => Ordering::Greater,
(Some(_), None) => Ordering::Less,
(Some(lhs), Some(rhs)) => NaturalComparator.compare(lhs, rhs),
}
}
}
impl Comparator<u32> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &u32, rhs: &u32) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<u64> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &u64, rhs: &u64) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<f64> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &f64, rhs: &f64) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<f32> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &f32, rhs: &f32) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<i64> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &i64, rhs: &i64) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<String> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &String, rhs: &String) -> Ordering {
NaturalComparator.compare(lhs, rhs)
}
}
impl Comparator<OwnedValue> for NaturalNoneIsHigherComparator {
#[inline(always)]
fn compare(&self, lhs: &OwnedValue, rhs: &OwnedValue) -> Ordering {
compare_owned_value::</* NULLS_FIRST= */ false>(lhs, rhs)
}
}
/// An enum representing the different sort orders.
#[derive(Debug, Clone, Copy, Eq, PartialEq, Default)]
pub enum ComparatorEnum {
/// Natural order (See [NaturalComparator])
#[default]
Natural,
/// Reverse order (See [ReverseComparator])
Reverse,
/// Reverse order by treating None as the lowest value. (See [ReverseNoneLowerComparator])
ReverseNoneLower,
/// Natural order but treating None as the highest value. (See [NaturalNoneIsHigherComparator])
NaturalNoneHigher,
}
impl From<Order> for ComparatorEnum {
fn from(order: Order) -> Self {
match order {
Order::Asc => ComparatorEnum::ReverseNoneLower,
Order::Desc => ComparatorEnum::Natural,
}
}
}
impl<T> Comparator<T> for ComparatorEnum
where
ReverseNoneIsLowerComparator: Comparator<T>,
NaturalComparator: Comparator<T>,
ReverseComparator: Comparator<T>,
NaturalNoneIsHigherComparator: Comparator<T>,
{
#[inline(always)]
fn compare(&self, lhs: &T, rhs: &T) -> Ordering {
match self {
ComparatorEnum::Natural => NaturalComparator.compare(lhs, rhs),
ComparatorEnum::Reverse => ReverseComparator.compare(lhs, rhs),
ComparatorEnum::ReverseNoneLower => ReverseNoneIsLowerComparator.compare(lhs, rhs),
ComparatorEnum::NaturalNoneHigher => NaturalNoneIsHigherComparator.compare(lhs, rhs),
}
}
}
impl<Head, Tail, LeftComparator, RightComparator> Comparator<(Head, Tail)>
for (LeftComparator, RightComparator)
where
LeftComparator: Comparator<Head>,
RightComparator: Comparator<Tail>,
{
#[inline(always)]
fn compare(&self, lhs: &(Head, Tail), rhs: &(Head, Tail)) -> Ordering {
self.0
.compare(&lhs.0, &rhs.0)
.then_with(|| self.1.compare(&lhs.1, &rhs.1))
}
}
impl<Type1, Type2, Type3, Comparator1, Comparator2, Comparator3> Comparator<(Type1, (Type2, Type3))>
for (Comparator1, Comparator2, Comparator3)
where
Comparator1: Comparator<Type1>,
Comparator2: Comparator<Type2>,
Comparator3: Comparator<Type3>,
{
#[inline(always)]
fn compare(&self, lhs: &(Type1, (Type2, Type3)), rhs: &(Type1, (Type2, Type3))) -> Ordering {
self.0
.compare(&lhs.0, &rhs.0)
.then_with(|| self.1.compare(&lhs.1 .0, &rhs.1 .0))
.then_with(|| self.2.compare(&lhs.1 .1, &rhs.1 .1))
}
}
impl<Type1, Type2, Type3, Comparator1, Comparator2, Comparator3> Comparator<(Type1, Type2, Type3)>
for (Comparator1, Comparator2, Comparator3)
where
Comparator1: Comparator<Type1>,
Comparator2: Comparator<Type2>,
Comparator3: Comparator<Type3>,
{
#[inline(always)]
fn compare(&self, lhs: &(Type1, Type2, Type3), rhs: &(Type1, Type2, Type3)) -> Ordering {
self.0
.compare(&lhs.0, &rhs.0)
.then_with(|| self.1.compare(&lhs.1, &rhs.1))
.then_with(|| self.2.compare(&lhs.2, &rhs.2))
}
}
impl<Type1, Type2, Type3, Type4, Comparator1, Comparator2, Comparator3, Comparator4>
Comparator<(Type1, (Type2, (Type3, Type4)))>
for (Comparator1, Comparator2, Comparator3, Comparator4)
where
Comparator1: Comparator<Type1>,
Comparator2: Comparator<Type2>,
Comparator3: Comparator<Type3>,
Comparator4: Comparator<Type4>,
{
#[inline(always)]
fn compare(
&self,
lhs: &(Type1, (Type2, (Type3, Type4))),
rhs: &(Type1, (Type2, (Type3, Type4))),
) -> Ordering {
self.0
.compare(&lhs.0, &rhs.0)
.then_with(|| self.1.compare(&lhs.1 .0, &rhs.1 .0))
.then_with(|| self.2.compare(&lhs.1 .1 .0, &rhs.1 .1 .0))
.then_with(|| self.3.compare(&lhs.1 .1 .1, &rhs.1 .1 .1))
}
}
impl<Type1, Type2, Type3, Type4, Comparator1, Comparator2, Comparator3, Comparator4>
Comparator<(Type1, Type2, Type3, Type4)>
for (Comparator1, Comparator2, Comparator3, Comparator4)
where
Comparator1: Comparator<Type1>,
Comparator2: Comparator<Type2>,
Comparator3: Comparator<Type3>,
Comparator4: Comparator<Type4>,
{
#[inline(always)]
fn compare(
&self,
lhs: &(Type1, Type2, Type3, Type4),
rhs: &(Type1, Type2, Type3, Type4),
) -> Ordering {
self.0
.compare(&lhs.0, &rhs.0)
.then_with(|| self.1.compare(&lhs.1, &rhs.1))
.then_with(|| self.2.compare(&lhs.2, &rhs.2))
.then_with(|| self.3.compare(&lhs.3, &rhs.3))
}
}
impl<TSortKeyComputer> SortKeyComputer for (TSortKeyComputer, ComparatorEnum)
where
TSortKeyComputer: SortKeyComputer,
ComparatorEnum: Comparator<TSortKeyComputer::SortKey>,
ComparatorEnum: Comparator<
<<TSortKeyComputer as SortKeyComputer>::Child as SegmentSortKeyComputer>::SegmentSortKey,
>,
{
type SortKey = TSortKeyComputer::SortKey;
type Child = SegmentSortKeyComputerWithComparator<TSortKeyComputer::Child, Self::Comparator>;
type Comparator = ComparatorEnum;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)
}
fn requires_scoring(&self) -> bool {
self.0.requires_scoring()
}
fn comparator(&self) -> Self::Comparator {
self.1
}
fn segment_sort_key_computer(
&self,
segment_reader: &crate::SegmentReader,
) -> crate::Result<Self::Child> {
let child = self.0.segment_sort_key_computer(segment_reader)?;
Ok(SegmentSortKeyComputerWithComparator {
segment_sort_key_computer: child,
comparator: self.comparator(),
})
}
}
impl<TSortKeyComputer> SortKeyComputer for (TSortKeyComputer, Order)
where
TSortKeyComputer: SortKeyComputer,
ComparatorEnum: Comparator<TSortKeyComputer::SortKey>,
ComparatorEnum: Comparator<
<<TSortKeyComputer as SortKeyComputer>::Child as SegmentSortKeyComputer>::SegmentSortKey,
>,
{
type SortKey = TSortKeyComputer::SortKey;
type Child = SegmentSortKeyComputerWithComparator<TSortKeyComputer::Child, Self::Comparator>;
type Comparator = ComparatorEnum;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)
}
fn requires_scoring(&self) -> bool {
self.0.requires_scoring()
}
fn comparator(&self) -> Self::Comparator {
self.1.into()
}
fn segment_sort_key_computer(
&self,
segment_reader: &crate::SegmentReader,
) -> crate::Result<Self::Child> {
let child = self.0.segment_sort_key_computer(segment_reader)?;
Ok(SegmentSortKeyComputerWithComparator {
segment_sort_key_computer: child,
comparator: self.comparator(),
})
}
}
/// A segment sort key computer with a custom ordering.
pub struct SegmentSortKeyComputerWithComparator<TSegmentSortKeyComputer, TComparator> {
segment_sort_key_computer: TSegmentSortKeyComputer,
comparator: TComparator,
}
impl<TSegmentSortKeyComputer, TSegmentSortKey, TComparator> SegmentSortKeyComputer
for SegmentSortKeyComputerWithComparator<TSegmentSortKeyComputer, TComparator>
where
TSegmentSortKeyComputer: SegmentSortKeyComputer<SegmentSortKey = TSegmentSortKey>,
TSegmentSortKey: Clone + 'static + Sync + Send,
TComparator: Comparator<TSegmentSortKey> + 'static + Sync + Send,
{
type SortKey = TSegmentSortKeyComputer::SortKey;
type SegmentSortKey = TSegmentSortKey;
type SegmentComparator = TComparator;
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Self::SegmentSortKey {
self.segment_sort_key_computer.segment_sort_key(doc, score)
}
#[inline(always)]
fn compare_segment_sort_key(
&self,
left: &Self::SegmentSortKey,
right: &Self::SegmentSortKey,
) -> Ordering {
self.comparator.compare(left, right)
}
fn convert_segment_sort_key(&self, sort_key: Self::SegmentSortKey) -> Self::SortKey {
self.segment_sort_key_computer
.convert_segment_sort_key(sort_key)
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::schema::OwnedValue;
#[test]
fn test_natural_none_is_higher() {
let comp = NaturalNoneIsHigherComparator;
let null = None;
let v1 = Some(1_u64);
let v2 = Some(2_u64);
// NaturalNoneIsGreaterComparator logic:
// 1. Delegates to NaturalComparator for non-nulls.
// NaturalComparator compare(2, 1) -> 2.cmp(1) -> Greater.
assert_eq!(comp.compare(&v2, &v1), Ordering::Greater);
// 2. Treats None (Null) as Greater than any value.
// compare(None, Some(2)) should be Greater.
assert_eq!(comp.compare(&null, &v2), Ordering::Greater);
// compare(Some(1), None) should be Less.
assert_eq!(comp.compare(&v1, &null), Ordering::Less);
// compare(None, None) should be Equal.
assert_eq!(comp.compare(&null, &null), Ordering::Equal);
}
#[test]
fn test_mixed_ownedvalue_compare() {
let u = OwnedValue::U64(10);
let i = OwnedValue::I64(10);
let f = OwnedValue::F64(10.0);
let nc = NaturalComparator;
assert_eq!(nc.compare(&u, &i), Ordering::Equal);
assert_eq!(nc.compare(&u, &f), Ordering::Equal);
assert_eq!(nc.compare(&i, &f), Ordering::Equal);
let u2 = OwnedValue::U64(11);
assert_eq!(nc.compare(&u2, &f), Ordering::Greater);
let s = OwnedValue::Str("a".to_string());
// Str < U64
assert_eq!(nc.compare(&s, &u), Ordering::Less);
// Str < I64
assert_eq!(nc.compare(&s, &i), Ordering::Less);
// Str < F64
assert_eq!(nc.compare(&s, &f), Ordering::Less);
}
}

