rust/tantivy
1
0
Fork 0
mirror of https://github.com/quickwit-oss/tantivy.git synced 2026-01-03 15:52:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

Compare commits

base: rust:document-deserialize-seed
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:experimental-do-not-merge
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

2 Commits
document-d ... experiment

Author SHA1 Message Date
Paul Masurel
4e64801ec9 Experimental do not merge 2024-04-19 18:32:03 +09:00
Paul Masurel
34cb0c48bf Owned value relying on Vec instead of BTreeMap 2024-04-19 15:08:50 +09:00
89 changed files with 1757 additions and 2447 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-2024-04-10 --profile minimal --component llvm-tools-preview
run: rustup toolchain install nightly-2023-09-10 --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-2024-04-10 llvm-cov --all-features --workspace --doctests --lcov --output-path lcov.info
run: cargo +nightly-2023-09-10 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

62
CHANGELOG.md
View File

@@ -1,65 +1,3 @@
Tantivy 0.22
================================
Tantivy 0.22 will be able to read indices created with Tantivy 0.21.
#### Bugfixes
- Fix null byte handling in JSON paths (null bytes in json keys caused panic during indexing) [#2345](https://github.com/quickwit-oss/tantivy/pull/2345)(@PSeitz)
- Fix bug that can cause `get_docids_for_value_range` to panic. [#2295](https://github.com/quickwit-oss/tantivy/pull/2295)(@fulmicoton)
- Avoid 1 document indices by increase min memory to 15MB for indexing [#2176](https://github.com/quickwit-oss/tantivy/pull/2176)(@PSeitz)
- Fix merge panic for JSON fields [#2284](https://github.com/quickwit-oss/tantivy/pull/2284)(@PSeitz)
- Fix bug occuring when merging JSON object indexed with positions. [#2253](https://github.com/quickwit-oss/tantivy/pull/2253)(@fulmicoton)
- Fix empty DateHistogram gap bug [#2183](https://github.com/quickwit-oss/tantivy/pull/2183)(@PSeitz)
- Fix range query end check (fields with less than 1 value per doc are affected) [#2226](https://github.com/quickwit-oss/tantivy/pull/2226)(@PSeitz)
- Handle exclusive out of bounds ranges on fastfield range queries [#2174](https://github.com/quickwit-oss/tantivy/pull/2174)(@PSeitz)
#### Breaking API Changes
- rename ReloadPolicy onCommit to onCommitWithDelay [#2235](https://github.com/quickwit-oss/tantivy/pull/2235)(@giovannicuccu)
- Move exports from the root into modules [#2220](https://github.com/quickwit-oss/tantivy/pull/2220)(@PSeitz)
- Accept field name instead of `Field` in FilterCollector [#2196](https://github.com/quickwit-oss/tantivy/pull/2196)(@PSeitz)
- remove deprecated IntOptions and DateTime [#2353](https://github.com/quickwit-oss/tantivy/pull/2353)(@PSeitz)
#### Features/Improvements
- Tantivy documents as a trait: Index data directly without converting to tantivy types first [#2071](https://github.com/quickwit-oss/tantivy/pull/2071)(@ChillFish8)
- encode some part of posting list as -1 instead of direct values (smaller inverted indices) [#2185](https://github.com/quickwit-oss/tantivy/pull/2185)(@trinity-1686a)
- **Aggregation**
- Support to deserialize f64 from string [#2311](https://github.com/quickwit-oss/tantivy/pull/2311)(@PSeitz)
- Add a top_hits aggregator [#2198](https://github.com/quickwit-oss/tantivy/pull/2198)(@ditsuke)
- Support bool type in term aggregation [#2318](https://github.com/quickwit-oss/tantivy/pull/2318)(@PSeitz)
- Support ip adresses in term aggregation [#2319](https://github.com/quickwit-oss/tantivy/pull/2319)(@PSeitz)
- Support date type in term aggregation [#2172](https://github.com/quickwit-oss/tantivy/pull/2172)(@PSeitz)
- Support escaped dot when addressing field [#2250](https://github.com/quickwit-oss/tantivy/pull/2250)(@PSeitz)
- Add ExistsQuery to check documents that have a value [#2160](https://github.com/quickwit-oss/tantivy/pull/2160)(@imotov)
- Expose TopDocs::order_by_u64_field again [#2282](https://github.com/quickwit-oss/tantivy/pull/2282)(@ditsuke)
- **Memory/Performance**
- Faster TopN: replace BinaryHeap with TopNComputer [#2186](https://github.com/quickwit-oss/tantivy/pull/2186)(@PSeitz)
- reduce number of allocations during indexing [#2257](https://github.com/quickwit-oss/tantivy/pull/2257)(@PSeitz)
- Less Memory while indexing: docid deltas while indexing [#2249](https://github.com/quickwit-oss/tantivy/pull/2249)(@PSeitz)
- Faster indexing: use term hashmap in fastfield [#2243](https://github.com/quickwit-oss/tantivy/pull/2243)(@PSeitz)
- term hashmap remove copy in is_empty, unused unordered_id [#2229](https://github.com/quickwit-oss/tantivy/pull/2229)(@PSeitz)
- add method to fetch block of first values in columnar [#2330](https://github.com/quickwit-oss/tantivy/pull/2330)(@PSeitz)
- Faster aggregations: add fast path for full columns in fetch_block [#2328](https://github.com/quickwit-oss/tantivy/pull/2328)(@PSeitz)
- Faster sstable loading: use fst for sstable index [#2268](https://github.com/quickwit-oss/tantivy/pull/2268)(@trinity-1686a)
- **QueryParser**
- allow newline where we allow space in query parser [#2302](https://github.com/quickwit-oss/tantivy/pull/2302)(@trinity-1686a)
- allow some mixing of occur and bool in strict query parser [#2323](https://github.com/quickwit-oss/tantivy/pull/2323)(@trinity-1686a)
- handle * inside term in lenient query parser [#2228](https://github.com/quickwit-oss/tantivy/pull/2228)(@trinity-1686a)
- add support for exists query syntax in query parser [#2170](https://github.com/quickwit-oss/tantivy/pull/2170)(@trinity-1686a)
- Add shared search executor [#2312](https://github.com/quickwit-oss/tantivy/pull/2312)(@MochiXu)
- Truncate keys to u16::MAX in term hashmap [#2299](https://github.com/quickwit-oss/tantivy/pull/2299)(@PSeitz)
- report if a term matched when warming up posting list [#2309](https://github.com/quickwit-oss/tantivy/pull/2309)(@trinity-1686a)
- Support json fields in FuzzyTermQuery [#2173](https://github.com/quickwit-oss/tantivy/pull/2173)(@PingXia-at)
- Read list of fields encoded in term dictionary for JSON fields [#2184](https://github.com/quickwit-oss/tantivy/pull/2184)(@PSeitz)
- add collect_block to BoxableSegmentCollector [#2331](https://github.com/quickwit-oss/tantivy/pull/2331)(@PSeitz)
- expose collect_block buffer size [#2326](https://github.com/quickwit-oss/tantivy/pull/2326)(@PSeitz)
- Forward regex parser errors [#2288](https://github.com/quickwit-oss/tantivy/pull/2288)(@adamreichold)
- Make FacetCounts defaultable and cloneable. [#2322](https://github.com/quickwit-oss/tantivy/pull/2322)(@adamreichold)
- Derive Debug for SchemaBuilder [#2254](https://github.com/quickwit-oss/tantivy/pull/2254)(@GodTamIt)
- add missing inlines to tantivy options [#2245](https://github.com/quickwit-oss/tantivy/pull/2245)(@PSeitz)
Tantivy 0.21.1
================================
#### Bugfixes