361
src/collector/sort_key/sort_by_erased_type.rs
View File

@@ -1,361 +0,0 @@
use columnar::{ColumnType, MonotonicallyMappableToU64};
use crate::collector::sort_key::{
NaturalComparator, SortBySimilarityScore, SortByStaticFastValue, SortByString,
};
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer};
use crate::fastfield::FastFieldNotAvailableError;
use crate::schema::OwnedValue;
use crate::{DateTime, DocId, Score};
/// Sort by the boxed / OwnedValue representation of either a fast field, or of the score.
///
/// Using the OwnedValue representation allows for type erasure, and can be useful when sort orders
/// are not known until runtime. But it comes with a performance cost: wherever possible, prefer to
/// use a SortKeyComputer implementation with a known-type at compile time.
#[derive(Debug, Clone)]
pub enum SortByErasedType {
/// Sort by a fast field
Field(String),
/// Sort by score
Score,
}
impl SortByErasedType {
/// Creates a new sort key computer which will sort by the given fast field column, with type
/// erasure.
pub fn for_field(column_name: impl ToString) -> Self {
Self::Field(column_name.to_string())
}
/// Creates a new sort key computer which will sort by score, with type erasure.
pub fn for_score() -> Self {
Self::Score
}
}
trait ErasedSegmentSortKeyComputer: Send + Sync {
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Option<u64>;
fn convert_segment_sort_key(&self, sort_key: Option<u64>) -> OwnedValue;
}
struct ErasedSegmentSortKeyComputerWrapper<C, F> {
inner: C,
converter: F,
}
impl<C, F> ErasedSegmentSortKeyComputer for ErasedSegmentSortKeyComputerWrapper<C, F>
where
C: SegmentSortKeyComputer<SegmentSortKey = Option<u64>> + Send + Sync,
F: Fn(C::SortKey) -> OwnedValue + Send + Sync + 'static,
{
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Option<u64> {
self.inner.segment_sort_key(doc, score)
}
fn convert_segment_sort_key(&self, sort_key: Option<u64>) -> OwnedValue {
let val = self.inner.convert_segment_sort_key(sort_key);
(self.converter)(val)
}
}
struct ScoreSegmentSortKeyComputer {
segment_computer: SortBySimilarityScore,
}
impl ErasedSegmentSortKeyComputer for ScoreSegmentSortKeyComputer {
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Option<u64> {
let score_value: f64 = self.segment_computer.segment_sort_key(doc, score).into();
Some(score_value.to_u64())
}
fn convert_segment_sort_key(&self, sort_key: Option<u64>) -> OwnedValue {
let score_value: u64 = sort_key.expect("This implementation always produces a score.");
OwnedValue::F64(f64::from_u64(score_value))
}
}
impl SortKeyComputer for SortByErasedType {
type SortKey = OwnedValue;
type Child = ErasedColumnSegmentSortKeyComputer;
type Comparator = NaturalComparator;
fn requires_scoring(&self) -> bool {
matches!(self, Self::Score)
}
fn segment_sort_key_computer(
&self,
segment_reader: &crate::SegmentReader,
) -> crate::Result<Self::Child> {
let inner: Box<dyn ErasedSegmentSortKeyComputer> = match self {
Self::Field(column_name) => {
let fast_fields = segment_reader.fast_fields();
// TODO: We currently double-open the column to avoid relying on the implementation
// details of `SortByString` or `SortByStaticFastValue`. Once
// https://github.com/quickwit-oss/tantivy/issues/2776 is resolved, we should
// consider directly constructing the appropriate `SegmentSortKeyComputer` type for
// the column that we open here.
let (_column, column_type) =
fast_fields.u64_lenient(column_name)?.ok_or_else(|| {
FastFieldNotAvailableError {
field_name: column_name.to_owned(),
}
})?;
match column_type {
ColumnType::Str => {
let computer = SortByString::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<String>| {
val.map(OwnedValue::Str).unwrap_or(OwnedValue::Null)
},
})
}
ColumnType::U64 => {
let computer = SortByStaticFastValue::<u64>::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<u64>| {
val.map(OwnedValue::U64).unwrap_or(OwnedValue::Null)
},
})
}
ColumnType::I64 => {
let computer = SortByStaticFastValue::<i64>::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<i64>| {
val.map(OwnedValue::I64).unwrap_or(OwnedValue::Null)
},
})
}
ColumnType::F64 => {
let computer = SortByStaticFastValue::<f64>::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<f64>| {
val.map(OwnedValue::F64).unwrap_or(OwnedValue::Null)
},
})
}
ColumnType::Bool => {
let computer = SortByStaticFastValue::<bool>::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<bool>| {
val.map(OwnedValue::Bool).unwrap_or(OwnedValue::Null)
},
})
}
ColumnType::DateTime => {
let computer = SortByStaticFastValue::<DateTime>::for_field(column_name);
let inner = computer.segment_sort_key_computer(segment_reader)?;
Box::new(ErasedSegmentSortKeyComputerWrapper {
inner,
converter: |val: Option<DateTime>| {
val.map(OwnedValue::Date).unwrap_or(OwnedValue::Null)
},
})
}
column_type => {
return Err(crate::TantivyError::SchemaError(format!(
"Field `{}` is of type {column_type:?}, which is not supported for \
sorting by owned value yet.",
column_name
)))
}
}
}
Self::Score => Box::new(ScoreSegmentSortKeyComputer {
segment_computer: SortBySimilarityScore,
}),
};
Ok(ErasedColumnSegmentSortKeyComputer { inner })
}
}
pub struct ErasedColumnSegmentSortKeyComputer {
inner: Box<dyn ErasedSegmentSortKeyComputer>,
}
impl SegmentSortKeyComputer for ErasedColumnSegmentSortKeyComputer {
type SortKey = OwnedValue;
type SegmentSortKey = Option<u64>;
type SegmentComparator = NaturalComparator;
#[inline(always)]
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Option<u64> {
self.inner.segment_sort_key(doc, score)
}
fn convert_segment_sort_key(&self, segment_sort_key: Self::SegmentSortKey) -> OwnedValue {
self.inner.convert_segment_sort_key(segment_sort_key)
}
}
#[cfg(test)]
mod tests {
use crate::collector::sort_key::{ComparatorEnum, SortByErasedType};
use crate::collector::TopDocs;
use crate::query::AllQuery;
use crate::schema::{OwnedValue, Schema, FAST, TEXT};
use crate::Index;
#[test]
fn test_sort_by_owned_u64() {
let mut schema_builder = Schema::builder();
let id_field = schema_builder.add_u64_field("id", FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut writer = index.writer_for_tests().unwrap();
writer.add_document(doc!(id_field => 10u64)).unwrap();
writer.add_document(doc!(id_field => 2u64)).unwrap();
writer.add_document(doc!()).unwrap();
writer.commit().unwrap();
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let collector = TopDocs::with_limit(10)
.order_by((SortByErasedType::for_field("id"), ComparatorEnum::Natural));
let top_docs = searcher.search(&AllQuery, &collector).unwrap();
let values: Vec<OwnedValue> = top_docs.into_iter().map(|(key, _)| key).collect();
assert_eq!(
values,
vec![OwnedValue::U64(10), OwnedValue::U64(2), OwnedValue::Null]
);
let collector = TopDocs::with_limit(10).order_by((
SortByErasedType::for_field("id"),
ComparatorEnum::ReverseNoneLower,
));
let top_docs = searcher.search(&AllQuery, &collector).unwrap();
let values: Vec<OwnedValue> = top_docs.into_iter().map(|(key, _)| key).collect();
assert_eq!(
values,
vec![OwnedValue::U64(2), OwnedValue::U64(10), OwnedValue::Null]
);
}
#[test]
fn test_sort_by_owned_string() {
let mut schema_builder = Schema::builder();
let city_field = schema_builder.add_text_field("city", FAST | TEXT);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut writer = index.writer_for_tests().unwrap();
writer.add_document(doc!(city_field => "tokyo")).unwrap();
writer.add_document(doc!(city_field => "austin")).unwrap();
writer.add_document(doc!()).unwrap();
writer.commit().unwrap();
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let collector = TopDocs::with_limit(10).order_by((
SortByErasedType::for_field("city"),
ComparatorEnum::ReverseNoneLower,
));
let top_docs = searcher.search(&AllQuery, &collector).unwrap();
let values: Vec<OwnedValue> = top_docs.into_iter().map(|(key, _)| key).collect();
assert_eq!(
values,
vec![
OwnedValue::Str("austin".to_string()),
OwnedValue::Str("tokyo".to_string()),
OwnedValue::Null
]
);
}
#[test]
fn test_sort_by_owned_reverse() {
let mut schema_builder = Schema::builder();
let id_field = schema_builder.add_u64_field("id", FAST);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut writer = index.writer_for_tests().unwrap();
writer.add_document(doc!(id_field => 10u64)).unwrap();
writer.add_document(doc!(id_field => 2u64)).unwrap();
writer.add_document(doc!()).unwrap();
writer.commit().unwrap();
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let collector = TopDocs::with_limit(10)
.order_by((SortByErasedType::for_field("id"), ComparatorEnum::Reverse));
let top_docs = searcher.search(&AllQuery, &collector).unwrap();
let values: Vec<OwnedValue> = top_docs.into_iter().map(|(key, _)| key).collect();
assert_eq!(
values,
vec![OwnedValue::Null, OwnedValue::U64(2), OwnedValue::U64(10)]
);
}
#[test]
fn test_sort_by_owned_score() {
let mut schema_builder = Schema::builder();
let body_field = schema_builder.add_text_field("body", TEXT);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema);
let mut writer = index.writer_for_tests().unwrap();
writer.add_document(doc!(body_field => "a a")).unwrap();
writer.add_document(doc!(body_field => "a")).unwrap();
writer.commit().unwrap();
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let query_parser = crate::query::QueryParser::for_index(&index, vec![body_field]);
let query = query_parser.parse_query("a").unwrap();
// Sort by score descending (Natural)
let collector = TopDocs::with_limit(10)
.order_by((SortByErasedType::for_score(), ComparatorEnum::Natural));
let top_docs = searcher.search(&query, &collector).unwrap();
let values: Vec<f64> = top_docs
.into_iter()
.map(|(key, _)| match key {
OwnedValue::F64(val) => val,
_ => panic!("Wrong type {key:?}"),
})
.collect();
assert_eq!(values.len(), 2);
assert!(values[0] > values[1]);
// Sort by score ascending (ReverseNoneLower)
let collector = TopDocs::with_limit(10).order_by((
SortByErasedType::for_score(),
ComparatorEnum::ReverseNoneLower,
));
let top_docs = searcher.search(&query, &collector).unwrap();
let values: Vec<f64> = top_docs
.into_iter()
.map(|(key, _)| match key {
OwnedValue::F64(val) => val,
_ => panic!("Wrong type {key:?}"),
})
.collect();
assert_eq!(values.len(), 2);
assert!(values[0] < values[1]);
}
}

77
src/collector/sort_key/sort_by_score.rs
View File

@@ -1,77 +0,0 @@
use crate::collector::sort_key::NaturalComparator;
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer, TopNComputer};
use crate::{DocAddress, DocId, Score};
/// Sort by similarity score.
#[derive(Clone, Debug, Copy)]
pub struct SortBySimilarityScore;
impl SortKeyComputer for SortBySimilarityScore {
type SortKey = Score;
type Child = SortBySimilarityScore;
type Comparator = NaturalComparator;
fn requires_scoring(&self) -> bool {
true
}
fn segment_sort_key_computer(
&self,
_segment_reader: &crate::SegmentReader,
) -> crate::Result<Self::Child> {
Ok(SortBySimilarityScore)
}
// Sorting by score is special in that it allows for the Block-Wand optimization.
fn collect_segment_top_k(
&self,
k: usize,
weight: &dyn crate::query::Weight,
reader: &crate::SegmentReader,
segment_ord: u32,
) -> crate::Result<Vec<(Self::SortKey, DocAddress)>> {
let mut top_n: TopNComputer<Score, DocId, Self::Comparator> =
TopNComputer::new_with_comparator(k, self.comparator());
if let Some(alive_bitset) = reader.alive_bitset() {
let mut threshold = Score::MIN;
top_n.threshold = Some(threshold);
weight.for_each_pruning(Score::MIN, reader, &mut |doc, score| {
if alive_bitset.is_deleted(doc) {
return threshold;
}
top_n.push(score, doc);
threshold = top_n.threshold.unwrap_or(Score::MIN);
threshold
})?;
} else {
weight.for_each_pruning(Score::MIN, reader, &mut |doc, score| {
top_n.push(score, doc);
top_n.threshold.unwrap_or(Score::MIN)
})?;
}
Ok(top_n
.into_vec()
.into_iter()
.map(|cid| (cid.sort_key, DocAddress::new(segment_ord, cid.doc)))
.collect())
}
}
impl SegmentSortKeyComputer for SortBySimilarityScore {
type SortKey = Score;
type SegmentSortKey = Score;
type SegmentComparator = NaturalComparator;
#[inline(always)]
fn segment_sort_key(&mut self, _doc: DocId, score: Score) -> Score {
score
}
fn convert_segment_sort_key(&self, score: Score) -> Score {
score
}
}

96
src/collector/sort_key/sort_by_static_fast_value.rs
View File

@@ -1,96 +0,0 @@
use std::marker::PhantomData;
use columnar::Column;
use crate::collector::sort_key::NaturalComparator;
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer};
use crate::fastfield::{FastFieldNotAvailableError, FastValue};
use crate::{DocId, Score, SegmentReader};
/// Sorts by a fast value (u64, i64, f64, bool).
///
/// The field must appear explicitly in the schema, with the right type, and declared as
/// a fast field..
///
/// If the field is multivalued, only the first value is considered.
///
/// Documents that do not have this value are still considered.
/// Their sort key will simply be `None`.
#[derive(Debug, Clone)]
pub struct SortByStaticFastValue<T: FastValue> {
field: String,
typ: PhantomData<T>,
}
impl<T: FastValue> SortByStaticFastValue<T> {
/// Creates a new `SortByStaticFastValue` instance for the given field.
pub fn for_field(column_name: impl ToString) -> SortByStaticFastValue<T> {
Self {
field: column_name.to_string(),
typ: PhantomData,
}
}
}
impl<T: FastValue> SortKeyComputer for SortByStaticFastValue<T> {
type Child = SortByFastValueSegmentSortKeyComputer<T>;
type SortKey = Option<T>;
type Comparator = NaturalComparator;
fn check_schema(&self, schema: &crate::schema::Schema) -> crate::Result<()> {
// At the segment sort key computer level, we rely on the u64 representation.
// The mapping is monotonic, so it is sufficient to compute our top-K docs.
let field = schema.get_field(&self.field)?;
let field_entry = schema.get_field_entry(field);
if !field_entry.is_fast() {
return Err(crate::TantivyError::SchemaError(format!(
"Field `{}` is not a fast field.",
self.field,
)));
}
let schema_type = field_entry.field_type().value_type();
if schema_type != T::to_type() {
return Err(crate::TantivyError::SchemaError(format!(
"Field `{}` is of type {schema_type:?}, not of the type {:?}.",
&self.field,
T::to_type()
)));
}
Ok(())
}
fn segment_sort_key_computer(
&self,
segment_reader: &SegmentReader,
) -> crate::Result<Self::Child> {
let sort_column_opt = segment_reader.fast_fields().u64_lenient(&self.field)?;
let (sort_column, _sort_column_type) =
sort_column_opt.ok_or_else(|| FastFieldNotAvailableError {
field_name: self.field.clone(),
})?;
Ok(SortByFastValueSegmentSortKeyComputer {
sort_column,
typ: PhantomData,
})
}
}
pub struct SortByFastValueSegmentSortKeyComputer<T> {
sort_column: Column<u64>,
typ: PhantomData<T>,
}
impl<T: FastValue> SegmentSortKeyComputer for SortByFastValueSegmentSortKeyComputer<T> {
type SortKey = Option<T>;
type SegmentSortKey = Option<u64>;
type SegmentComparator = NaturalComparator;
#[inline(always)]
fn segment_sort_key(&mut self, doc: DocId, _score: Score) -> Self::SegmentSortKey {
self.sort_column.first(doc)
}
fn convert_segment_sort_key(&self, sort_key: Self::SegmentSortKey) -> Self::SortKey {
sort_key.map(T::from_u64)
}
}

72
src/collector/sort_key/sort_by_string.rs
View File

@@ -1,72 +0,0 @@
use columnar::StrColumn;
use crate::collector::sort_key::NaturalComparator;
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer};
use crate::termdict::TermOrdinal;
use crate::{DocId, Score};
/// Sort by the first value of a string column.
///
/// The string can be dynamic (coming from a json field)
/// or static (being specificaly defined in the configuration).
///
/// If the field is multivalued, only the first value is considered.
///
/// Documents that do not have this value are still considered.
/// Their sort key will simply be `None`.
#[derive(Debug, Clone)]
pub struct SortByString {
column_name: String,
}
impl SortByString {
/// Creates a new sort by string sort key computer.
pub fn for_field(column_name: impl ToString) -> Self {
SortByString {
column_name: column_name.to_string(),
}
}
}
impl SortKeyComputer for SortByString {
type SortKey = Option<String>;
type Child = ByStringColumnSegmentSortKeyComputer;
type Comparator = NaturalComparator;
fn segment_sort_key_computer(
&self,
segment_reader: &crate::SegmentReader,
) -> crate::Result<Self::Child> {
let str_column_opt = segment_reader.fast_fields().str(&self.column_name)?;
Ok(ByStringColumnSegmentSortKeyComputer { str_column_opt })
}
}
pub struct ByStringColumnSegmentSortKeyComputer {
str_column_opt: Option<StrColumn>,
}
impl SegmentSortKeyComputer for ByStringColumnSegmentSortKeyComputer {
type SortKey = Option<String>;
type SegmentSortKey = Option<TermOrdinal>;
type SegmentComparator = NaturalComparator;
#[inline(always)]
fn segment_sort_key(&mut self, doc: DocId, _score: Score) -> Option<TermOrdinal> {
let str_column = self.str_column_opt.as_ref()?;
str_column.ords().first(doc)
}
fn convert_segment_sort_key(&self, term_ord_opt: Option<TermOrdinal>) -> Option<String> {
// TODO: Individual lookups to the dictionary like this are very likely to repeatedly
// decompress the same blocks. See https://github.com/quickwit-oss/tantivy/issues/2776
let term_ord = term_ord_opt?;
let str_column = self.str_column_opt.as_ref()?;
let mut bytes = Vec::new();
str_column
.dictionary()
.ord_to_term(term_ord, &mut bytes)
.ok()?;
String::try_from(bytes).ok()
}
}

643
src/collector/sort_key/sort_key_computer.rs
View File

@@ -1,643 +0,0 @@
use std::cmp::Ordering;
use crate::collector::sort_key::{Comparator, NaturalComparator};
use crate::collector::sort_key_top_collector::TopBySortKeySegmentCollector;
use crate::collector::{default_collect_segment_impl, SegmentCollector as _, TopNComputer};
use crate::schema::Schema;
use crate::{DocAddress, DocId, Result, Score, SegmentReader};
/// A `SegmentSortKeyComputer` makes it possible to modify the default score
/// for a given document belonging to a specific segment.
///
/// It is the segment local version of the [`SortKeyComputer`].
pub trait SegmentSortKeyComputer: 'static {
/// The final score being emitted.
type SortKey: 'static + Send + Sync + Clone;
/// Sort key used by at the segment level by the `SegmentSortKeyComputer`.
///
/// It is typically small like a `u64`, and is meant to be converted
/// to the final score at the end of the collection of the segment.
type SegmentSortKey: 'static + Clone + Send + Sync + Clone;
/// Comparator type.
type SegmentComparator: Comparator<Self::SegmentSortKey> + 'static;
/// Returns the segment sort key comparator.
fn segment_comparator(&self) -> Self::SegmentComparator {
Self::SegmentComparator::default()
}
/// Computes the sort key for the given document and score.
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Self::SegmentSortKey;
/// Computes the sort key and pushes the document in a TopN Computer.
///
/// When using a tuple as the sorting key, the sort key is evaluated in a lazy manner.
#[inline(always)]
fn compute_sort_key_and_collect<C: Comparator<Self::SegmentSortKey>>(
&mut self,
doc: DocId,
score: Score,
top_n_computer: &mut TopNComputer<Self::SegmentSortKey, DocId, C>,
) {
let sort_key = self.segment_sort_key(doc, score);
top_n_computer.push(sort_key, doc);
}
/// A SegmentSortKeyComputer maps to a SegmentSortKey, but it can also decide on
/// its ordering.
///
/// This method must be consistent with the `SortKey` ordering.
#[inline(always)]
fn compare_segment_sort_key(
&self,
left: &Self::SegmentSortKey,
right: &Self::SegmentSortKey,
) -> Ordering {
self.segment_comparator().compare(left, right)
}
/// Implementing this method makes it possible to avoid computing
/// a sort_key entirely if we can assess that it won't pass a threshold
/// with a partial computation.
///
/// This is currently used for lexicographic sorting.
fn accept_sort_key_lazy(
&mut self,
doc_id: DocId,
score: Score,
threshold: &Self::SegmentSortKey,
) -> Option<(Ordering, Self::SegmentSortKey)> {
let sort_key = self.segment_sort_key(doc_id, score);
let cmp = self.compare_segment_sort_key(&sort_key, threshold);
if cmp == Ordering::Less {
None
} else {
Some((cmp, sort_key))
}
}
/// Convert a segment level sort key into the global sort key.
fn convert_segment_sort_key(&self, sort_key: Self::SegmentSortKey) -> Self::SortKey;
}
/// `SortKeyComputer` defines the sort key to be used by a TopK Collector.
///
/// The `SortKeyComputer` itself does not make much of the computation itself.
/// Instead, it helps constructing `Self::Child` instances that will compute
/// the sort key at a segment scale.
pub trait SortKeyComputer: Sync {
/// The sort key type.
type SortKey: 'static + Send + Sync + Clone + std::fmt::Debug;
/// Type of the associated [`SegmentSortKeyComputer`].
type Child: SegmentSortKeyComputer<SortKey = Self::SortKey>;
/// Comparator type.
type Comparator: Comparator<Self::SortKey>
+ Comparator<<Self::Child as SegmentSortKeyComputer>::SegmentSortKey>
+ 'static;
/// Checks whether the schema is compatible with the sort key computer.
fn check_schema(&self, _schema: &Schema) -> crate::Result<()> {
Ok(())
}
/// Returns the sort key comparator.
fn comparator(&self) -> Self::Comparator {
Self::Comparator::default()
}
/// Indicates whether the sort key actually uses the similarity score (by default BM25).
/// If set to false, the similary score might not be computed (as an optimization),
/// and the score fed in the segment sort key computer could take any value.
fn requires_scoring(&self) -> bool {
false
}
/// Sorting by score has a overriding implementation for BM25 scores, using Block-WAND.
fn collect_segment_top_k(
&self,
k: usize,
weight: &dyn crate::query::Weight,
reader: &crate::SegmentReader,
segment_ord: u32,
) -> crate::Result<Vec<(Self::SortKey, DocAddress)>> {
let with_scoring = self.requires_scoring();
let segment_sort_key_computer = self.segment_sort_key_computer(reader)?;
let topn_computer = TopNComputer::new_with_comparator(k, self.comparator());
let mut segment_top_key_collector = TopBySortKeySegmentCollector {
topn_computer,
segment_ord,
segment_sort_key_computer,
};
default_collect_segment_impl(&mut segment_top_key_collector, weight, reader, with_scoring)?;
Ok(segment_top_key_collector.harvest())
}
/// Builds a child sort key computer for a specific segment.
fn segment_sort_key_computer(&self, segment_reader: &SegmentReader) -> Result<Self::Child>;
}
impl<HeadSortKeyComputer, TailSortKeyComputer> SortKeyComputer
for (HeadSortKeyComputer, TailSortKeyComputer)
where
HeadSortKeyComputer: SortKeyComputer,
TailSortKeyComputer: SortKeyComputer,
{
type SortKey = (HeadSortKeyComputer::SortKey, TailSortKeyComputer::SortKey);
type Child = (HeadSortKeyComputer::Child, TailSortKeyComputer::Child);
type Comparator = (
HeadSortKeyComputer::Comparator,
TailSortKeyComputer::Comparator,
);
fn comparator(&self) -> Self::Comparator {
(self.0.comparator(), self.1.comparator())
}
fn segment_sort_key_computer(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
Ok((
self.0.segment_sort_key_computer(segment_reader)?,
self.1.segment_sort_key_computer(segment_reader)?,
))
}
/// Checks whether the schema is compatible with the sort key computer.
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
Ok(())
}
/// Indicates whether the sort key actually uses the similarity score (by default BM25).
/// If set to false, the similary score might not be computed (as an optimization),
/// and the score fed in the segment sort key computer could take any value.
fn requires_scoring(&self) -> bool {
self.0.requires_scoring() || self.1.requires_scoring()
}
}
impl<HeadSegmentSortKeyComputer, TailSegmentSortKeyComputer> SegmentSortKeyComputer
for (HeadSegmentSortKeyComputer, TailSegmentSortKeyComputer)
where
HeadSegmentSortKeyComputer: SegmentSortKeyComputer,
TailSegmentSortKeyComputer: SegmentSortKeyComputer,
{
type SortKey = (
HeadSegmentSortKeyComputer::SortKey,
TailSegmentSortKeyComputer::SortKey,
);
type SegmentSortKey = (
HeadSegmentSortKeyComputer::SegmentSortKey,
TailSegmentSortKeyComputer::SegmentSortKey,
);
type SegmentComparator = (
HeadSegmentSortKeyComputer::SegmentComparator,
TailSegmentSortKeyComputer::SegmentComparator,
);
/// A SegmentSortKeyComputer maps to a SegmentSortKey, but it can also decide on
/// its ordering.
///
/// By default, it uses the natural ordering.
#[inline]
fn compare_segment_sort_key(
&self,
left: &Self::SegmentSortKey,
right: &Self::SegmentSortKey,
) -> Ordering {
self.0
.compare_segment_sort_key(&left.0, &right.0)
.then_with(|| self.1.compare_segment_sort_key(&left.1, &right.1))
}
#[inline(always)]
fn compute_sort_key_and_collect<C: Comparator<Self::SegmentSortKey>>(
&mut self,
doc: DocId,
score: Score,
top_n_computer: &mut TopNComputer<Self::SegmentSortKey, DocId, C>,
) {
let sort_key: Self::SegmentSortKey;
if let Some(threshold) = &top_n_computer.threshold {
if let Some((_cmp, lazy_sort_key)) = self.accept_sort_key_lazy(doc, score, threshold) {
sort_key = lazy_sort_key;
} else {
return;
}
} else {
sort_key = self.segment_sort_key(doc, score);
};
top_n_computer.append_doc(doc, sort_key);
}
#[inline(always)]
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Self::SegmentSortKey {
let head_sort_key = self.0.segment_sort_key(doc, score);
let tail_sort_key = self.1.segment_sort_key(doc, score);
(head_sort_key, tail_sort_key)
}
fn accept_sort_key_lazy(
&mut self,
doc_id: DocId,
score: Score,
threshold: &Self::SegmentSortKey,
) -> Option<(Ordering, Self::SegmentSortKey)> {
let (head_threshold, tail_threshold) = threshold;
let (head_cmp, head_sort_key) =
self.0.accept_sort_key_lazy(doc_id, score, head_threshold)?;
if head_cmp == Ordering::Equal {
let (tail_cmp, tail_sort_key) =
self.1.accept_sort_key_lazy(doc_id, score, tail_threshold)?;
Some((tail_cmp, (head_sort_key, tail_sort_key)))
} else {
let tail_sort_key = self.1.segment_sort_key(doc_id, score);
Some((head_cmp, (head_sort_key, tail_sort_key)))
}
}
fn convert_segment_sort_key(&self, sort_key: Self::SegmentSortKey) -> Self::SortKey {
let (head_sort_key, tail_sort_key) = sort_key;
(
self.0.convert_segment_sort_key(head_sort_key),
self.1.convert_segment_sort_key(tail_sort_key),
)
}
}
/// This struct is used as an adapter to take a sort key computer and map its score to another
/// new sort key.
pub struct MappedSegmentSortKeyComputer<T, PreviousSortKey, NewSortKey> {
sort_key_computer: T,
map: fn(PreviousSortKey) -> NewSortKey,
}
impl<T, PreviousScore, NewScore> SegmentSortKeyComputer
for MappedSegmentSortKeyComputer<T, PreviousScore, NewScore>
where
T: SegmentSortKeyComputer<SortKey = PreviousScore>,
PreviousScore: 'static + Clone + Send + Sync,
NewScore: 'static + Clone + Send + Sync,
{
type SortKey = NewScore;
type SegmentSortKey = T::SegmentSortKey;
type SegmentComparator = T::SegmentComparator;
fn segment_sort_key(&mut self, doc: DocId, score: Score) -> Self::SegmentSortKey {
self.sort_key_computer.segment_sort_key(doc, score)
}
fn accept_sort_key_lazy(
&mut self,
doc_id: DocId,
score: Score,
threshold: &Self::SegmentSortKey,
) -> Option<(Ordering, Self::SegmentSortKey)> {
self.sort_key_computer
.accept_sort_key_lazy(doc_id, score, threshold)
}
#[inline(always)]
fn compute_sort_key_and_collect<C: Comparator<Self::SegmentSortKey>>(
&mut self,
doc: DocId,
score: Score,
top_n_computer: &mut TopNComputer<Self::SegmentSortKey, DocId, C>,
) {
self.sort_key_computer
.compute_sort_key_and_collect(doc, score, top_n_computer);
}
fn convert_segment_sort_key(&self, segment_sort_key: Self::SegmentSortKey) -> Self::SortKey {
(self.map)(
self.sort_key_computer
.convert_segment_sort_key(segment_sort_key),
)
}
}
// We then re-use our (head, tail) implement and our mapper by seeing mapping any tuple (a, b, c,
// ...) as the chain (a, (b, (c, ...)))
impl<SortKeyComputer1, SortKeyComputer2, SortKeyComputer3> SortKeyComputer
for (SortKeyComputer1, SortKeyComputer2, SortKeyComputer3)
where
SortKeyComputer1: SortKeyComputer,
SortKeyComputer2: SortKeyComputer,
SortKeyComputer3: SortKeyComputer,
{
type SortKey = (
SortKeyComputer1::SortKey,
SortKeyComputer2::SortKey,
SortKeyComputer3::SortKey,
);
type Child = MappedSegmentSortKeyComputer<
<(SortKeyComputer1, (SortKeyComputer2, SortKeyComputer3)) as SortKeyComputer>::Child,
(
SortKeyComputer1::SortKey,
(SortKeyComputer2::SortKey, SortKeyComputer3::SortKey),
),
Self::SortKey,
>;
type Comparator = (
SortKeyComputer1::Comparator,
SortKeyComputer2::Comparator,
SortKeyComputer3::Comparator,
);
fn comparator(&self) -> Self::Comparator {
(
self.0.comparator(),
self.1.comparator(),
self.2.comparator(),
)
}
fn segment_sort_key_computer(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
let sort_key_computer1 = self.0.segment_sort_key_computer(segment_reader)?;
let sort_key_computer2 = self.1.segment_sort_key_computer(segment_reader)?;
let sort_key_computer3 = self.2.segment_sort_key_computer(segment_reader)?;
let map = |(sort_key1, (sort_key2, sort_key3))| (sort_key1, sort_key2, sort_key3);
Ok(MappedSegmentSortKeyComputer {
sort_key_computer: (sort_key_computer1, (sort_key_computer2, sort_key_computer3)),
map,
})
}
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
self.2.check_schema(schema)?;
Ok(())
}
fn requires_scoring(&self) -> bool {
self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring()
}
}
impl<SortKeyComputer1, SortKeyComputer2, SortKeyComputer3, SortKeyComputer4> SortKeyComputer
for (
SortKeyComputer1,
SortKeyComputer2,
SortKeyComputer3,
SortKeyComputer4,
)
where
SortKeyComputer1: SortKeyComputer,
SortKeyComputer2: SortKeyComputer,
SortKeyComputer3: SortKeyComputer,
SortKeyComputer4: SortKeyComputer,
{
type Child = MappedSegmentSortKeyComputer<
<(
SortKeyComputer1,
(SortKeyComputer2, (SortKeyComputer3, SortKeyComputer4)),
) as SortKeyComputer>::Child,
(
SortKeyComputer1::SortKey,
(
SortKeyComputer2::SortKey,
(SortKeyComputer3::SortKey, SortKeyComputer4::SortKey),
),
),
Self::SortKey,
>;
type SortKey = (
SortKeyComputer1::SortKey,
SortKeyComputer2::SortKey,
SortKeyComputer3::SortKey,
SortKeyComputer4::SortKey,
);
type Comparator = (
SortKeyComputer1::Comparator,
SortKeyComputer2::Comparator,
SortKeyComputer3::Comparator,
SortKeyComputer4::Comparator,
);
fn segment_sort_key_computer(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
let sort_key_computer1 = self.0.segment_sort_key_computer(segment_reader)?;
let sort_key_computer2 = self.1.segment_sort_key_computer(segment_reader)?;
let sort_key_computer3 = self.2.segment_sort_key_computer(segment_reader)?;
let sort_key_computer4 = self.3.segment_sort_key_computer(segment_reader)?;
Ok(MappedSegmentSortKeyComputer {
sort_key_computer: (
sort_key_computer1,
(sort_key_computer2, (sort_key_computer3, sort_key_computer4)),
),
map: |(sort_key1, (sort_key2, (sort_key3, sort_key4)))| {
(sort_key1, sort_key2, sort_key3, sort_key4)
},
})
}
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.0.check_schema(schema)?;
self.1.check_schema(schema)?;
self.2.check_schema(schema)?;
self.3.check_schema(schema)?;
Ok(())
}
fn requires_scoring(&self) -> bool {
self.0.requires_scoring()
|| self.1.requires_scoring()
|| self.2.requires_scoring()
|| self.3.requires_scoring()
}
}
impl<F, SegmentF, TSortKey> SortKeyComputer for F
where
F: 'static + Send + Sync + Fn(&SegmentReader) -> SegmentF,
SegmentF: 'static + FnMut(DocId) -> TSortKey,
TSortKey: 'static + PartialOrd + Clone + Send + Sync + std::fmt::Debug,
{
type SortKey = TSortKey;
type Child = SegmentF;
type Comparator = NaturalComparator;
fn segment_sort_key_computer(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
Ok((self)(segment_reader))
}
}
impl<F, TSortKey> SegmentSortKeyComputer for F
where
F: 'static + FnMut(DocId) -> TSortKey,
TSortKey: 'static + PartialOrd + Clone + Send + Sync,
{
type SortKey = TSortKey;
type SegmentSortKey = TSortKey;
type SegmentComparator = NaturalComparator;
fn segment_sort_key(&mut self, doc: DocId, _score: Score) -> TSortKey {
(self)(doc)
}
/// Convert a segment level score into the global level score.
fn convert_segment_sort_key(&self, sort_key: Self::SegmentSortKey) -> Self::SortKey {
sort_key
}
}
#[cfg(test)]
mod tests {
use std::cmp::Ordering;
use std::sync::atomic::{AtomicUsize, Ordering as AtomicOrdering};
use std::sync::Arc;
use crate::collector::{SegmentSortKeyComputer, SortKeyComputer};
use crate::schema::Schema;
use crate::{DocId, Index, Order, SegmentReader};
fn build_test_index() -> Index {
let schema = Schema::builder().build();
let index = Index::create_in_ram(schema);
let mut index_writer = index.writer_for_tests().unwrap();
index_writer
.add_document(crate::TantivyDocument::default())
.unwrap();
index_writer.commit().unwrap();
index
}
#[test]
fn test_lazy_score_computer() {
let score_computer_primary = |_segment_reader: &SegmentReader| |_doc: DocId| 200u32;
let call_count = Arc::new(AtomicUsize::new(0));
let call_count_clone = call_count.clone();
let score_computer_secondary = move |_segment_reader: &SegmentReader| {
let call_count_new_clone = call_count_clone.clone();
move |_doc: DocId| {
call_count_new_clone.fetch_add(1, AtomicOrdering::SeqCst);
"b"
}
};
let lazy_score_computer = (score_computer_primary, score_computer_secondary);
let index = build_test_index();
let searcher = index.reader().unwrap().searcher();
let mut segment_sort_key_computer = lazy_score_computer
.segment_sort_key_computer(searcher.segment_reader(0))
.unwrap();
let expected_sort_key = (200, "b");
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(100u32, "a"));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 1);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(100u32, "c"));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 2);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(200u32, "a"));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 3);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(200u32, "c"));
assert!(sort_key_opt.is_none());
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(300u32, "a"));
assert_eq!(sort_key_opt, None);
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(300u32, "c"));
assert_eq!(sort_key_opt, None);
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &expected_sort_key);
assert_eq!(sort_key_opt, Some((Ordering::Equal, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 5);
}
}
#[test]
fn test_lazy_score_computer_dynamic_ordering() {
let score_computer_primary = |_segment_reader: &SegmentReader| |_doc: DocId| 200u32;
let call_count = Arc::new(AtomicUsize::new(0));
let call_count_clone = call_count.clone();
let score_computer_secondary = move |_segment_reader: &SegmentReader| {
let call_count_new_clone = call_count_clone.clone();
move |_doc: DocId| {
call_count_new_clone.fetch_add(1, AtomicOrdering::SeqCst);
2u32
}
};
let lazy_score_computer = (
(score_computer_primary, Order::Desc),
(score_computer_secondary, Order::Asc),
);
let index = build_test_index();
let searcher = index.reader().unwrap().searcher();
let mut segment_sort_key_computer = lazy_score_computer
.segment_sort_key_computer(searcher.segment_reader(0))
.unwrap();
let expected_sort_key = (200, 2u32);
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(100u32, 1u32));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 1);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(100u32, 3u32));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 2);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(200u32, 1u32));
assert!(sort_key_opt.is_none());
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 3);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(200u32, 3u32));
assert_eq!(sort_key_opt, Some((Ordering::Greater, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(300u32, 1u32));
assert_eq!(sort_key_opt, None);
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &(300u32, 3u32));
assert_eq!(sort_key_opt, None);
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 4);
}
{
let sort_key_opt =
segment_sort_key_computer.accept_sort_key_lazy(0u32, 1f32, &expected_sort_key);
assert_eq!(sort_key_opt, Some((Ordering::Equal, expected_sort_key)));
assert_eq!(call_count.load(AtomicOrdering::SeqCst), 5);
}
assert_eq!(
segment_sort_key_computer.convert_segment_sort_key(expected_sort_key),
(200u32, 2u32)
);
}
}