51
Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy"
version = "0.23.0"
version = "0.22.0-dev"
authors = ["Paul Masurel <paul.masurel@gmail.com>"]
license = "MIT"
categories = ["database-implementations", "data-structures"]
@@ -15,16 +15,12 @@ rust-version = "1.63"
exclude = ["benches/*.json", "benches/*.txt"]
[dependencies]
# Switch back to the non-forked oneshot crate once https://github.com/faern/oneshot/pull/35 is merged
oneshot = { git = "https://github.com/fulmicoton/oneshot.git", rev = "c10a3ba" }
oneshot = "0.1.5"
base64 = "0.22.0"
byteorder = "1.4.3"
crc32fast = "1.3.2"
once_cell = "1.10.0"
regex = { version = "1.5.5", default-features = false, features = [
"std",
"unicode",
] }
regex = { version = "1.5.5", default-features = false, features = ["std", "unicode"] }
aho-corasick = "1.0"
tantivy-fst = "0.5"
memmap2 = { version = "0.9.0", optional = true }
@@ -34,15 +30,14 @@ tempfile = { version = "3.3.0", optional = true }
log = "0.4.16"
serde = { version = "1.0.136", features = ["derive"] }
serde_json = "1.0.79"
num_cpus = "1.13.1"
fs4 = { version = "0.8.0", optional = true }
levenshtein_automata = "0.2.1"
uuid = { version = "1.0.0", features = ["v4", "serde"] }
crossbeam-channel = "0.5.4"
rust-stemmers = "1.2.0"
downcast-rs = "1.2.0"
bitpacking = { version = "0.9.2", default-features = false, features = [
"bitpacker4x",
] }
bitpacking = { version = "0.9.2", default-features = false, features = ["bitpacker4x"] }
census = "0.4.2"
rustc-hash = "1.1.0"
thiserror = "1.0.30"
@@ -53,17 +48,17 @@ smallvec = "1.8.0"
rayon = "1.5.2"
lru = "0.12.0"
fastdivide = "0.4.0"
itertools = "0.13.0"
itertools = "0.12.0"
measure_time = "0.8.2"
arc-swap = "1.5.0"
columnar = { version = "0.3", path = "./columnar", package = "tantivy-columnar" }
sstable = { version = "0.3", path = "./sstable", package = "tantivy-sstable", optional = true }
stacker = { version = "0.3", path = "./stacker", package = "tantivy-stacker" }
query-grammar = { version = "0.22.0", path = "./query-grammar", package = "tantivy-query-grammar" }
tantivy-bitpacker = { version = "0.6", path = "./bitpacker" }
common = { version = "0.7", path = "./common/", package = "tantivy-common" }
tokenizer-api = { version = "0.3", path = "./tokenizer-api", package = "tantivy-tokenizer-api" }
columnar = { version= "0.2", path="./columnar", package ="tantivy-columnar" }
sstable = { version= "0.2", path="./sstable", package ="tantivy-sstable", optional = true }
stacker = { version= "0.2", path="./stacker", package ="tantivy-stacker" }
query-grammar = { version= "0.21.0", path="./query-grammar", package = "tantivy-query-grammar" }
tantivy-bitpacker = { version= "0.5", path="./bitpacker" }
common = { version= "0.6", path = "./common/", package = "tantivy-common" }
tokenizer-api = { version= "0.2", path="./tokenizer-api", package="tantivy-tokenizer-api" }
sketches-ddsketch = { version = "0.2.1", features = ["use_serde"] }
futures-util = { version = "0.3.28", optional = true }
fnv = "1.0.7"
@@ -72,7 +67,6 @@ fnv = "1.0.7"
winapi = "0.3.9"
[dev-dependencies]
binggan = "0.8.0"
rand = "0.8.5"
maplit = "1.0.2"
matches = "0.1.9"
@@ -118,26 +112,17 @@ lz4-compression = ["lz4_flex"]
zstd-compression = ["zstd"]
failpoints = ["fail", "fail/failpoints"]
unstable = [] # useful for benches.
unstable = [] # useful for benches.
quickwit = ["sstable", "futures-util"]
# Compares only the hash of a string when indexing data.
# Compares only the hash of a string when indexing data.
# Increases indexing speed, but may lead to extremely rare missing terms, when there's a hash collision.
# Uses 64bit ahash.
compare_hash_only = ["stacker/compare_hash_only"]
[workspace]
members = [
"query-grammar",
"bitpacker",
"common",
"ownedbytes",
"stacker",
"sstable",
"tokenizer-api",
"columnar",
]
members = ["query-grammar", "bitpacker", "common", "ownedbytes", "stacker", "sstable", "tokenizer-api", "columnar"]
# Following the "fail" crate best practises, we isolate
# tests that define specific behavior in fail check points
@@ -158,7 +143,3 @@ harness = false
[[bench]]
name = "index-bench"
harness = false
[[bench]]
name = "agg_bench"
harness = false