193
src/collector/sort_key_top_collector.rs
View File

@@ -1,193 +0,0 @@
use std::ops::Range;
use crate::collector::sort_key::{Comparator, SegmentSortKeyComputer, SortKeyComputer};
use crate::collector::{Collector, SegmentCollector, TopNComputer};
use crate::query::Weight;
use crate::schema::Schema;
use crate::{DocAddress, DocId, Result, Score, SegmentReader};
pub(crate) struct TopBySortKeyCollector<TSortKeyComputer> {
sort_key_computer: TSortKeyComputer,
doc_range: Range<usize>,
}
impl<TSortKeyComputer> TopBySortKeyCollector<TSortKeyComputer> {
pub fn new(sort_key_computer: TSortKeyComputer, doc_range: Range<usize>) -> Self {
TopBySortKeyCollector {
sort_key_computer,
doc_range,
}
}
}
impl<TSortKeyComputer> Collector for TopBySortKeyCollector<TSortKeyComputer>
where TSortKeyComputer: SortKeyComputer + Send + Sync + 'static
{
type Fruit = Vec<(TSortKeyComputer::SortKey, DocAddress)>;
type Child =
TopBySortKeySegmentCollector<TSortKeyComputer::Child, TSortKeyComputer::Comparator>;
fn check_schema(&self, schema: &Schema) -> crate::Result<()> {
self.sort_key_computer.check_schema(schema)
}
fn for_segment(&self, segment_ord: u32, segment_reader: &SegmentReader) -> Result<Self::Child> {
let segment_sort_key_computer = self
.sort_key_computer
.segment_sort_key_computer(segment_reader)?;
let topn_computer = TopNComputer::new_with_comparator(
self.doc_range.end,
self.sort_key_computer.comparator(),
);
Ok(TopBySortKeySegmentCollector {
topn_computer,
segment_ord,
segment_sort_key_computer,
})
}
fn requires_scoring(&self) -> bool {
self.sort_key_computer.requires_scoring()
}
fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit> {
Ok(merge_top_k(
segment_fruits.into_iter().flatten(),
self.doc_range.clone(),
self.sort_key_computer.comparator(),
))
}
fn collect_segment(
&self,
weight: &dyn Weight,
segment_ord: u32,
reader: &SegmentReader,
) -> crate::Result<Vec<(TSortKeyComputer::SortKey, DocAddress)>> {
let k = self.doc_range.end;
let docs = self
.sort_key_computer
.collect_segment_top_k(k, weight, reader, segment_ord)?;
Ok(docs)
}
}
fn merge_top_k<D: Ord, TSortKey: Clone + std::fmt::Debug, C: Comparator<TSortKey>>(
sort_key_docs: impl Iterator<Item = (TSortKey, D)>,
doc_range: Range<usize>,
comparator: C,
) -> Vec<(TSortKey, D)> {
if doc_range.is_empty() {
return Vec::new();
}
let mut top_collector: TopNComputer<TSortKey, D, C> =
TopNComputer::new_with_comparator(doc_range.end, comparator);
for (sort_key, doc) in sort_key_docs {
top_collector.push(sort_key, doc);
}
top_collector
.into_sorted_vec()
.into_iter()
.skip(doc_range.start)
.map(|cdoc| (cdoc.sort_key, cdoc.doc))
.collect()
}
pub struct TopBySortKeySegmentCollector<TSegmentSortKeyComputer, C>
where
TSegmentSortKeyComputer: SegmentSortKeyComputer,
C: Comparator<TSegmentSortKeyComputer::SegmentSortKey>,
{
pub(crate) topn_computer: TopNComputer<TSegmentSortKeyComputer::SegmentSortKey, DocId, C>,
pub(crate) segment_ord: u32,
pub(crate) segment_sort_key_computer: TSegmentSortKeyComputer,
}
impl<TSegmentSortKeyComputer, C> SegmentCollector
for TopBySortKeySegmentCollector<TSegmentSortKeyComputer, C>
where
TSegmentSortKeyComputer: 'static + SegmentSortKeyComputer,
C: Comparator<TSegmentSortKeyComputer::SegmentSortKey> + 'static,
{
type Fruit = Vec<(TSegmentSortKeyComputer::SortKey, DocAddress)>;
fn collect(&mut self, doc: DocId, score: Score) {
self.segment_sort_key_computer.compute_sort_key_and_collect(
doc,
score,
&mut self.topn_computer,
);
}
fn harvest(self) -> Self::Fruit {
let segment_ord = self.segment_ord;
let segment_hits: Vec<(TSegmentSortKeyComputer::SortKey, DocAddress)> = self
.topn_computer
.into_vec()
.into_iter()
.map(|comparable_doc| {
let sort_key = self
.segment_sort_key_computer
.convert_segment_sort_key(comparable_doc.sort_key);
(
sort_key,
DocAddress {
segment_ord,
doc_id: comparable_doc.doc,
},
)
})
.collect();
segment_hits
}
}
#[cfg(test)]
mod tests {
use std::ops::Range;
use rand;
use rand::seq::SliceRandom as _;
use super::merge_top_k;
use crate::collector::sort_key::ComparatorEnum;
use crate::Order;
fn test_merge_top_k_aux(
order: Order,
doc_range: Range<usize>,
expected: &[(crate::Score, usize)],
) {
let mut vals: Vec<(crate::Score, usize)> = (0..10).map(|val| (val as f32, val)).collect();
vals.shuffle(&mut rand::thread_rng());
let vals_merged = merge_top_k(vals.into_iter(), doc_range, ComparatorEnum::from(order));
assert_eq!(&vals_merged, expected);
}
#[test]
fn test_merge_top_k() {
test_merge_top_k_aux(Order::Asc, 0..0, &[]);
test_merge_top_k_aux(Order::Asc, 3..3, &[]);
test_merge_top_k_aux(Order::Asc, 0..3, &[(0.0f32, 0), (1.0f32, 1), (2.0f32, 2)]);
test_merge_top_k_aux(
Order::Asc,
0..11,
&[
(0.0f32, 0),
(1.0f32, 1),
(2.0f32, 2),
(3.0f32, 3),
(4.0f32, 4),
(5.0f32, 5),
(6.0f32, 6),
(7.0f32, 7),
(8.0f32, 8),
(9.0f32, 9),
],
);
test_merge_top_k_aux(Order::Asc, 1..3, &[(1.0f32, 1), (2.0f32, 2)]);
test_merge_top_k_aux(Order::Desc, 0..2, &[(9.0f32, 9), (8.0f32, 8)]);
test_merge_top_k_aux(Order::Desc, 2..4, &[(7.0f32, 7), (6.0f32, 6)]);
}
}