413
benches/agg_bench.rs
View File

@@ -1,413 +0,0 @@
use binggan::{black_box, InputGroup, PeakMemAlloc, INSTRUMENTED_SYSTEM};
use rand::prelude::SliceRandom;
use rand::rngs::StdRng;
use rand::{Rng, SeedableRng};
use rand_distr::Distribution;
use serde_json::json;
use tantivy::aggregation::agg_req::Aggregations;
use tantivy::aggregation::AggregationCollector;
use tantivy::query::{AllQuery, TermQuery};
use tantivy::schema::{IndexRecordOption, Schema, TextFieldIndexing, FAST, STRING};
use tantivy::{doc, Index, Term};
#[global_allocator]
pub static GLOBAL: &PeakMemAlloc<std::alloc::System> = &INSTRUMENTED_SYSTEM;
/// Mini macro to register a function via its name
/// runner.register("average_u64", move |index| average_u64(index));
macro_rules! register {
($runner:expr, $func:ident) => {
$runner.register(stringify!($func), move |index| $func(index))
};
}
fn main() {
let inputs = vec![
("full", get_test_index_bench(Cardinality::Full).unwrap()),
(
"dense",
get_test_index_bench(Cardinality::OptionalDense).unwrap(),
),
(
"sparse",
get_test_index_bench(Cardinality::OptionalSparse).unwrap(),
),
(
"multivalue",
get_test_index_bench(Cardinality::Multivalued).unwrap(),
),
];
bench_agg(InputGroup::new_with_inputs(inputs));
}
fn bench_agg(mut group: InputGroup<Index>) {
group.set_alloc(GLOBAL); // Set the peak mem allocator. This will enable peak memory reporting.
register!(group, average_u64);
register!(group, average_f64);
register!(group, average_f64_u64);
register!(group, stats_f64);
register!(group, percentiles_f64);
register!(group, terms_few);
register!(group, terms_many);
register!(group, terms_many_order_by_term);
register!(group, terms_many_with_top_hits);
register!(group, terms_many_with_avg_sub_agg);
register!(group, terms_many_json_mixed_type_with_sub_agg_card);
register!(group, range_agg);
register!(group, range_agg_with_avg_sub_agg);
register!(group, range_agg_with_term_agg_few);
register!(group, range_agg_with_term_agg_many);
register!(group, histogram);
register!(group, histogram_hard_bounds);
register!(group, histogram_with_avg_sub_agg);
register!(group, avg_and_range_with_avg_sub_agg);
group.run();
}
fn exec_term_with_agg(index: &Index, agg_req: serde_json::Value) {
let agg_req: Aggregations = serde_json::from_value(agg_req).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let collector = get_collector(agg_req);
let searcher = reader.searcher();
black_box(searcher.search(&term_query, &collector).unwrap());
}
fn average_u64(index: &Index) {
let agg_req = json!({
"average": { "avg": { "field": "score", } }
});
exec_term_with_agg(index, agg_req)
}
fn average_f64(index: &Index) {
let agg_req = json!({
"average": { "avg": { "field": "score_f64", } }
});
exec_term_with_agg(index, agg_req)
}
fn average_f64_u64(index: &Index) {
let agg_req = json!({
"average_f64": { "avg": { "field": "score_f64" } },
"average": { "avg": { "field": "score" } },
});
exec_term_with_agg(index, agg_req)
}
fn stats_f64(index: &Index) {
let agg_req = json!({
"average_f64": { "stats": { "field": "score_f64", } }
});
exec_term_with_agg(index, agg_req)
}
fn percentiles_f64(index: &Index) {
let agg_req = json!({
"mypercentiles": {
"percentiles": {
"field": "score_f64",
"percents": [ 95, 99, 99.9 ]
}
}
});
execute_agg(index, agg_req);
}
fn terms_few(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_few_terms" } },
});
execute_agg(index, agg_req);
}
fn terms_many(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_many_terms" } },
});
execute_agg(index, agg_req);
}
fn terms_many_order_by_term(index: &Index) {
let agg_req = json!({
"my_texts": { "terms": { "field": "text_many_terms", "order": { "_key": "desc" } } },
});
execute_agg(index, agg_req);
}
fn terms_many_with_top_hits(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_many_terms" },
"aggs": {
"top_hits": { "top_hits":
{
"sort": [
{ "score": "desc" }
],
"size": 2,
"doc_value_fields": ["score_f64"]
}
}
}
},
});
execute_agg(index, agg_req);
}
fn terms_many_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "text_many_terms" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn terms_many_json_mixed_type_with_sub_agg_card(index: &Index) {
let agg_req = json!({
"my_texts": {
"terms": { "field": "json.mixed_type" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn execute_agg(index: &Index, agg_req: serde_json::Value) {
let agg_req: Aggregations = serde_json::from_value(agg_req).unwrap();
let collector = get_collector(agg_req);
let reader = index.reader().unwrap();
let searcher = reader.searcher();
black_box(searcher.search(&AllQuery, &collector).unwrap());
}
fn range_agg(index: &Index) {
let agg_req = json!({
"range_f64": { "range": { "field": "score_f64", "ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
] } },
});
execute_agg(index, agg_req);
}
fn range_agg_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
]
},
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
});
execute_agg(index, agg_req);
}
fn range_agg_with_term_agg_few(index: &Index) {
let agg_req = json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
]
},
"aggs": {
"my_texts": { "terms": { "field": "text_few_terms" } },
}
},
});
execute_agg(index, agg_req);
}
fn range_agg_with_term_agg_many(index: &Index) {
let agg_req = json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
]
},
"aggs": {
"my_texts": { "terms": { "field": "text_many_terms" } },
}
},
});
execute_agg(index, agg_req);
}
fn histogram(index: &Index) {
let agg_req = json!({
"rangef64": {
"histogram": {
"field": "score_f64",
"interval": 100 // 1000 buckets
},
}
});
execute_agg(index, agg_req);
}
fn histogram_hard_bounds(index: &Index) {
let agg_req = json!({
"rangef64": { "histogram": { "field": "score_f64", "interval": 100, "hard_bounds": { "min": 1000, "max": 300000 } } },
});
execute_agg(index, agg_req);
}
fn histogram_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"rangef64": {
"histogram": { "field": "score_f64", "interval": 100 },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
}
});
execute_agg(index, agg_req);
}
fn avg_and_range_with_avg_sub_agg(index: &Index) {
let agg_req = json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 60000 }
]
},
"aggs": {
"average_in_range": { "avg": { "field": "score" } }
}
},
"average": { "avg": { "field": "score" } }
});
execute_agg(index, agg_req);
}
#[derive(Clone, Copy, Hash, Default, Debug, PartialEq, Eq, PartialOrd, Ord)]
enum Cardinality {
/// All documents contain exactly one value.
/// `Full` is the default for auto-detecting the Cardinality, since it is the most strict.
#[default]
Full = 0,
/// All documents contain at most one value.
OptionalDense = 1,
/// All documents may contain any number of values.
Multivalued = 2,
/// 1 / 20 documents has a value
OptionalSparse = 3,
}
fn get_collector(agg_req: Aggregations) -> AggregationCollector {
AggregationCollector::from_aggs(agg_req, Default::default())
}
fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
let mut schema_builder = Schema::builder();
let text_fieldtype = tantivy::schema::TextOptions::default()
.set_indexing_options(
TextFieldIndexing::default().set_index_option(IndexRecordOption::WithFreqs),
)
.set_stored();
let text_field = schema_builder.add_text_field("text", text_fieldtype);
let json_field = schema_builder.add_json_field("json", 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 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"];
let lg_norm = rand_distr::LogNormal::new(2.996f64, 0.979f64).unwrap();
let many_terms_data = (0..150_000)
.map(|num| format!("author{}", num))
.collect::<Vec<_>>();
{
let mut rng = StdRng::from_seed([1u8; 32]);
let mut index_writer = index.writer_with_num_threads(1, 200_000_000)?;
// To make the different test cases comparable we just change one doc to force the
// cardinality
if cardinality == Cardinality::OptionalDense {
index_writer.add_document(doc!())?;
}
if cardinality == Cardinality::Multivalued {
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_many_terms => "cool",
text_field_many_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms => "cool",
score_field => 1u64,
score_field => 1u64,
score_field_f64 => lg_norm.sample(&mut rng),
score_field_f64 => lg_norm.sample(&mut rng),
score_field_i64 => 1i64,
score_field_i64 => 1i64,
))?;
}
let mut doc_with_value = 1_000_000;
if cardinality == Cardinality::OptionalSparse {
doc_with_value /= 20;
}
let _val_max = 1_000_000.0;
for _ in 0..doc_with_value {
let val: f64 = rng.gen_range(0.0..1_000_000.0);
let json = if rng.gen_bool(0.1) {
// 10% are numeric values
json!({ "mixed_type": val })
} else {
json!({"mixed_type": many_terms_data.choose(&mut rng).unwrap().to_string()})
};
index_writer.add_document(doc!(
text_field => "cool",
json_field => json,
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(),
score_field => val as u64,
score_field_f64 => lg_norm.sample(&mut rng),
score_field_i64 => val as i64,
))?;
if cardinality == Cardinality::OptionalSparse {
for _ in 0..20 {
index_writer.add_document(doc!(text_field => "cool"))?;
}
}
}
// writing the segment
index_writer.commit()?;
}
Ok(index)
}

10
benches/index-bench.rs
View File

@@ -18,7 +18,7 @@ fn benchmark(
benchmark_dynamic_json(b, input, schema, commit, parse_json)
} else {
_benchmark(b, input, schema, commit, parse_json, |schema, doc_json| {
TantivyDocument::parse_json(schema, doc_json).unwrap()
TantivyDocument::parse_json(&schema, doc_json).unwrap()
})
}
}
@@ -90,7 +90,8 @@ fn benchmark_dynamic_json(
) {
let json_field = schema.get_field("json").unwrap();
_benchmark(b, input, schema, commit, parse_json, |_schema, doc_json| {
let json_val: serde_json::Value = serde_json::from_str(doc_json).unwrap();
let json_val: serde_json::Map<String, serde_json::Value> =
serde_json::from_str(doc_json).unwrap();
tantivy::doc!(json_field=>json_val)
})
}
@@ -137,13 +138,12 @@ pub fn hdfs_index_benchmark(c: &mut Criterion) {
for (prefix, schema, is_dynamic) in benches {
for commit in [false, true] {
let suffix = if commit { "with-commit" } else { "no-commit" };
{
let parse_json = false;
for parse_json in [false] {
// for parse_json in [false, true] {
let suffix = if parse_json {
format!("{}-with-json-parsing", suffix)
} else {
suffix.to_string()
format!("{}", suffix)
};
let bench_name = format!("{}{}", prefix, suffix);

2
bitpacker/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-bitpacker"
version = "0.6.0"
version = "0.5.0"
edition = "2021"
authors = ["Paul Masurel <paul.masurel@gmail.com>"]
license = "MIT"

83
cliff.toml
View File

@@ -1,10 +1,6 @@
# configuration file for git-cliff{ pattern = "foo", replace = "bar"}
# see https://github.com/orhun/git-cliff#configuration-file
[remote.github]
owner = "quickwit-oss"
repo = "tantivy"
[changelog]
# changelog header
header = """
@@ -12,43 +8,15 @@ header = """
# template for the changelog body
# https://tera.netlify.app/docs/#introduction
body = """
## What's Changed
{%- if version %} in {{ version }}{%- endif -%}
{% if version %}\
{{ version | trim_start_matches(pat="v") }} ({{ timestamp | date(format="%Y-%m-%d") }})
==================
{% else %}\
## [unreleased]
{% endif %}\
{% for commit in commits %}
{% if commit.github.pr_title -%}
{%- set commit_message = commit.github.pr_title -%}
{%- else -%}
{%- set commit_message = commit.message -%}
{%- endif -%}
- {{ commit_message | split(pat="\n") | first | trim }}\
{% if commit.github.pr_number %} \
[#{{ commit.github.pr_number }}]({{ self::remote_url() }}/pull/{{ commit.github.pr_number }}){% if commit.github.username %}(@{{ commit.github.username }}){%- endif -%} \
{%- endif %}
{%- endfor -%}
{% if github.contributors | filter(attribute="is_first_time", value=true) | length != 0 %}
{% raw %}\n{% endraw -%}
## New Contributors
{%- endif %}\
{% for contributor in github.contributors | filter(attribute="is_first_time", value=true) %}
* @{{ contributor.username }} made their first contribution
{%- if contributor.pr_number %} in \
[#{{ contributor.pr_number }}]({{ self::remote_url() }}/pull/{{ contributor.pr_number }}) \
{%- endif %}
{%- endfor -%}
{% if version %}
{% if previous.version %}
**Full Changelog**: {{ self::remote_url() }}/compare/{{ previous.version }}...{{ version }}
{% endif %}
{% else -%}
{% raw %}\n{% endraw %}
{% endif %}
{%- macro remote_url() -%}
https://github.com/{{ remote.github.owner }}/{{ remote.github.repo }}
{%- endmacro -%}
- {% if commit.breaking %}[**breaking**] {% endif %}{{ commit.message | split(pat="\n") | first | trim | upper_first }}(@{{ commit.author.name }})\
{% endfor %}
"""
# remove the leading and trailing whitespace from the template
trim = true
@@ -57,24 +25,53 @@ footer = """
"""
postprocessors = [
{ pattern = 'Paul Masurel', replace = "fulmicoton"}, # replace with github user
{ pattern = 'PSeitz', replace = "PSeitz"}, # replace with github user
{ pattern = 'Adam Reichold', replace = "adamreichold"}, # replace with github user
{ pattern = 'trinity-1686a', replace = "trinity-1686a"}, # replace with github user
{ pattern = 'Michael Kleen', replace = "mkleen"}, # replace with github user
{ pattern = 'Adrien Guillo', replace = "guilload"}, # replace with github user
{ pattern = 'François Massot', replace = "fmassot"}, # replace with github user
{ pattern = 'Naveen Aiathurai', replace = "naveenann"}, # replace with github user
{ pattern = '', replace = ""}, # replace with github user
]
[git]
# parse the commits based on https://www.conventionalcommits.org
# This is required or commit.message contains the whole commit message and not just the title
conventional_commits = false
conventional_commits = true
# filter out the commits that are not conventional
filter_unconventional = true
filter_unconventional = false
# process each line of a commit as an individual commit
split_commits = false
# regex for preprocessing the commit messages
commit_preprocessors = [
{ pattern = '\((\w+\s)?#([0-9]+)\)', replace = ""},
{ pattern = '\((\w+\s)?#([0-9]+)\)', replace = "[#${2}](https://github.com/quickwit-oss/tantivy/issues/${2})"}, # replace issue numbers
]
#link_parsers = [
#{ pattern = "#(\\d+)", href = "https://github.com/quickwit-oss/tantivy/pulls/$1"},
#]
# regex for parsing and grouping commits
commit_parsers = [
{ message = "^feat", group = "Features"},
{ message = "^fix", group = "Bug Fixes"},
{ message = "^doc", group = "Documentation"},
{ message = "^perf", group = "Performance"},
{ message = "^refactor", group = "Refactor"},
{ message = "^style", group = "Styling"},
{ message = "^test", group = "Testing"},
{ message = "^chore\\(release\\): prepare for", skip = true},
{ message = "(?i)clippy", skip = true},
{ message = "(?i)dependabot", skip = true},
{ message = "(?i)fmt", skip = true},
{ message = "(?i)bump", skip = true},
{ message = "(?i)readme", skip = true},
{ message = "(?i)comment", skip = true},
{ message = "(?i)spelling", skip = true},
{ message = "^chore", group = "Miscellaneous Tasks"},
{ body = ".*security", group = "Security"},
{ message = ".*", group = "Other", default_scope = "other"},
]
# protect breaking changes from being skipped due to matching a skipping commit_parser
protect_breaking_commits = false
# filter out the commits that are not matched by commit parsers

12
columnar/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-columnar"
version = "0.3.0"
version = "0.2.0"
edition = "2021"
license = "MIT"
homepage = "https://github.com/quickwit-oss/tantivy"
@@ -9,13 +9,13 @@ description = "column oriented storage for tantivy"
categories = ["database-implementations", "data-structures", "compression"]
[dependencies]
itertools = "0.13.0"
itertools = "0.12.0"
fastdivide = "0.4.0"
stacker = { version= "0.3", path = "../stacker", package="tantivy-stacker"}
sstable = { version= "0.3", path = "../sstable", package = "tantivy-sstable" }
common = { version= "0.7", path = "../common", package = "tantivy-common" }
tantivy-bitpacker = { version= "0.6", path = "../bitpacker/" }
stacker = { version= "0.2", path = "../stacker", package="tantivy-stacker"}
sstable = { version= "0.2", path = "../sstable", package = "tantivy-sstable" }
common = { version= "0.6", path = "../common", package = "tantivy-common" }
tantivy-bitpacker = { version= "0.5", path = "../bitpacker/" }
serde = "1.0.152"
downcast-rs = "1.2.0"

88
columnar/src/columnar/writer/mod.rs
View File

@@ -59,6 +59,22 @@ pub struct ColumnarWriter {
buffers: SpareBuffers,
}
#[inline]
fn mutate_or_create_column<V, TMutator>(
arena_hash_map: &mut ArenaHashMap,
column_name: &str,
updater: TMutator,
) where
V: Copy + 'static,
TMutator: FnMut(Option<V>) -> V,
{
assert!(
!column_name.as_bytes().contains(&0u8),
"key may not contain the 0 byte"
);
arena_hash_map.mutate_or_create(column_name.as_bytes(), updater);
}
impl ColumnarWriter {
pub fn mem_usage(&self) -> usize {
self.arena.mem_usage()
@@ -159,8 +175,9 @@ impl ColumnarWriter {
},
&mut self.dictionaries,
);
hash_map.mutate_or_create(
column_name.as_bytes(),
mutate_or_create_column(
hash_map,
column_name,
|column_opt: Option<StrOrBytesColumnWriter>| {
let mut column_writer = if let Some(column_writer) = column_opt {
column_writer
@@ -175,21 +192,24 @@ impl ColumnarWriter {
);
}
ColumnType::Bool => {
self.bool_field_hash_map.mutate_or_create(
column_name.as_bytes(),
mutate_or_create_column(
&mut self.bool_field_hash_map,
column_name,
|column_opt: Option<ColumnWriter>| column_opt.unwrap_or_default(),
);
}
ColumnType::DateTime => {
self.datetime_field_hash_map.mutate_or_create(
column_name.as_bytes(),
mutate_or_create_column(
&mut self.datetime_field_hash_map,
column_name,
|column_opt: Option<ColumnWriter>| column_opt.unwrap_or_default(),
);
}
ColumnType::I64 | ColumnType::F64 | ColumnType::U64 => {
let numerical_type = column_type.numerical_type().unwrap();
self.numerical_field_hash_map.mutate_or_create(
column_name.as_bytes(),
mutate_or_create_column(
&mut self.numerical_field_hash_map,
column_name,
|column_opt: Option<NumericalColumnWriter>| {
let mut column: NumericalColumnWriter = column_opt.unwrap_or_default();
column.force_numerical_type(numerical_type);
@@ -197,8 +217,9 @@ impl ColumnarWriter {
},
);
}
ColumnType::IpAddr => self.ip_addr_field_hash_map.mutate_or_create(
column_name.as_bytes(),
ColumnType::IpAddr => mutate_or_create_column(
&mut self.ip_addr_field_hash_map,
column_name,
|column_opt: Option<ColumnWriter>| column_opt.unwrap_or_default(),
),
}
@@ -211,8 +232,9 @@ impl ColumnarWriter {
numerical_value: T,
) {
let (hash_map, arena) = (&mut self.numerical_field_hash_map, &mut self.arena);
hash_map.mutate_or_create(
column_name.as_bytes(),
mutate_or_create_column(
hash_map,
column_name,
|column_opt: Option<NumericalColumnWriter>| {
let mut column: NumericalColumnWriter = column_opt.unwrap_or_default();
column.record_numerical_value(doc, numerical_value.into(), arena);
@@ -222,6 +244,10 @@ impl ColumnarWriter {
}
pub fn record_ip_addr(&mut self, doc: RowId, column_name: &str, ip_addr: Ipv6Addr) {
assert!(
!column_name.as_bytes().contains(&0u8),
"key may not contain the 0 byte"
);
let (hash_map, arena) = (&mut self.ip_addr_field_hash_map, &mut self.arena);
hash_map.mutate_or_create(
column_name.as_bytes(),
@@ -235,30 +261,24 @@ impl ColumnarWriter {
pub fn record_bool(&mut self, doc: RowId, column_name: &str, val: bool) {
let (hash_map, arena) = (&mut self.bool_field_hash_map, &mut self.arena);
hash_map.mutate_or_create(
column_name.as_bytes(),
|column_opt: Option<ColumnWriter>| {
let mut column: ColumnWriter = column_opt.unwrap_or_default();
column.record(doc, val, arena);
column
},
);
mutate_or_create_column(hash_map, column_name, |column_opt: Option<ColumnWriter>| {
let mut column: ColumnWriter = column_opt.unwrap_or_default();
column.record(doc, val, arena);
column
});
}
pub fn record_datetime(&mut self, doc: RowId, column_name: &str, datetime: common::DateTime) {
let (hash_map, arena) = (&mut self.datetime_field_hash_map, &mut self.arena);
hash_map.mutate_or_create(
column_name.as_bytes(),
|column_opt: Option<ColumnWriter>| {
let mut column: ColumnWriter = column_opt.unwrap_or_default();
column.record(
doc,
NumericalValue::I64(datetime.into_timestamp_nanos()),
arena,
);
column
},
);
mutate_or_create_column(hash_map, column_name, |column_opt: Option<ColumnWriter>| {
let mut column: ColumnWriter = column_opt.unwrap_or_default();
column.record(
doc,
NumericalValue::I64(datetime.into_timestamp_nanos()),
arena,
);
column
});
}
pub fn record_str(&mut self, doc: RowId, column_name: &str, value: &str) {
@@ -283,6 +303,10 @@ impl ColumnarWriter {
}
pub fn record_bytes(&mut self, doc: RowId, column_name: &str, value: &[u8]) {
assert!(
!column_name.as_bytes().contains(&0u8),
"key may not contain the 0 byte"
);
let (hash_map, arena, dictionaries) = (
&mut self.bytes_field_hash_map,
&mut self.arena,