11
src/collector/tests.rs
View File

@@ -40,7 +40,7 @@ pub fn test_filter_collector() -> crate::Result<()> {
let filter_some_collector = FilterCollector::new(
"price".to_string(),
&|value: u64| value > 20_120u64,
TopDocs::with_limit(2).order_by_score(),
TopDocs::with_limit(2),
);
let top_docs = searcher.search(&query, &filter_some_collector)?;
@@ -50,7 +50,7 @@ pub fn test_filter_collector() -> crate::Result<()> {
let filter_all_collector: FilterCollector<_, _, u64> = FilterCollector::new(
"price".to_string(),
&|value| value < 5u64,
TopDocs::with_limit(2).order_by_score(),
TopDocs::with_limit(2),
);
let filtered_top_docs = searcher.search(&query, &filter_all_collector).unwrap();
@@ -62,11 +62,8 @@ pub fn test_filter_collector() -> crate::Result<()> {
> 0
}
let filter_dates_collector = FilterCollector::new(
"date".to_string(),
&date_filter,
TopDocs::with_limit(5).order_by_score(),
);
let filter_dates_collector =
FilterCollector::new("date".to_string(), &date_filter, TopDocs::with_limit(5));
let filtered_date_docs = searcher.search(&query, &filter_dates_collector)?;
assert_eq!(filtered_date_docs.len(), 2);

374
src/collector/top_collector.rs
View File

@@ -1,22 +1,374 @@
use std::cmp::Ordering;
use std::marker::PhantomData;
use serde::{Deserialize, Serialize};
use super::top_score_collector::TopNComputer;
use crate::index::SegmentReader;
use crate::{DocAddress, DocId, SegmentOrdinal};
/// Contains a feature (field, score, etc.) of a document along with the document address.
///
/// Used only by TopNComputer, which implements the actual comparison via a `Comparator`.
#[derive(Clone, Default, Eq, PartialEq, Serialize, Deserialize)]
pub struct ComparableDoc<T, D> {
/// It guarantees stable sorting: in case of a tie on the feature, the document
/// address is used.
///
/// The REVERSE_ORDER generic parameter controls whether the by-feature order
/// should be reversed, which is useful for achieving for example largest-first
/// semantics without having to wrap the feature in a `Reverse`.
#[derive(Clone, Default, Serialize, Deserialize)]
pub struct ComparableDoc<T, D, const REVERSE_ORDER: bool = false> {
/// The feature of the document. In practice, this is
/// is a type which can be compared with a `Comparator<T>`.
pub sort_key: T,
/// The document address. In practice, this is either a `DocId` or `DocAddress`.
/// is any type that implements `PartialOrd`.
pub feature: T,
/// The document address. In practice, this is any
/// type that implements `PartialOrd`, and is guaranteed
/// to be unique for each document.
pub doc: D,
}
impl<T: std::fmt::Debug, D: std::fmt::Debug> std::fmt::Debug for ComparableDoc<T, D> {
fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
f.debug_struct("ComparableDoc")
.field("feature", &self.sort_key)
impl<T: std::fmt::Debug, D: std::fmt::Debug, const R: bool> std::fmt::Debug
for ComparableDoc<T, D, R>
{
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.debug_struct(format!("ComparableDoc<_, _ {R}").as_str())
.field("feature", &self.feature)
.field("doc", &self.doc)
.finish()
}
}
impl<T: PartialOrd, D: PartialOrd, const R: bool> PartialOrd for ComparableDoc<T, D, R> {
fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
Some(self.cmp(other))
}
}
impl<T: PartialOrd, D: PartialOrd, const R: bool> Ord for ComparableDoc<T, D, R> {
#[inline]
fn cmp(&self, other: &Self) -> Ordering {
let by_feature = self
.feature
.partial_cmp(&other.feature)
.map(|ord| if R { ord.reverse() } else { ord })
.unwrap_or(Ordering::Equal);
let lazy_by_doc_address = || self.doc.partial_cmp(&other.doc).unwrap_or(Ordering::Equal);
// In case of a tie on the feature, we sort by ascending
// `DocAddress` in order to ensure a stable sorting of the
// documents.
by_feature.then_with(lazy_by_doc_address)
}
}
impl<T: PartialOrd, D: PartialOrd, const R: bool> PartialEq for ComparableDoc<T, D, R> {
fn eq(&self, other: &Self) -> bool {
self.cmp(other) == Ordering::Equal
}
}
impl<T: PartialOrd, D: PartialOrd, const R: bool> Eq for ComparableDoc<T, D, R> {}
pub(crate) struct TopCollector<T> {
pub limit: usize,
pub offset: usize,
_marker: PhantomData<T>,
}
impl<T> TopCollector<T>
where T: PartialOrd + Clone
{
/// Creates a top collector, with a number of documents equal to "limit".
///
/// # Panics
/// The method panics if limit is 0
pub fn with_limit(limit: usize) -> TopCollector<T> {
assert!(limit >= 1, "Limit must be strictly greater than 0.");
Self {
limit,
offset: 0,
_marker: PhantomData,
}
}
/// Skip the first "offset" documents when collecting.
///
/// This is equivalent to `OFFSET` in MySQL or PostgreSQL and `start` in
/// Lucene's TopDocsCollector.
pub fn and_offset(mut self, offset: usize) -> TopCollector<T> {
self.offset = offset;
self
}
pub fn merge_fruits(
&self,
children: Vec<Vec<(T, DocAddress)>>,
) -> crate::Result<Vec<(T, DocAddress)>> {
if self.limit == 0 {
return Ok(Vec::new());
}
let mut top_collector: TopNComputer<_, _> = TopNComputer::new(self.limit + self.offset);
for child_fruit in children {
for (feature, doc) in child_fruit {
top_collector.push(feature, doc);
}
}
Ok(top_collector
.into_sorted_vec()
.into_iter()
.skip(self.offset)
.map(|cdoc| (cdoc.feature, cdoc.doc))
.collect())
}
pub(crate) fn for_segment<F: PartialOrd + Clone>(
&self,
segment_id: SegmentOrdinal,
_: &SegmentReader,
) -> TopSegmentCollector<F> {
TopSegmentCollector::new(segment_id, self.limit + self.offset)
}
/// Create a new TopCollector with the same limit and offset.
///
/// Ideally we would use Into but the blanket implementation seems to cause the Scorer traits
/// to fail.
#[doc(hidden)]
pub(crate) fn into_tscore<TScore: PartialOrd + Clone>(self) -> TopCollector<TScore> {
TopCollector {
limit: self.limit,
offset: self.offset,
_marker: PhantomData,
}
}
}
/// The Top Collector keeps track of the K documents
/// sorted by type `T`.
///
/// The implementation is based on a repeatedly truncating on the median after K * 2 documents
/// The theoretical complexity for collecting the top `K` out of `n` documents
/// is `O(n + K)`.
pub(crate) struct TopSegmentCollector<T> {
/// We reverse the order of the feature in order to
/// have top-semantics instead of bottom semantics.
topn_computer: TopNComputer<T, DocId>,
segment_ord: u32,
}
impl<T: PartialOrd + Clone> TopSegmentCollector<T> {
fn new(segment_ord: SegmentOrdinal, limit: usize) -> TopSegmentCollector<T> {
TopSegmentCollector {
topn_computer: TopNComputer::new(limit),
segment_ord,
}
}
}
impl<T: PartialOrd + Clone> TopSegmentCollector<T> {
pub fn harvest(self) -> Vec<(T, DocAddress)> {
let segment_ord = self.segment_ord;
self.topn_computer
.into_sorted_vec()
.into_iter()
.map(|comparable_doc| {
(
comparable_doc.feature,
DocAddress {
segment_ord,
doc_id: comparable_doc.doc,
},
)
})
.collect()
}
/// Collects a document scored by the given feature
///
/// It collects documents until it has reached the max capacity. Once it reaches capacity, it
/// will compare the lowest scoring item with the given one and keep whichever is greater.
#[inline]
pub fn collect(&mut self, doc: DocId, feature: T) {
self.topn_computer.push(feature, doc);
}
}
#[cfg(test)]
mod tests {
use super::{TopCollector, TopSegmentCollector};
use crate::DocAddress;
#[test]
fn test_top_collector_not_at_capacity() {
let mut top_collector = TopSegmentCollector::new(0, 4);
top_collector.collect(1, 0.8);
top_collector.collect(3, 0.2);
top_collector.collect(5, 0.3);
assert_eq!(
top_collector.harvest(),
vec![
(0.8, DocAddress::new(0, 1)),
(0.3, DocAddress::new(0, 5)),
(0.2, DocAddress::new(0, 3))
]
);
}
#[test]
fn test_top_collector_at_capacity() {
let mut top_collector = TopSegmentCollector::new(0, 4);
top_collector.collect(1, 0.8);
top_collector.collect(3, 0.2);
top_collector.collect(5, 0.3);
top_collector.collect(7, 0.9);
top_collector.collect(9, -0.2);
assert_eq!(
top_collector.harvest(),
vec![
(0.9, DocAddress::new(0, 7)),
(0.8, DocAddress::new(0, 1)),
(0.3, DocAddress::new(0, 5)),
(0.2, DocAddress::new(0, 3))
]
);
}
#[test]
fn test_top_segment_collector_stable_ordering_for_equal_feature() {
// given that the documents are collected in ascending doc id order,
// when harvesting we have to guarantee stable sorting in case of a tie
// on the score
let doc_ids_collection = [4, 5, 6];
let score = 3.3f32;
let mut top_collector_limit_2 = TopSegmentCollector::new(0, 2);
for id in &doc_ids_collection {
top_collector_limit_2.collect(*id, score);
}
let mut top_collector_limit_3 = TopSegmentCollector::new(0, 3);
for id in &doc_ids_collection {
top_collector_limit_3.collect(*id, score);
}
assert_eq!(
top_collector_limit_2.harvest(),
top_collector_limit_3.harvest()[..2].to_vec(),
);
}
#[test]
fn test_top_collector_with_limit_and_offset() {
let collector = TopCollector::with_limit(2).and_offset(1);
let results = collector
.merge_fruits(vec![vec![
(0.9, DocAddress::new(0, 1)),
(0.8, DocAddress::new(0, 2)),
(0.7, DocAddress::new(0, 3)),
(0.6, DocAddress::new(0, 4)),
(0.5, DocAddress::new(0, 5)),
]])
.unwrap();
assert_eq!(
results,
vec![(0.8, DocAddress::new(0, 2)), (0.7, DocAddress::new(0, 3)),]
);
}
#[test]
fn test_top_collector_with_limit_larger_than_set_and_offset() {
let collector = TopCollector::with_limit(2).and_offset(1);
let results = collector
.merge_fruits(vec![vec![
(0.9, DocAddress::new(0, 1)),
(0.8, DocAddress::new(0, 2)),
]])
.unwrap();
assert_eq!(results, vec![(0.8, DocAddress::new(0, 2)),]);
}
#[test]
fn test_top_collector_with_limit_and_offset_larger_than_set() {
let collector = TopCollector::with_limit(2).and_offset(20);
let results = collector
.merge_fruits(vec![vec![
(0.9, DocAddress::new(0, 1)),
(0.8, DocAddress::new(0, 2)),
]])
.unwrap();
assert_eq!(results, vec![]);
}
}
#[cfg(all(test, feature = "unstable"))]
mod bench {
use test::Bencher;
use super::TopSegmentCollector;
#[bench]
fn bench_top_segment_collector_collect_not_at_capacity(b: &mut Bencher) {
let mut top_collector = TopSegmentCollector::new(0, 400);
b.iter(|| {
for i in 0..100 {
top_collector.collect(i, 0.8);
}
});
}
#[bench]
fn bench_top_segment_collector_collect_at_capacity(b: &mut Bencher) {
let mut top_collector = TopSegmentCollector::new(0, 100);
for i in 0..100 {
top_collector.collect(i, 0.8);
}
b.iter(|| {
for i in 0..100 {
top_collector.collect(i, 0.8);
}
});
}
#[bench]
fn bench_top_segment_collector_collect_and_harvest_many_ties(b: &mut Bencher) {
b.iter(|| {
let mut top_collector = TopSegmentCollector::new(0, 100);
for i in 0..100 {
top_collector.collect(i, 0.8);
}
// it would be nice to be able to do the setup N times but still
// measure only harvest(). We can't since harvest() consumes
// the top_collector.
top_collector.harvest()
});
}
#[bench]
fn bench_top_segment_collector_collect_and_harvest_no_tie(b: &mut Bencher) {
b.iter(|| {
let mut top_collector = TopSegmentCollector::new(0, 100);
let mut score = 1.0;
for i in 0..100 {
score += 1.0;
top_collector.collect(i, score);
}
// it would be nice to be able to do the setup N times but still
// measure only harvest(). We can't since harvest() consumes
// the top_collector.
top_collector.harvest()
});
}
}