4
common/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-common"
version = "0.7.0"
version = "0.6.0"
authors = ["Paul Masurel <paul@quickwit.io>", "Pascal Seitz <pascal@quickwit.io>"]
license = "MIT"
edition = "2021"
@@ -14,7 +14,7 @@ repository = "https://github.com/quickwit-oss/tantivy"
[dependencies]
byteorder = "1.4.3"
ownedbytes = { version= "0.7", path="../ownedbytes" }
ownedbytes = { version= "0.6", path="../ownedbytes" }
async-trait = "0.1"
time = { version = "0.3.10", features = ["serde-well-known"] }
serde = { version = "1.0.136", features = ["derive"] }

2
common/src/bitset.rs
View File

@@ -1,5 +1,5 @@
use std::io::Write;
use std::{fmt, io};
use std::{fmt, io, u64};
use ownedbytes::OwnedBytes;

5
common/src/datetime.rs
View File

@@ -1,3 +1,5 @@
#![allow(deprecated)]
use std::fmt;
use std::io::{Read, Write};
@@ -25,6 +27,9 @@ pub enum DateTimePrecision {
Nanoseconds,
}
#[deprecated(since = "0.20.0", note = "Use `DateTimePrecision` instead")]
pub type DatePrecision = DateTimePrecision;
/// A date/time value with nanoseconds precision.
///
/// This timestamp does not carry any explicit time zone information.

36
common/src/json_path_writer.rs
View File

@@ -5,12 +5,6 @@ pub const JSON_PATH_SEGMENT_SEP: u8 = 1u8;
pub const JSON_PATH_SEGMENT_SEP_STR: &str =
unsafe { std::str::from_utf8_unchecked(&[JSON_PATH_SEGMENT_SEP]) };
/// Separates the json path and the value in
/// a JSON term binary representation.
pub const JSON_END_OF_PATH: u8 = 0u8;
pub const JSON_END_OF_PATH_STR: &str =
unsafe { std::str::from_utf8_unchecked(&[JSON_END_OF_PATH]) };
/// Create a new JsonPathWriter, that creates flattened json paths for tantivy.
#[derive(Clone, Debug, Default)]
pub struct JsonPathWriter {
@@ -20,14 +14,6 @@ pub struct JsonPathWriter {
}
impl JsonPathWriter {
pub fn with_expand_dots(expand_dots: bool) -> Self {
JsonPathWriter {
path: String::new(),
indices: Vec::new(),
expand_dots,
}
}
pub fn new() -> Self {
JsonPathWriter {
path: String::new(),
@@ -53,8 +39,8 @@ impl JsonPathWriter {
pub fn push(&mut self, segment: &str) {
let len_path = self.path.len();
self.indices.push(len_path);
if self.indices.len() > 1 {
self.path.push(JSON_PATH_SEGMENT_SEP as char);
if !self.path.is_empty() {
self.path.push_str(JSON_PATH_SEGMENT_SEP_STR);
}
self.path.push_str(segment);
if self.expand_dots {
@@ -69,12 +55,6 @@ impl JsonPathWriter {
}
}
/// Set the end of JSON path marker.
#[inline]
pub fn set_end(&mut self) {
self.path.push_str(JSON_END_OF_PATH_STR);
}
/// Remove the last segment. Does nothing if the path is empty.
#[inline]
pub fn pop(&mut self) {
@@ -111,7 +91,6 @@ mod tests {
#[test]
fn json_path_writer_test() {
let mut writer = JsonPathWriter::new();
writer.set_expand_dots(false);
writer.push("root");
assert_eq!(writer.as_str(), "root");
@@ -130,15 +109,4 @@ mod tests {
writer.push("k8s.node.id");
assert_eq!(writer.as_str(), "root\u{1}k8s\u{1}node\u{1}id");
}
#[test]
fn test_json_path_expand_dots_enabled_pop_segment() {
let mut json_writer = JsonPathWriter::with_expand_dots(true);
json_writer.push("hello");
assert_eq!(json_writer.as_str(), "hello");
json_writer.push("color.hue");
assert_eq!(json_writer.as_str(), "hello\x01color\x01hue");
json_writer.pop();
assert_eq!(json_writer.as_str(), "hello");
}
}

4
common/src/lib.rs
View File

@@ -9,12 +9,14 @@ mod byte_count;
mod datetime;
pub mod file_slice;
mod group_by;
pub mod json_path_writer;
mod json_path_writer;
mod serialize;
mod vint;
mod writer;
pub use bitset::*;
pub use byte_count::ByteCount;
#[allow(deprecated)]
pub use datetime::DatePrecision;
pub use datetime::{DateTime, DateTimePrecision};
pub use group_by::GroupByIteratorExtended;
pub use json_path_writer::JsonPathWriter;

2
common/src/vint.rs
View File

@@ -151,7 +151,7 @@ pub fn read_u32_vint_no_advance(data: &[u8]) -> (u32, usize) {
(result, vlen)
}
/// Write a `u32` as a vint payload.
pub fn write_u32_vint<W: io::Write + ?Sized>(val: u32, writer: &mut W) -> io::Result<()> {
pub fn write_u32_vint<W: io::Write>(val: u32, writer: &mut W) -> io::Result<()> {
let mut buf = [0u8; 8];
let data = serialize_vint_u32(val, &mut buf);
writer.write_all(data)

3
examples/custom_collector.rs
View File

@@ -11,10 +11,9 @@ use columnar::Column;
// ---
// Importing tantivy...
use tantivy::collector::{Collector, SegmentCollector};
use tantivy::index::SegmentReader;
use tantivy::query::QueryParser;
use tantivy::schema::{Schema, FAST, INDEXED, TEXT};
use tantivy::{doc, Index, IndexWriter, Score};
use tantivy::{doc, Index, IndexWriter, Score, SegmentReader};
#[derive(Default)]
struct Stats {

13
examples/date_time_field.rs
View File

@@ -4,7 +4,7 @@
use tantivy::collector::TopDocs;
use tantivy::query::QueryParser;
use tantivy::schema::{DateOptions, Document, Schema, INDEXED, STORED, STRING};
use tantivy::schema::{DateOptions, Document, OwnedValue, Schema, INDEXED, STORED, STRING};
use tantivy::{Index, IndexWriter, TantivyDocument};
fn main() -> tantivy::Result<()> {
@@ -13,7 +13,7 @@ fn main() -> tantivy::Result<()> {
let opts = DateOptions::from(INDEXED)
.set_stored()
.set_fast()
.set_precision(tantivy::schema::DateTimePrecision::Seconds);
.set_precision(tantivy::DateTimePrecision::Seconds);
// Add `occurred_at` date field type
let occurred_at = schema_builder.add_date_field("occurred_at", opts);
let event_type = schema_builder.add_text_field("event", STRING | STORED);
@@ -61,11 +61,10 @@ fn main() -> tantivy::Result<()> {
assert_eq!(count_docs.len(), 1);
for (_score, doc_address) in count_docs {
let retrieved_doc = searcher.doc::<TantivyDocument>(doc_address)?;
assert!(retrieved_doc
.get_first(occurred_at)
.unwrap()
.as_datetime()
.is_some(),);
assert!(matches!(
retrieved_doc.get_first(occurred_at),
Some(OwnedValue::Date(_))
));
assert_eq!(
retrieved_doc.to_json(&schema),
r#"{"event":["comment"],"occurred_at":["2022-06-22T13:00:00.22Z"]}"#

5
examples/faceted_search_with_tweaked_score.rs
View File

@@ -51,7 +51,7 @@ fn main() -> tantivy::Result<()> {
let reader = index.reader()?;
let searcher = reader.searcher();
{
let facets = [
let facets = vec![
Facet::from("/ingredient/egg"),
Facet::from("/ingredient/oil"),
Facet::from("/ingredient/garlic"),
@@ -94,8 +94,9 @@ fn main() -> tantivy::Result<()> {
.doc::<TantivyDocument>(*doc_id)
.unwrap()
.get_first(title)
.and_then(|v| v.as_str().map(|el| el.to_string()))
.and_then(|v| v.as_str())
.unwrap()
.to_owned()
})
.collect();
assert_eq!(titles, vec!["Fried egg", "Egg rolls"]);

3
examples/iterating_docs_and_positions.rs
View File

@@ -7,11 +7,10 @@
// the list of documents containing a term, getting
// its term frequency, and accessing its positions.
use tantivy::postings::Postings;
// ---
// Importing tantivy...
use tantivy::schema::*;
use tantivy::{doc, DocSet, Index, IndexWriter, TERMINATED};
use tantivy::{doc, DocSet, Index, IndexWriter, Postings, TERMINATED};
fn main() -> tantivy::Result<()> {
// We first create a schema for the sake of the

3
examples/warmer.rs
View File

@@ -3,11 +3,10 @@ use std::collections::{HashMap, HashSet};
use std::sync::{Arc, RwLock, Weak};
use tantivy::collector::TopDocs;
use tantivy::index::SegmentId;
use tantivy::query::QueryParser;
use tantivy::schema::{Schema, FAST, TEXT};
use tantivy::{
doc, DocAddress, DocId, Index, IndexWriter, Opstamp, Searcher, SearcherGeneration,
doc, DocAddress, DocId, Index, IndexWriter, Opstamp, Searcher, SearcherGeneration, SegmentId,
SegmentReader, Warmer,
};

2
ownedbytes/Cargo.toml
View File

@@ -1,7 +1,7 @@
[package]
authors = ["Paul Masurel <paul@quickwit.io>", "Pascal Seitz <pascal@quickwit.io>"]
name = "ownedbytes"
version = "0.7.0"
version = "0.6.0"
edition = "2021"
description = "Expose data as static slice"
license = "MIT"

2
query-grammar/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-query-grammar"
version = "0.22.0"
version = "0.21.0"
authors = ["Paul Masurel <paul.masurel@gmail.com>"]
license = "MIT"
categories = ["database-implementations", "data-structures"]

81
query-grammar/src/query_grammar.rs
View File

@@ -218,14 +218,27 @@ fn term_or_phrase_infallible(inp: &str) -> JResult<&str, Option<UserInputLeaf>>
}
fn term_group(inp: &str) -> IResult<&str, UserInputAst> {
let occur_symbol = alt((
value(Occur::MustNot, char('-')),
value(Occur::Must, char('+')),
));
map(
tuple((
terminated(field_name, multispace0),
delimited(tuple((char('('), multispace0)), ast, char(')')),
delimited(
tuple((char('('), multispace0)),
separated_list0(multispace1, tuple((opt(occur_symbol), term_or_phrase))),
char(')'),
),
)),
|(field_name, mut ast)| {
ast.set_default_field(field_name);
ast
|(field_name, terms)| {
UserInputAst::Clause(
terms
.into_iter()
.map(|(occur, leaf)| (occur, leaf.set_field(Some(field_name.clone())).into()))
.collect(),
)
},
)(inp)
}
@@ -245,18 +258,46 @@ fn term_group_precond(inp: &str) -> IResult<&str, (), ()> {
}
fn term_group_infallible(inp: &str) -> JResult<&str, UserInputAst> {
let (inp, (field_name, _, _, _)) =
let (mut inp, (field_name, _, _, _)) =
tuple((field_name, multispace0, char('('), multispace0))(inp).expect("precondition failed");
let res = delimited_infallible(
nothing,
map(ast_infallible, |(mut ast, errors)| {
ast.set_default_field(field_name.to_string());
(ast, errors)
}),
opt_i_err(char(')'), "expected ')'"),
)(inp);
res
let mut terms = Vec::new();
let mut errs = Vec::new();
let mut first_round = true;
loop {
let mut space_error = if first_round {
first_round = false;
Vec::new()
} else {
let (rest, (_, err)) = space1_infallible(inp)?;
inp = rest;
err
};
if inp.is_empty() {
errs.push(LenientErrorInternal {
pos: inp.len(),
message: "missing )".to_string(),
});
break Ok((inp, (UserInputAst::Clause(terms), errs)));
}
if let Some(inp) = inp.strip_prefix(')') {
break Ok((inp, (UserInputAst::Clause(terms), errs)));
}
// only append missing space error if we did not reach the end of group
errs.append(&mut space_error);
// here we do the assumption term_or_phrase_infallible always consume something if the
// first byte is not `)` or ' '. If it did not, we would end up looping.
let (rest, ((occur, leaf), mut err)) =
tuple_infallible((occur_symbol, term_or_phrase_infallible))(inp)?;
errs.append(&mut err);
if let Some(leaf) = leaf {
terms.push((occur, leaf.set_field(Some(field_name.clone())).into()));
}
inp = rest;
}
}
fn exists(inp: &str) -> IResult<&str, UserInputLeaf> {
@@ -1427,18 +1468,8 @@ mod test {
#[test]
fn test_parse_query_term_group() {
test_parse_query_to_ast_helper(r#"field:(abc)"#, r#""field":abc"#);
test_parse_query_to_ast_helper(r#"field:(abc)"#, r#"(*"field":abc)"#);
test_parse_query_to_ast_helper(r#"field:(+a -"b c")"#, r#"(+"field":a -"field":"b c")"#);
test_parse_query_to_ast_helper(r#"field:(a AND "b c")"#, r#"(+"field":a +"field":"b c")"#);
test_parse_query_to_ast_helper(r#"field:(a OR "b c")"#, r#"(?"field":a ?"field":"b c")"#);
test_parse_query_to_ast_helper(
r#"field:(a OR (b AND c))"#,
r#"(?"field":a ?(+"field":b +"field":c))"#,
);
test_parse_query_to_ast_helper(
r#"field:(a [b TO c])"#,
r#"(*"field":a *"field":["b" TO "c"])"#,
);
test_is_parse_err(r#"field:(+a -"b c""#, r#"(+"field":a -"field":"b c")"#);
}

30
query-grammar/src/user_input_ast.rs
View File

@@ -44,26 +44,6 @@ impl UserInputLeaf {
},
}
}
pub(crate) fn set_default_field(&mut self, default_field: String) {
match self {
UserInputLeaf::Literal(ref mut literal) if literal.field_name.is_none() => {
literal.field_name = Some(default_field)
}
UserInputLeaf::All => {
*self = UserInputLeaf::Exists {
field: default_field,
}
}
UserInputLeaf::Range { ref mut field, .. } if field.is_none() => {
*field = Some(default_field)
}
UserInputLeaf::Set { ref mut field, .. } if field.is_none() => {
*field = Some(default_field)
}
_ => (), // field was already set, do nothing
}
}
}
impl Debug for UserInputLeaf {
@@ -225,16 +205,6 @@ impl UserInputAst {
pub fn or(asts: Vec<UserInputAst>) -> UserInputAst {
UserInputAst::compose(Occur::Should, asts)
}
pub(crate) fn set_default_field(&mut self, field: String) {
match self {
UserInputAst::Clause(clauses) => clauses
.iter_mut()
.for_each(|(_, ast)| ast.set_default_field(field.clone())),
UserInputAst::Leaf(leaf) => leaf.set_default_field(field),
UserInputAst::Boost(ref mut ast, _) => ast.set_default_field(field),
}
}
}
impl From<UserInputLiteral> for UserInputLeaf {