1276
src/collector/top_score_collector.rs
View File

File diff suppressed because it is too large Load Diff

124
src/collector/tweak_score_top_collector.rs Normal file
View File

@@ -0,0 +1,124 @@
use crate::collector::top_collector::{TopCollector, TopSegmentCollector};
use crate::collector::{Collector, SegmentCollector};
use crate::{DocAddress, DocId, Result, Score, SegmentReader};
pub(crate) struct TweakedScoreTopCollector<TScoreTweaker, TScore = Score> {
score_tweaker: TScoreTweaker,
collector: TopCollector<TScore>,
}
impl<TScoreTweaker, TScore> TweakedScoreTopCollector<TScoreTweaker, TScore>
where TScore: Clone + PartialOrd
{
pub fn new(
score_tweaker: TScoreTweaker,
collector: TopCollector<TScore>,
) -> TweakedScoreTopCollector<TScoreTweaker, TScore> {
TweakedScoreTopCollector {
score_tweaker,
collector,
}
}
}
/// A `ScoreSegmentTweaker` makes it possible to modify the default score
/// for a given document belonging to a specific segment.
///
/// It is the segment local version of the [`ScoreTweaker`].
pub trait ScoreSegmentTweaker<TScore>: 'static {
/// Tweak the given `score` for the document `doc`.
fn score(&mut self, doc: DocId, score: Score) -> TScore;
}
/// `ScoreTweaker` makes it possible to tweak the score
/// emitted by the scorer into another one.
///
/// The `ScoreTweaker` itself does not make much of the computation itself.
/// Instead, it helps constructing `Self::Child` instances that will compute
/// the score at a segment scale.
pub trait ScoreTweaker<TScore>: Sync {
/// Type of the associated [`ScoreSegmentTweaker`].
type Child: ScoreSegmentTweaker<TScore>;
/// Builds a child tweaker for a specific segment. The child scorer is associated with
/// a specific segment.
fn segment_tweaker(&self, segment_reader: &SegmentReader) -> Result<Self::Child>;
}
impl<TScoreTweaker, TScore> Collector for TweakedScoreTopCollector<TScoreTweaker, TScore>
where
TScoreTweaker: ScoreTweaker<TScore> + Send + Sync,
TScore: 'static + PartialOrd + Clone + Send + Sync,
{
type Fruit = Vec<(TScore, DocAddress)>;
type Child = TopTweakedScoreSegmentCollector<TScoreTweaker::Child, TScore>;
fn for_segment(
&self,
segment_local_id: u32,
segment_reader: &SegmentReader,
) -> Result<Self::Child> {
let segment_scorer = self.score_tweaker.segment_tweaker(segment_reader)?;
let segment_collector = self.collector.for_segment(segment_local_id, segment_reader);
Ok(TopTweakedScoreSegmentCollector {
segment_collector,
segment_scorer,
})
}
fn requires_scoring(&self) -> bool {
true
}
fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit> {
self.collector.merge_fruits(segment_fruits)
}
}
pub struct TopTweakedScoreSegmentCollector<TSegmentScoreTweaker, TScore>
where
TScore: 'static + PartialOrd + Clone + Send + Sync + Sized,
TSegmentScoreTweaker: ScoreSegmentTweaker<TScore>,
{
segment_collector: TopSegmentCollector<TScore>,
segment_scorer: TSegmentScoreTweaker,
}
impl<TSegmentScoreTweaker, TScore> SegmentCollector
for TopTweakedScoreSegmentCollector<TSegmentScoreTweaker, TScore>
where
TScore: 'static + PartialOrd + Clone + Send + Sync,
TSegmentScoreTweaker: 'static + ScoreSegmentTweaker<TScore>,
{
type Fruit = Vec<(TScore, DocAddress)>;
fn collect(&mut self, doc: DocId, score: Score) {
let score = self.segment_scorer.score(doc, score);
self.segment_collector.collect(doc, score);
}
fn harvest(self) -> Vec<(TScore, DocAddress)> {
self.segment_collector.harvest()
}
}
impl<F, TScore, TSegmentScoreTweaker> ScoreTweaker<TScore> for F
where
F: 'static + Send + Sync + Fn(&SegmentReader) -> TSegmentScoreTweaker,
TSegmentScoreTweaker: ScoreSegmentTweaker<TScore>,
{
type Child = TSegmentScoreTweaker;
fn segment_tweaker(&self, segment_reader: &SegmentReader) -> Result<Self::Child> {
Ok((self)(segment_reader))
}
}
impl<F, TScore> ScoreSegmentTweaker<TScore> for F
where F: 'static + FnMut(DocId, Score) -> TScore
{
fn score(&mut self, doc: DocId, score: Score) -> TScore {
(self)(doc, score)
}
}

2
src/compat_tests.rs
View File

@@ -69,7 +69,7 @@ fn assert_date_time_precision(index: &Index, doc_store_precision: DateTimePrecis
.parse_query("dateformat")
.expect("Failed to parse query");
let top_docs = searcher
.search(&query, &TopDocs::with_limit(1).order_by_score())
.search(&query, &TopDocs::with_limit(1))
.expect("Search failed");
assert_eq!(top_docs.len(), 1, "Expected 1 search result");

27
src/core/json_utils.rs
View File

@@ -3,7 +3,6 @@ use common::json_path_writer::{JSON_END_OF_PATH, JSON_PATH_SEGMENT_SEP};
use common::{replace_in_place, JsonPathWriter};
use rustc_hash::FxHashMap;
use crate::indexer::indexing_term::IndexingTerm;
use crate::postings::{IndexingContext, IndexingPosition, PostingsWriter};
use crate::schema::document::{ReferenceValue, ReferenceValueLeaf, Value};
use crate::schema::{Type, DATE_TIME_PRECISION_INDEXED};
@@ -78,7 +77,7 @@ fn index_json_object<'a, V: Value<'a>>(
doc: DocId,
json_visitor: V::ObjectIter,
text_analyzer: &mut TextAnalyzer,
term_buffer: &mut IndexingTerm,
term_buffer: &mut Term,
json_path_writer: &mut JsonPathWriter,
postings_writer: &mut dyn PostingsWriter,
ctx: &mut IndexingContext,
@@ -108,17 +107,17 @@ pub(crate) fn index_json_value<'a, V: Value<'a>>(
doc: DocId,
json_value: V,
text_analyzer: &mut TextAnalyzer,
term_buffer: &mut IndexingTerm,
term_buffer: &mut Term,
json_path_writer: &mut JsonPathWriter,
postings_writer: &mut dyn PostingsWriter,
ctx: &mut IndexingContext,
positions_per_path: &mut IndexingPositionsPerPath,
) {
let set_path_id = |term_buffer: &mut IndexingTerm, unordered_id: u32| {
let set_path_id = |term_buffer: &mut Term, unordered_id: u32| {
term_buffer.truncate_value_bytes(0);
term_buffer.append_bytes(&unordered_id.to_be_bytes());
};
let set_type = |term_buffer: &mut IndexingTerm, typ: Type| {
let set_type = |term_buffer: &mut Term, typ: Type| {
term_buffer.append_bytes(&[typ.to_code()]);
};
@@ -406,7 +405,7 @@ mod tests {
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_str("red");
assert_eq!(term.serialized_value_bytes(), b"color\x00sred".to_vec())
assert_eq!(term.serialized_term(), b"\x00\x00\x00\x01jcolor\x00sred")
}
#[test]
@@ -416,8 +415,8 @@ mod tests {
term.append_type_and_fast_value(-4i64);
assert_eq!(
term.serialized_value_bytes(),
b"color\x00i\x7f\xff\xff\xff\xff\xff\xff\xfc".to_vec()
term.serialized_term(),
b"\x00\x00\x00\x01jcolor\x00i\x7f\xff\xff\xff\xff\xff\xff\xfc"
)
}
@@ -428,8 +427,8 @@ mod tests {
term.append_type_and_fast_value(4u64);
assert_eq!(
term.serialized_value_bytes(),
b"color\x00u\x00\x00\x00\x00\x00\x00\x00\x04".to_vec()
term.serialized_term(),
b"\x00\x00\x00\x01jcolor\x00u\x00\x00\x00\x00\x00\x00\x00\x04"
)
}
@@ -439,8 +438,8 @@ mod tests {
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(4.0f64);
assert_eq!(
term.serialized_value_bytes(),
b"color\x00f\xc0\x10\x00\x00\x00\x00\x00\x00".to_vec()
term.serialized_term(),
b"\x00\x00\x00\x01jcolor\x00f\xc0\x10\x00\x00\x00\x00\x00\x00"
)
}
@@ -450,8 +449,8 @@ mod tests {
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(true);
assert_eq!(
term.serialized_value_bytes(),
b"color\x00o\x00\x00\x00\x00\x00\x00\x00\x01".to_vec()
term.serialized_term(),
b"\x00\x00\x00\x01jcolor\x00o\x00\x00\x00\x00\x00\x00\x00\x01"
)
}

1
src/core/searcher.rs
View File

@@ -225,7 +225,6 @@ impl Searcher {
enabled_scoring: EnableScoring,
) -> crate::Result<C::Fruit> {
let weight = query.weight(enabled_scoring)?;
collector.check_schema(self.schema())?;
let segment_readers = self.segment_readers();
let fruits = executor.map(
|(segment_ord, segment_reader)| {

7
src/directory/composite_file.rs
View File

@@ -5,7 +5,7 @@ use std::ops::Range;
use common::{BinarySerializable, CountingWriter, HasLen, VInt};
use crate::directory::{FileSlice, TerminatingWrite, WritePtr};
use crate::schema::{Field, Schema};
use crate::schema::Field;
use crate::space_usage::{FieldUsage, PerFieldSpaceUsage};
#[derive(Eq, PartialEq, Hash, Copy, Ord, PartialOrd, Clone, Debug)]
@@ -167,11 +167,10 @@ impl CompositeFile {
.map(|byte_range| self.data.slice(byte_range.clone()))
}
pub fn space_usage(&self, schema: &Schema) -> PerFieldSpaceUsage {
pub fn space_usage(&self) -> PerFieldSpaceUsage {
let mut fields = Vec::new();
for (&field_addr, byte_range) in &self.offsets_index {
let field_name = schema.get_field_name(field_addr.field).to_string();
let mut field_usage = FieldUsage::empty(field_name);
let mut field_usage = FieldUsage::empty(field_addr.field);
field_usage.add_field_idx(field_addr.idx, byte_range.len().into());
fields.push(field_usage);
}

2
src/directory/directory.rs
View File

@@ -108,7 +108,7 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
/// Opens a file and returns a boxed `FileHandle`.
///
/// Users of `Directory` should typically call `Directory::open_read(...)`,
/// while `Directory` implementer should implement `get_file_handle()`.
/// while `Directory` implementor should implement `get_file_handle()`.
fn get_file_handle(&self, path: &Path) -> Result<Arc<dyn FileHandle>, OpenReadError>;
/// Once a virtual file is open, its data may not

0
src/directory/mmap_directory/file_watcher.rs → src/directory/file_watcher.rs
View File

4
src/directory/mmap_directory/mod.rs → src/directory/mmap_directory.rs
View File

@@ -1,5 +1,3 @@
mod file_watcher;
use std::collections::HashMap;
use std::fmt;
use std::fs::{self, File, OpenOptions};
@@ -9,7 +7,6 @@ use std::path::{Path, PathBuf};
use std::sync::{Arc, RwLock, Weak};
use common::StableDeref;
use file_watcher::FileWatcher;
use fs4::fs_std::FileExt;
#[cfg(all(feature = "mmap", unix))]
pub use memmap2::Advice;
@@ -21,6 +18,7 @@ use crate::core::META_FILEPATH;
use crate::directory::error::{
DeleteError, LockError, OpenDirectoryError, OpenReadError, OpenWriteError,
};
use crate::directory::file_watcher::FileWatcher;
use crate::directory::{
AntiCallToken, Directory, DirectoryLock, FileHandle, Lock, OwnedBytes, TerminatingWrite,
WatchCallback, WatchHandle, WritePtr,

1
src/directory/mod.rs
View File

@@ -5,6 +5,7 @@ mod mmap_directory;
mod directory;
mod directory_lock;
mod file_watcher;
pub mod footer;
mod managed_directory;
mod ram_directory;