585
src/aggregation/agg_bench.rs Normal file
View File

@@ -0,0 +1,585 @@
#[cfg(all(test, feature = "unstable"))]
mod bench {
use rand::prelude::SliceRandom;
use rand::rngs::StdRng;
use rand::{Rng, SeedableRng};
use rand_distr::Distribution;
use serde_json::json;
use test::{self, Bencher};
use crate::aggregation::agg_req::Aggregations;
use crate::aggregation::AggregationCollector;
use crate::query::{AllQuery, TermQuery};
use crate::schema::{IndexRecordOption, Schema, TextFieldIndexing, FAST, STRING};
use crate::{Index, Term};
#[derive(Clone, Copy, Hash, Default, Debug, PartialEq, Eq, PartialOrd, Ord)]
enum Cardinality {
/// All documents contain exactly one value.
/// `Full` is the default for auto-detecting the Cardinality, since it is the most strict.
#[default]
Full = 0,
/// All documents contain at most one value.
Optional = 1,
/// All documents may contain any number of values.
Multivalued = 2,
/// 1 / 20 documents has a value
Sparse = 3,
}
fn get_collector(agg_req: Aggregations) -> AggregationCollector {
AggregationCollector::from_aggs(agg_req, Default::default())
}
fn get_test_index_bench(cardinality: Cardinality) -> crate::Result<Index> {
let mut schema_builder = Schema::builder();
let text_fieldtype = crate::schema::TextOptions::default()
.set_indexing_options(
TextFieldIndexing::default().set_index_option(IndexRecordOption::WithFreqs),
)
.set_stored();
let text_field = schema_builder.add_text_field("text", text_fieldtype);
let json_field = schema_builder.add_json_field("json", 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 score_fieldtype = crate::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"];
let lg_norm = rand_distr::LogNormal::new(2.996f64, 0.979f64).unwrap();
let many_terms_data = (0..150_000)
.map(|num| format!("author{}", num))
.collect::<Vec<_>>();
{
let mut rng = StdRng::from_seed([1u8; 32]);
let mut index_writer = index.writer_with_num_threads(1, 200_000_000)?;
// To make the different test cases comparable we just change one doc to force the
// cardinality
if cardinality == Cardinality::Optional {
index_writer.add_document(doc!())?;
}
if cardinality == Cardinality::Multivalued {
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_many_terms => "cool",
text_field_many_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms => "cool",
score_field => 1u64,
score_field => 1u64,
score_field_f64 => lg_norm.sample(&mut rng),
score_field_f64 => lg_norm.sample(&mut rng),
score_field_i64 => 1i64,
score_field_i64 => 1i64,
))?;
}
let mut doc_with_value = 1_000_000;
if cardinality == Cardinality::Sparse {
doc_with_value /= 20;
}
let _val_max = 1_000_000.0;
for _ in 0..doc_with_value {
let val: f64 = rng.gen_range(0.0..1_000_000.0);
let json = if rng.gen_bool(0.1) {
// 10% are numeric values
json!({ "mixed_type": val })
} else {
json!({"mixed_type": many_terms_data.choose(&mut rng).unwrap().to_string()})
};
index_writer.add_document(doc!(
text_field => "cool",
json_field => json,
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(),
score_field => val as u64,
score_field_f64 => lg_norm.sample(&mut rng),
score_field_i64 => val as i64,
))?;
if cardinality == Cardinality::Sparse {
for _ in 0..20 {
index_writer.add_document(doc!(text_field => "cool"))?;
}
}
}
// writing the segment
index_writer.commit()?;
}
Ok(index)
}
use paste::paste;
#[macro_export]
macro_rules! bench_all_cardinalities {
( $x:ident ) => {
paste! {
#[bench]
fn $x(b: &mut Bencher) {
[<$x _card>](b, Cardinality::Full)
}
#[bench]
fn [<$x _opt>](b: &mut Bencher) {
[<$x _card>](b, Cardinality::Optional)
}
#[bench]
fn [<$x _multi>](b: &mut Bencher) {
[<$x _card>](b, Cardinality::Multivalued)
}
#[bench]
fn [<$x _sparse>](b: &mut Bencher) {
[<$x _card>](b, Cardinality::Sparse)
}
}
};
}
bench_all_cardinalities!(bench_aggregation_average_u64);
fn bench_aggregation_average_u64_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
b.iter(|| {
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let agg_req_1: Aggregations = serde_json::from_value(json!({
"average": { "avg": { "field": "score", } }
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&term_query, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_stats_f64);
fn bench_aggregation_stats_f64_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
b.iter(|| {
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let agg_req_1: Aggregations = serde_json::from_value(json!({
"average_f64": { "stats": { "field": "score_f64", } }
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&term_query, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_average_f64);
fn bench_aggregation_average_f64_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
b.iter(|| {
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let agg_req_1: Aggregations = serde_json::from_value(json!({
"average_f64": { "avg": { "field": "score_f64", } }
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&term_query, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_percentiles_f64);
fn bench_aggregation_percentiles_f64_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_str = r#"
{
"mypercentiles": {
"percentiles": {
"field": "score_f64",
"percents": [ 95, 99, 99.9 ]
}
}
} "#;
let agg_req_1: Aggregations = serde_json::from_str(agg_req_str).unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_average_u64_and_f64);
fn bench_aggregation_average_u64_and_f64_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
b.iter(|| {
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let agg_req_1: Aggregations = serde_json::from_value(json!({
"average_f64": { "avg": { "field": "score_f64" } },
"average": { "avg": { "field": "score" } },
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&term_query, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_few);
fn bench_aggregation_terms_few_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": { "terms": { "field": "text_few_terms" } },
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_many_with_top_hits_agg);
fn bench_aggregation_terms_many_with_top_hits_agg_card(
b: &mut Bencher,
cardinality: Cardinality,
) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": {
"terms": { "field": "text_many_terms" },
"aggs": {
"top_hits": { "top_hits":
{
"sort": [
{ "score": "desc" }
],
"size": 2,
"doc_value_fields": ["score_f64"]
}
}
}
},
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_many_with_sub_agg);
fn bench_aggregation_terms_many_with_sub_agg_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": {
"terms": { "field": "text_many_terms" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_many_json_mixed_type_with_sub_agg);
fn bench_aggregation_terms_many_json_mixed_type_with_sub_agg_card(
b: &mut Bencher,
cardinality: Cardinality,
) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": {
"terms": { "field": "json.mixed_type" },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_many2);
fn bench_aggregation_terms_many2_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": { "terms": { "field": "text_many_terms" } },
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_terms_many_order_by_term);
fn bench_aggregation_terms_many_order_by_term_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req: Aggregations = serde_json::from_value(json!({
"my_texts": { "terms": { "field": "text_many_terms", "order": { "_key": "desc" } } },
}))
.unwrap();
let collector = get_collector(agg_req);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_range_only);
fn bench_aggregation_range_only_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_1: Aggregations = serde_json::from_value(json!({
"range_f64": { "range": { "field": "score_f64", "ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
] } },
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_range_with_avg);
fn bench_aggregation_range_with_avg_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_1: Aggregations = serde_json::from_value(json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 30000 },
{ "from": 30000, "to": 40000 },
{ "from": 40000, "to": 50000 },
{ "from": 50000, "to": 60000 }
]
},
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
},
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
// hard bounds has a different algorithm, because it actually limits collection range
//
bench_all_cardinalities!(bench_aggregation_histogram_only_hard_bounds);
fn bench_aggregation_histogram_only_hard_bounds_card(
b: &mut Bencher,
cardinality: Cardinality,
) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_1: Aggregations = serde_json::from_value(json!({
"rangef64": { "histogram": { "field": "score_f64", "interval": 100, "hard_bounds": { "min": 1000, "max": 300000 } } },
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_histogram_with_avg);
fn bench_aggregation_histogram_with_avg_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_1: Aggregations = serde_json::from_value(json!({
"rangef64": {
"histogram": { "field": "score_f64", "interval": 100 },
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
}
}
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_histogram_only);
fn bench_aggregation_histogram_only_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
b.iter(|| {
let agg_req_1: Aggregations = serde_json::from_value(json!({
"rangef64": {
"histogram": {
"field": "score_f64",
"interval": 100 // 1000 buckets
},
}
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&AllQuery, &collector).unwrap()
});
}
bench_all_cardinalities!(bench_aggregation_avg_and_range_with_avg);
fn bench_aggregation_avg_and_range_with_avg_card(b: &mut Bencher, cardinality: Cardinality) {
let index = get_test_index_bench(cardinality).unwrap();
let reader = index.reader().unwrap();
let text_field = reader.searcher().schema().get_field("text").unwrap();
b.iter(|| {
let term_query = TermQuery::new(
Term::from_field_text(text_field, "cool"),
IndexRecordOption::Basic,
);
let agg_req_1: Aggregations = serde_json::from_value(json!({
"rangef64": {
"range": {
"field": "score_f64",
"ranges": [
{ "from": 3, "to": 7000 },
{ "from": 7000, "to": 20000 },
{ "from": 20000, "to": 60000 }
]
},
"aggs": {
"average_in_range": { "avg": { "field": "score" } }
}
},
"average": { "avg": { "field": "score" } }
}))
.unwrap();
let collector = get_collector(agg_req_1);
let searcher = reader.searcher();
searcher.search(&term_query, &collector).unwrap()
});
}
}