67
src/docset.rs
View File

@@ -40,8 +40,6 @@ pub trait DocSet: Send {
/// of `DocSet` should support it.
///
/// Calling `seek(TERMINATED)` is also legal and is the normal way to consume a `DocSet`.
///
/// `target` has to be larger or equal to `.doc()` when calling `seek`.
fn seek(&mut self, target: DocId) -> DocId {
let mut doc = self.doc();
debug_assert!(doc <= target);
@@ -51,33 +49,6 @@ pub trait DocSet: Send {
doc
}
/// Seeks to the target if possible and returns true if the target is in the DocSet.
///
/// DocSets that already have an efficient `seek` method don't need to implement
/// `seek_into_the_danger_zone`. All wrapper DocSets should forward
/// `seek_into_the_danger_zone` to the underlying DocSet.
///
/// ## API Behaviour
/// If `seek_into_the_danger_zone` is returning true, a call to `doc()` has to return target.
/// If `seek_into_the_danger_zone` is returning false, a call to `doc()` may return any doc
/// between the last doc that matched and target or a doc that is a valid next hit after
/// target. The DocSet is considered to be in an invalid state until
/// `seek_into_the_danger_zone` returns true again.
///
/// `target` needs to be equal or larger than `doc` when in a valid state.
///
/// Consecutive calls are not allowed to have decreasing `target` values.
///
/// # Warning
/// This is an advanced API used by intersection. The API contract is tricky, avoid using it.
fn seek_into_the_danger_zone(&mut self, target: DocId) -> bool {
let current_doc = self.doc();
if current_doc < target {
self.seek(target);
}
self.doc() == target
}
/// Fills a given mutable buffer with the next doc ids from the
/// `DocSet`
///
@@ -116,26 +87,6 @@ pub trait DocSet: Send {
/// length of the docset.
fn size_hint(&self) -> u32;
/// Returns a best-effort hint of the cost to consume the entire docset.
///
/// Consuming means calling advance until [`TERMINATED`] is returned.
/// The cost should be relative to the cost of driving a Term query,
/// which would be the number of documents in the DocSet.
///
/// By default this returns `size_hint()`.
///
/// DocSets may have vastly different cost depending on their type,
/// e.g. an intersection with 10 hits is much cheaper than
/// a phrase search with 10 hits, since it needs to load positions.
///
/// ### Future Work
/// We may want to differentiate `DocSet` costs more more granular, e.g.
/// creation_cost, advance_cost, seek_cost on to get a good estimation
/// what query types to choose.
fn cost(&self) -> u64 {
self.size_hint() as u64
}
/// Returns the number documents matching.
/// Calling this method consumes the `DocSet`.
fn count(&mut self, alive_bitset: &AliveBitSet) -> u32 {
@@ -175,10 +126,6 @@ impl DocSet for &mut dyn DocSet {
(**self).seek(target)
}
fn seek_into_the_danger_zone(&mut self, target: DocId) -> bool {
(**self).seek_into_the_danger_zone(target)
}
fn doc(&self) -> u32 {
(**self).doc()
}
@@ -187,10 +134,6 @@ impl DocSet for &mut dyn DocSet {
(**self).size_hint()
}
fn cost(&self) -> u64 {
(**self).cost()
}
fn count(&mut self, alive_bitset: &AliveBitSet) -> u32 {
(**self).count(alive_bitset)
}
@@ -211,11 +154,6 @@ impl<TDocSet: DocSet + ?Sized> DocSet for Box<TDocSet> {
unboxed.seek(target)
}
fn seek_into_the_danger_zone(&mut self, target: DocId) -> bool {
let unboxed: &mut TDocSet = self.borrow_mut();
unboxed.seek_into_the_danger_zone(target)
}
fn fill_buffer(&mut self, buffer: &mut [DocId; COLLECT_BLOCK_BUFFER_LEN]) -> usize {
let unboxed: &mut TDocSet = self.borrow_mut();
unboxed.fill_buffer(buffer)
@@ -231,11 +169,6 @@ impl<TDocSet: DocSet + ?Sized> DocSet for Box<TDocSet> {
unboxed.size_hint()
}
fn cost(&self) -> u64 {
let unboxed: &TDocSet = self.borrow();
unboxed.cost()
}
fn count(&mut self, alive_bitset: &AliveBitSet) -> u32 {
let unboxed: &mut TDocSet = self.borrow_mut();
unboxed.count(alive_bitset)

2
src/error.rs
View File

@@ -104,7 +104,7 @@ pub enum TantivyError {
#[error("{0:?}")]
IncompatibleIndex(Incompatibility),
/// An internal error occurred. This is are internal states that should not be reached.
/// e.g. a datastructure is incorrectly initialized.
/// e.g. a datastructure is incorrectly inititalized.
#[error("Internal error: '{0}'")]
InternalError(String),
#[error("Deserialize error: {0}")]

8
src/fastfield/mod.rs
View File

@@ -726,22 +726,22 @@ mod tests {
.column_opt::<DateTime>("multi_date")
.unwrap()
.unwrap();
let mut dates = Vec::new();
{
assert_eq!(date_fast_field.get_val(0).into_timestamp_nanos(), 1i64);
let dates: Vec<DateTime> = dates_fast_field.values_for_doc(0u32).collect();
dates_fast_field.fill_vals(0u32, &mut dates);
assert_eq!(dates.len(), 2);
assert_eq!(dates[0].into_timestamp_nanos(), 2i64);
assert_eq!(dates[1].into_timestamp_nanos(), 3i64);
}
{
assert_eq!(date_fast_field.get_val(1).into_timestamp_nanos(), 4i64);
let dates: Vec<DateTime> = dates_fast_field.values_for_doc(1u32).collect();
dates_fast_field.fill_vals(1u32, &mut dates);
assert!(dates.is_empty());
}
{
assert_eq!(date_fast_field.get_val(2).into_timestamp_nanos(), 0i64);
let dates: Vec<DateTime> = dates_fast_field.values_for_doc(2u32).collect();
dates_fast_field.fill_vals(2u32, &mut dates);
assert_eq!(dates.len(), 2);
assert_eq!(dates[0].into_timestamp_nanos(), 5i64);
assert_eq!(dates[1].into_timestamp_nanos(), 6i64);

18
src/fastfield/readers.rs
View File

@@ -8,7 +8,7 @@ use columnar::{
};
use common::ByteCount;
use crate::core::json_utils::{encode_column_name, json_path_sep_to_dot};
use crate::core::json_utils::encode_column_name;
use crate::directory::FileSlice;
use crate::schema::{Field, FieldEntry, FieldType, Schema};
use crate::space_usage::{FieldUsage, PerFieldSpaceUsage};
@@ -39,15 +39,19 @@ impl FastFieldReaders {
self.resolve_column_name_given_default_field(column_name, default_field_opt)
}
pub(crate) fn space_usage(&self) -> io::Result<PerFieldSpaceUsage> {
pub(crate) fn space_usage(&self, schema: &Schema) -> io::Result<PerFieldSpaceUsage> {
let mut per_field_usages: Vec<FieldUsage> = Default::default();
for (mut field_name, column_handle) in self.columnar.iter_columns()? {
json_path_sep_to_dot(&mut field_name);
let space_usage = column_handle.space_usage()?;
let mut field_usage = FieldUsage::empty(field_name);
field_usage.set_column_usage(space_usage);
for (field, field_entry) in schema.fields() {
let column_handles = self.columnar.read_columns(field_entry.name())?;
let num_bytes: ByteCount = column_handles
.iter()
.map(|column_handle| column_handle.num_bytes())
.sum();
let mut field_usage = FieldUsage::empty(field);
field_usage.add_field_idx(0, num_bytes);
per_field_usages.push(field_usage);
}
// TODO fix space usage for JSON fields.
Ok(PerFieldSpaceUsage::new(per_field_usages))
}

6
src/fieldnorm/reader.rs
View File

@@ -2,7 +2,7 @@ use std::sync::Arc;
use super::{fieldnorm_to_id, id_to_fieldnorm};
use crate::directory::{CompositeFile, FileSlice, OwnedBytes};
use crate::schema::{Field, Schema};
use crate::schema::Field;
use crate::space_usage::PerFieldSpaceUsage;
use crate::DocId;
@@ -37,8 +37,8 @@ impl FieldNormReaders {
}
/// Return a break down of the space usage per field.
pub fn space_usage(&self, schema: &Schema) -> PerFieldSpaceUsage {
self.data.space_usage(schema)
pub fn space_usage(&self) -> PerFieldSpaceUsage {
self.data.space_usage()
}
/// Returns a handle to inner file

18
src/index/index_meta.rs
View File

@@ -13,9 +13,9 @@ use crate::store::Compressor;
use crate::{Inventory, Opstamp, TrackedObject};
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct DeleteMeta {
struct DeleteMeta {
num_deleted_docs: u32,
pub opstamp: Opstamp,
opstamp: Opstamp,
}
#[derive(Clone, Default)]
@@ -213,7 +213,7 @@ impl SegmentMeta {
struct InnerSegmentMeta {
segment_id: SegmentId,
max_doc: u32,
pub deletes: Option<DeleteMeta>,
deletes: Option<DeleteMeta>,
/// If you want to avoid the SegmentComponent::TempStore file to be covered by
/// garbage collection and deleted, set this to true. This is used during merge.
#[serde(skip)]
@@ -276,14 +276,13 @@ impl Default for IndexSettings {
}
/// The order to sort by
#[derive(Clone, Copy, Debug, Serialize, Deserialize, Eq, PartialEq)]
#[derive(Clone, Debug, Serialize, Deserialize, Eq, PartialEq)]
pub enum Order {
/// Ascending Order
Asc,
/// Descending Order
Desc,
}
impl Order {
/// return if the Order is ascending
pub fn is_asc(&self) -> bool {
@@ -404,10 +403,7 @@ mod tests {
schema_builder.build()
};
let index_metas = IndexMeta {
index_settings: IndexSettings {
docstore_compression: Compressor::None,
..Default::default()
},
index_settings: IndexSettings::default(),
segments: Vec::new(),
schema,
opstamp: 0u64,
@@ -416,7 +412,7 @@ mod tests {
let json = serde_json::ser::to_string(&index_metas).expect("serialization failed");
assert_eq!(
json,
r#"{"index_settings":{"docstore_compression":"none","docstore_blocksize":16384},"segments":[],"schema":[{"name":"text","type":"text","options":{"indexing":{"record":"position","fieldnorms":true,"tokenizer":"default"},"stored":false,"fast":false}}],"opstamp":0}"#
r#"{"index_settings":{"docstore_compression":"lz4","docstore_blocksize":16384},"segments":[],"schema":[{"name":"text","type":"text","options":{"indexing":{"record":"position","fieldnorms":true,"tokenizer":"default"},"stored":false,"fast":false}}],"opstamp":0}"#
);
let deser_meta: UntrackedIndexMeta = serde_json::from_str(&json).unwrap();
@@ -497,8 +493,6 @@ mod tests {
#[test]
#[cfg(feature = "lz4-compression")]
fn test_index_settings_default() {
use crate::store::Compressor;
let mut index_settings = IndexSettings::default();
assert_eq!(
index_settings,

2
src/index/segment.rs
View File

@@ -46,7 +46,7 @@ impl Segment {
///
/// This method is only used when updating `max_doc` from 0
/// as we finalize a fresh new segment.
pub fn with_max_doc(self, max_doc: u32) -> Segment {
pub(crate) fn with_max_doc(self, max_doc: u32) -> Segment {
Segment {
index: self.index,
meta: self.meta.with_max_doc(max_doc),

16
src/index/segment_reader.rs
View File

@@ -455,11 +455,11 @@ impl SegmentReader {
pub fn space_usage(&self) -> io::Result<SegmentSpaceUsage> {
Ok(SegmentSpaceUsage::new(
self.num_docs(),
self.termdict_composite.space_usage(self.schema()),
self.postings_composite.space_usage(self.schema()),
self.positions_composite.space_usage(self.schema()),
self.fast_fields_readers.space_usage()?,
self.fieldnorm_readers.space_usage(self.schema()),
self.termdict_composite.space_usage(),
self.postings_composite.space_usage(),
self.positions_composite.space_usage(),
self.fast_fields_readers.space_usage(self.schema())?,
self.fieldnorm_readers.space_usage(),
self.get_store_reader(0)?.space_usage(),
self.alive_bitset_opt
.as_ref()
@@ -608,7 +608,7 @@ mod test {
term_dictionary_size: Some(ByteCount::from(100u64)),
postings_size: Some(ByteCount::from(1_000u64)),
positions_size: Some(ByteCount::from(2_000u64)),
fast_size: Some(ByteCount::from(1_000u64)),
fast_size: Some(ByteCount::from(1_000u64).into()),
};
let field_metadata2 = FieldMetadata {
field_name: "a".to_string(),
@@ -617,7 +617,7 @@ mod test {
term_dictionary_size: Some(ByteCount::from(80u64)),
postings_size: Some(ByteCount::from(1_500u64)),
positions_size: Some(ByteCount::from(2_500u64)),
fast_size: Some(ByteCount::from(3_000u64)),
fast_size: Some(ByteCount::from(3_000u64).into()),
};
let expected = FieldMetadata {
field_name: "a".to_string(),
@@ -626,7 +626,7 @@ mod test {
term_dictionary_size: Some(ByteCount::from(180u64)),
postings_size: Some(ByteCount::from(2_500u64)),
positions_size: Some(ByteCount::from(4_500u64)),
fast_size: Some(ByteCount::from(4_000u64)),
fast_size: Some(ByteCount::from(4_000u64).into()),
};
assert_merge(
&[vec![field_metadata1.clone()], vec![field_metadata2]],

54
src/indexer/delete_queue.rs
View File

@@ -4,37 +4,38 @@ use std::sync::{Arc, RwLock, Weak};
use super::operation::DeleteOperation;
use crate::Opstamp;
/// The DeleteQueue is similar in conceptually to a multiple
/// consumer single producer broadcast channel.
///
/// All consumer will receive all messages.
///
/// Consumer of the delete queue are holding a `DeleteCursor`,
/// which points to a specific place of the `DeleteQueue`.
///
/// New consumer can be created in two ways
/// - calling `delete_queue.cursor()` returns a cursor, that will include all future delete
/// operation (and some or none of the past operations... The client is in charge of checking the
/// opstamps.).
/// - cloning an existing cursor returns a new cursor, that is at the exact same position, and can
/// now advance independently from the original cursor.
// The DeleteQueue is similar in conceptually to a multiple
// consumer single producer broadcast channel.
//
// All consumer will receive all messages.
//
// Consumer of the delete queue are holding a `DeleteCursor`,
// which points to a specific place of the `DeleteQueue`.
//
// New consumer can be created in two ways
// - calling `delete_queue.cursor()` returns a cursor, that will include all future delete operation
// (and some or none of the past operations... The client is in charge of checking the opstamps.).
// - cloning an existing cursor returns a new cursor, that is at the exact same position, and can
// now advance independently from the original cursor.
#[derive(Default)]
struct InnerDeleteQueue {
writer: Vec<DeleteOperation>,
last_block: Weak<Block>,
}
/// The delete queue is a linked list storing delete operations.
///
/// Several consumers can hold a reference to it. Delete operations
/// get dropped/gc'ed when no more consumers are holding a reference
/// to them.
#[derive(Clone, Default)]
#[derive(Clone)]
pub struct DeleteQueue {
inner: Arc<RwLock<InnerDeleteQueue>>,
}
impl DeleteQueue {
// Creates a new delete queue.
pub fn new() -> DeleteQueue {
DeleteQueue {
inner: Arc::default(),
}
}
fn get_last_block(&self) -> Arc<Block> {
{
// try get the last block with simply acquiring the read lock.
@@ -57,10 +58,10 @@ impl DeleteQueue {
block
}
/// Creates a new cursor that makes it possible to
/// consume future delete operations.
///
/// Past delete operations are not accessible.
// Creates a new cursor that makes it possible to
// consume future delete operations.
//
// Past delete operations are not accessible.
pub fn cursor(&self) -> DeleteCursor {
let last_block = self.get_last_block();
let operations_len = last_block.operations.len();
@@ -70,7 +71,7 @@ impl DeleteQueue {
}
}
/// Appends a new delete operations.
// Appends a new delete operations.
pub fn push(&self, delete_operation: DeleteOperation) {
self.inner
.write()
@@ -168,7 +169,6 @@ struct Block {
next: NextBlock,
}
/// As we process delete operations, keeps track of our position.
#[derive(Clone)]
pub struct DeleteCursor {
block: Arc<Block>,
@@ -261,7 +261,7 @@ mod tests {
#[test]
fn test_deletequeue() {
let delete_queue = DeleteQueue::default();
let delete_queue = DeleteQueue::new();
let make_op = |i: usize| DeleteOperation {
opstamp: i as u64,

30
src/indexer/index_writer.rs
View File

@@ -128,7 +128,7 @@ fn compute_deleted_bitset(
/// is `==` target_opstamp.
/// For instance, there was no delete operation between the state of the `segment_entry` and
/// the `target_opstamp`, `segment_entry` is not updated.
pub fn advance_deletes(
pub(crate) fn advance_deletes(
mut segment: Segment,
segment_entry: &mut SegmentEntry,
target_opstamp: Opstamp,
@@ -303,7 +303,7 @@ impl<D: Document> IndexWriter<D> {
let (document_sender, document_receiver) =
crossbeam_channel::bounded(PIPELINE_MAX_SIZE_IN_DOCS);
let delete_queue = DeleteQueue::default();
let delete_queue = DeleteQueue::new();
let current_opstamp = index.load_metas()?.opstamp;
@@ -513,7 +513,7 @@ impl<D: Document> IndexWriter<D> {
/// let searcher = index.reader()?.searcher();
/// let query_parser = QueryParser::for_index(&index, vec![title]);
/// let query_promo = query_parser.parse_query("Prometheus")?;
/// let top_docs_promo = searcher.search(&query_promo, &TopDocs::with_limit(1).order_by_score())?;
/// let top_docs_promo = searcher.search(&query_promo, &TopDocs::with_limit(1))?;
///
/// assert!(top_docs_promo.is_empty());
/// Ok(())
@@ -946,11 +946,11 @@ mod tests {
let searcher = reader.searcher();
let a_docs = searcher
.search(&a_query, &TopDocs::with_limit(1).order_by_score())
.search(&a_query, &TopDocs::with_limit(1))
.expect("search for a failed");
let b_docs = searcher
.search(&b_query, &TopDocs::with_limit(1).order_by_score())
.search(&b_query, &TopDocs::with_limit(1))
.expect("search for b failed");
assert_eq!(a_docs.len(), 1);
@@ -2014,9 +2014,8 @@ mod tests {
let query = QueryParser::for_index(&index, vec![field])
.parse_query(term)
.unwrap();
let top_docs: Vec<(f32, DocAddress)> = searcher
.search(&query, &TopDocs::with_limit(1000).order_by_score())
.unwrap();
let top_docs: Vec<(f32, DocAddress)> =
searcher.search(&query, &TopDocs::with_limit(1000)).unwrap();
top_docs.iter().map(|el| el.1).collect::<Vec<_>>()
};
@@ -2450,9 +2449,8 @@ mod tests {
Term::from_field_u64(id_field, existing_id),
IndexRecordOption::Basic,
);
let top_docs: Vec<(f32, DocAddress)> = searcher
.search(&query, &TopDocs::with_limit(10).order_by_score())
.unwrap();
let top_docs: Vec<(f32, DocAddress)> =
searcher.search(&query, &TopDocs::with_limit(10)).unwrap();
assert_eq!(top_docs.len(), 1); // Was failing
@@ -2493,9 +2491,8 @@ mod tests {
Term::from_field_i64(id_field, 10i64),
IndexRecordOption::Basic,
);
let top_docs: Vec<(f32, DocAddress)> = searcher
.search(&query, &TopDocs::with_limit(10).order_by_score())
.unwrap();
let top_docs: Vec<(f32, DocAddress)> =
searcher.search(&query, &TopDocs::with_limit(10)).unwrap();
assert_eq!(top_docs.len(), 1); // Fails
@@ -2503,9 +2500,8 @@ mod tests {
Term::from_field_i64(id_field, 30i64),
IndexRecordOption::Basic,
);
let top_docs: Vec<(f32, DocAddress)> = searcher
.search(&query, &TopDocs::with_limit(10).order_by_score())
.unwrap();
let top_docs: Vec<(f32, DocAddress)> =
searcher.search(&query, &TopDocs::with_limit(10)).unwrap();
assert_eq!(top_docs.len(), 1); // Fails

214
src/indexer/indexing_term.rs
View File

@@ -1,214 +0,0 @@
use std::net::Ipv6Addr;
use columnar::MonotonicallyMappableToU128;
use crate::fastfield::FastValue;
use crate::schema::Field;
/// IndexingTerm is used to represent a term during indexing.
/// It's a serialized representation over field and value.
///
/// It actually wraps a `Vec<u8>`. The first 4 bytes are the field.
///
/// We serialize the field, because we index everything in a single
/// global term dictionary during indexing.
#[derive(Clone)]
pub(crate) struct IndexingTerm<B = Vec<u8>>(B)
where B: AsRef<[u8]>;
/// The number of bytes used as metadata by `Term`.
const TERM_METADATA_LENGTH: usize = 4;
impl IndexingTerm {
/// Create a new Term with a buffer with a given capacity.
pub fn with_capacity(capacity: usize) -> IndexingTerm {
let mut data = Vec::with_capacity(TERM_METADATA_LENGTH + capacity);
data.resize(TERM_METADATA_LENGTH, 0u8);
IndexingTerm(data)
}
/// Panics when the term is not empty... ie: some value is set.
/// Use `clear_with_field_and_type` in that case.
///
/// Sets field and the type.
pub(crate) fn set_field(&mut self, field: Field) {
assert!(self.is_empty());
self.0[0..4].clone_from_slice(field.field_id().to_be_bytes().as_ref());
}
/// Is empty if there are no value bytes.
pub fn is_empty(&self) -> bool {
self.0.len() == TERM_METADATA_LENGTH
}
/// Removes the value_bytes and set the field
pub(crate) fn clear_with_field(&mut self, field: Field) {
self.truncate_value_bytes(0);
self.set_field(field);
}
/// Sets a u64 value in the term.
///
/// U64 are serialized using (8-byte) BigEndian
/// representation.
/// The use of BigEndian has the benefit of preserving
/// the natural order of the values.
pub fn set_u64(&mut self, val: u64) {
self.set_fast_value(val);
}
/// Sets a `i64` value in the term.
pub fn set_i64(&mut self, val: i64) {
self.set_fast_value(val);
}
/// Sets a `f64` value in the term.
pub fn set_f64(&mut self, val: f64) {
self.set_fast_value(val);
}
/// Sets a `bool` value in the term.
pub fn set_bool(&mut self, val: bool) {
self.set_fast_value(val);
}
fn set_fast_value<T: FastValue>(&mut self, val: T) {
self.set_bytes(val.to_u64().to_be_bytes().as_ref());
}
/// Append a type marker + fast value to a term.
/// This is used in JSON type to append a fast value after the path.
///
/// It will not clear existing bytes.
pub fn append_type_and_fast_value<T: FastValue>(&mut self, val: T) {
self.0.push(T::to_type().to_code());
let value = val.to_u64();
self.0.extend(value.to_be_bytes().as_ref());
}
/// Sets a `Ipv6Addr` value in the term.
pub fn set_ip_addr(&mut self, val: Ipv6Addr) {
self.set_bytes(val.to_u128().to_be_bytes().as_ref());
}
/// Sets the value of a `Bytes` field.
pub fn set_bytes(&mut self, bytes: &[u8]) {
self.truncate_value_bytes(0);
self.0.extend(bytes);
}
/// Truncates the value bytes of the term. Value and field type stays the same.
pub fn truncate_value_bytes(&mut self, len: usize) {
self.0.truncate(len + TERM_METADATA_LENGTH);
}
/// The length of the bytes.
pub fn len_bytes(&self) -> usize {
self.0.len() - TERM_METADATA_LENGTH
}
/// Appends value bytes to the Term.
///
/// This function returns the segment that has just been added.
#[inline]
pub fn append_bytes(&mut self, bytes: &[u8]) -> &mut [u8] {
let len_before = self.0.len();
self.0.extend_from_slice(bytes);
&mut self.0[len_before..]
}
}
impl<B> IndexingTerm<B>
where B: AsRef<[u8]>
{
/// Wraps serialized term bytes.
///
/// The input buffer is expected to be the concatenation of the big endian encoded field id
/// followed by the serialized value bytes (type tag + payload).
#[inline]
pub fn wrap(serialized_term: B) -> IndexingTerm<B> {
debug_assert!(serialized_term.as_ref().len() >= TERM_METADATA_LENGTH);
IndexingTerm(serialized_term)
}
/// Returns the field this term belongs to.
#[inline]
pub fn field(&self) -> Field {
let field_id_bytes: [u8; 4] = self.0.as_ref()[..4].try_into().unwrap();
Field::from_field_id(u32::from_be_bytes(field_id_bytes))
}
/// Returns the serialized representation of Term.
/// This includes field_id, value type and value.
///
/// Do NOT rely on this byte representation in the index.
/// This value is likely to change in the future.
#[inline]
pub fn serialized_term(&self) -> &[u8] {
self.0.as_ref()
}
}
#[cfg(test)]
mod tests {
use super::IndexingTerm;
use crate::schema::*;
#[test]
pub fn test_term_str() {
let mut schema_builder = Schema::builder();
schema_builder.add_text_field("text", STRING);
let title_field = schema_builder.add_text_field("title", STRING);
let mut term = IndexingTerm::with_capacity(0);
term.set_field(title_field);
term.set_bytes(b"test");
assert_eq!(term.field(), title_field);
assert_eq!(term.serialized_term(), b"\x00\x00\x00\x01test".to_vec())
}
/// Size (in bytes) of the buffer of a fast value (u64, i64, f64, or date) term.
/// <field> + <type byte> + <value len>
///
/// - <field> is a big endian encoded u32 field id
/// - <value> is, if this is not the json term, a binary representation specific to the type.
/// If it is a JSON Term, then it is prepended with the path that leads to this leaf value.
const FAST_VALUE_TERM_LEN: usize = 4 + 8;
#[test]
pub fn test_term_u64() {
let mut schema_builder = Schema::builder();
let count_field = schema_builder.add_u64_field("count", INDEXED);
let mut term = IndexingTerm::with_capacity(0);
term.set_field(count_field);
term.set_u64(983u64);
assert_eq!(term.field(), count_field);
assert_eq!(term.serialized_term().len(), FAST_VALUE_TERM_LEN);
}
#[test]
pub fn test_term_bool() {
let mut schema_builder = Schema::builder();
let bool_field = schema_builder.add_bool_field("bool", INDEXED);
let term = {
let mut term = IndexingTerm::with_capacity(0);
term.set_field(bool_field);
term.set_bool(true);
term
};
assert_eq!(term.field(), bool_field);
assert_eq!(term.serialized_term().len(), FAST_VALUE_TERM_LEN);
}
#[test]
pub fn indexing_term_wrap_extracts_field() {
let field = Field::from_field_id(7u32);
let mut term = IndexingTerm::with_capacity(0);
term.set_field(field);
term.append_bytes(b"abc");
let wrapped = IndexingTerm::wrap(term.serialized_term());
assert_eq!(wrapped.field(), field);
assert_eq!(wrapped.serialized_term(), term.serialized_term());
}
}

5
src/indexer/merge_index_test.rs
View File

@@ -104,9 +104,8 @@ mod tests {
let query = QueryParser::for_index(&index, vec![my_text_field])
.parse_query(term)
.unwrap();
let top_docs: Vec<(f32, DocAddress)> = searcher
.search(&query, &TopDocs::with_limit(3).order_by_score())
.unwrap();
let top_docs: Vec<(f32, DocAddress)> =
searcher.search(&query, &TopDocs::with_limit(3)).unwrap();
top_docs.iter().map(|el| el.1.doc_id).collect::<Vec<_>>()
};

9
src/indexer/merger.rs
View File

@@ -1518,8 +1518,7 @@ mod tests {
let searcher = reader.searcher();
let mut term_scorer = term_query
.specialized_weight(EnableScoring::enabled_from_searcher(&searcher))?
.term_scorer_for_test(searcher.segment_reader(0u32), 1.0)?
.unwrap();
.specialized_scorer(searcher.segment_reader(0u32), 1.0)?;
assert_eq!(term_scorer.doc(), 0);
assert_nearly_equals!(term_scorer.block_max_score(), 0.0079681855);
assert_nearly_equals!(term_scorer.score(), 0.0079681855);
@@ -1534,8 +1533,7 @@ mod tests {
for segment_reader in searcher.segment_readers() {
let mut term_scorer = term_query
.specialized_weight(EnableScoring::enabled_from_searcher(&searcher))?
.term_scorer_for_test(segment_reader, 1.0)?
.unwrap();
.specialized_scorer(segment_reader, 1.0)?;
// the difference compared to before is intrinsic to the bm25 formula. no worries
// there.
for doc in segment_reader.doc_ids_alive() {
@@ -1560,8 +1558,7 @@ mod tests {
let segment_reader = searcher.segment_reader(0u32);
let mut term_scorer = term_query
.specialized_weight(EnableScoring::enabled_from_searcher(&searcher))?
.term_scorer_for_test(segment_reader, 1.0)?
.unwrap();
.specialized_scorer(segment_reader, 1.0)?;
// the difference compared to before is intrinsic to the bm25 formula. no worries there.
for doc in segment_reader.doc_ids_alive() {
assert_eq!(term_scorer.doc(), doc);

Some files were not shown because too many files have changed in this diff Show More