22
src/aggregation/agg_limits.rs
View File

@@ -81,11 +81,10 @@ impl AggregationLimits {
}
}
pub(crate) fn add_memory_consumed(&self, add_num_bytes: u64) -> crate::Result<()> {
let prev_value = self
.memory_consumption
.fetch_add(add_num_bytes, Ordering::Relaxed);
validate_memory_consumption(prev_value + add_num_bytes, self.memory_limit)?;
pub(crate) fn add_memory_consumed(&self, num_bytes: u64) -> crate::Result<()> {
self.memory_consumption
.fetch_add(num_bytes, Ordering::Relaxed);
validate_memory_consumption(&self.memory_consumption, self.memory_limit)?;
Ok(())
}
@@ -95,11 +94,11 @@ impl AggregationLimits {
}
fn validate_memory_consumption(
memory_consumption: u64,
memory_consumption: &AtomicU64,
memory_limit: ByteCount,
) -> Result<(), AggregationError> {
// Load the estimated memory consumed by the aggregations
let memory_consumed: ByteCount = memory_consumption.into();
let memory_consumed: ByteCount = memory_consumption.load(Ordering::Relaxed).into();
if memory_consumed > memory_limit {
return Err(AggregationError::MemoryExceeded {
limit: memory_limit,
@@ -119,11 +118,10 @@ pub struct ResourceLimitGuard {
}
impl ResourceLimitGuard {
pub(crate) fn add_memory_consumed(&self, add_num_bytes: u64) -> crate::Result<()> {
let prev_value = self
.memory_consumption
.fetch_add(add_num_bytes, Ordering::Relaxed);
validate_memory_consumption(prev_value + add_num_bytes, self.memory_limit)?;
pub(crate) fn add_memory_consumed(&self, num_bytes: u64) -> crate::Result<()> {
self.memory_consumption
.fetch_add(num_bytes, Ordering::Relaxed);
validate_memory_consumption(&self.memory_consumption, self.memory_limit)?;
Ok(())
}
}

3
src/aggregation/agg_req_with_accessor.rs
View File

@@ -17,8 +17,7 @@ use super::metric::{
use super::segment_agg_result::AggregationLimits;
use super::VecWithNames;
use crate::aggregation::{f64_to_fastfield_u64, Key};
use crate::index::SegmentReader;
use crate::SegmentOrdinal;
use crate::{SegmentOrdinal, SegmentReader};
#[derive(Default)]
pub(crate) struct AggregationsWithAccessor {

8
src/aggregation/bucket/histogram/histogram.rs
View File

@@ -331,11 +331,9 @@ impl SegmentAggregationCollector for SegmentHistogramCollector {
}
let mem_delta = self.get_memory_consumption() - mem_pre;
if mem_delta > 0 {
bucket_agg_accessor
.limits
.add_memory_consumed(mem_delta as u64)?;
}
bucket_agg_accessor
.limits
.add_memory_consumed(mem_delta as u64)?;
Ok(())
}

11
src/aggregation/bucket/mod.rs
View File

@@ -28,7 +28,6 @@ mod term_agg;
mod term_missing_agg;
use std::collections::HashMap;
use std::fmt;
pub use histogram::*;
pub use range::*;
@@ -73,12 +72,12 @@ impl From<&str> for OrderTarget {
}
}
impl fmt::Display for OrderTarget {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
impl ToString for OrderTarget {
fn to_string(&self) -> String {
match self {
OrderTarget::Key => f.write_str("_key"),
OrderTarget::Count => f.write_str("_count"),
OrderTarget::SubAggregation(agg) => agg.fmt(f),
OrderTarget::Key => "_key".to_string(),
OrderTarget::Count => "_count".to_string(),
OrderTarget::SubAggregation(agg) => agg.to_string(),
}
}
}

8
src/aggregation/bucket/term_agg.rs
View File

@@ -324,11 +324,9 @@ impl SegmentAggregationCollector for SegmentTermCollector {
}
let mem_delta = self.get_memory_consumption() - mem_pre;
if mem_delta > 0 {
bucket_agg_accessor
.limits
.add_memory_consumed(mem_delta as u64)?;
}
bucket_agg_accessor
.limits
.add_memory_consumed(mem_delta as u64)?;
Ok(())
}

3
src/aggregation/collector.rs
View File

@@ -8,8 +8,7 @@ use super::segment_agg_result::{
};
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};
use crate::{DocId, SegmentOrdinal, SegmentReader, TantivyError};
/// The default max bucket count, before the aggregation fails.
pub const DEFAULT_BUCKET_LIMIT: u32 = 65000;

48
src/aggregation/metric/top_hits.rs
View File

@@ -2,7 +2,6 @@ use std::collections::HashMap;
use std::net::Ipv6Addr;
use columnar::{ColumnarReader, DynamicColumn};
use common::json_path_writer::JSON_PATH_SEGMENT_SEP_STR;
use common::DateTime;
use regex::Regex;
use serde::ser::SerializeMap;
@@ -14,8 +13,8 @@ use crate::aggregation::intermediate_agg_result::{
IntermediateAggregationResult, IntermediateMetricResult,
};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::AggregationError;
use crate::collector::TopNComputer;
use crate::schema::term::JSON_PATH_SEGMENT_SEP_STR;
use crate::schema::OwnedValue;
use crate::{DocAddress, DocId, SegmentOrdinal};
@@ -97,14 +96,6 @@ pub struct TopHitsAggregation {
#[serde(rename = "docvalue_fields")]
#[serde(default)]
doc_value_fields: Vec<String>,
// Not supported
_source: Option<serde_json::Value>,
fields: Option<serde_json::Value>,
script_fields: Option<serde_json::Value>,
highlight: Option<serde_json::Value>,
explain: Option<serde_json::Value>,
version: Option<serde_json::Value>,
}
#[derive(Debug, Clone, PartialEq, Default)]
@@ -151,49 +142,12 @@ fn globbed_string_to_regex(glob: &str) -> Result<Regex, crate::TantivyError> {
})
}
fn use_doc_value_fields_err(parameter: &str) -> crate::Result<()> {
Err(crate::TantivyError::AggregationError(
AggregationError::InvalidRequest(format!(
"The `{}` parameter is not supported, only `docvalue_fields` is supported in \
`top_hits` aggregation",
parameter
)),
))
}
fn unsupported_err(parameter: &str) -> crate::Result<()> {
Err(crate::TantivyError::AggregationError(
AggregationError::InvalidRequest(format!(
"The `{}` parameter is not supported in the `top_hits` aggregation",
parameter
)),
))
}
impl TopHitsAggregation {
/// Validate and resolve field retrieval parameters
pub fn validate_and_resolve_field_names(
&mut self,
reader: &ColumnarReader,
) -> crate::Result<()> {
if self._source.is_some() {
use_doc_value_fields_err("_source")?;
}
if self.fields.is_some() {
use_doc_value_fields_err("fields")?;
}
if self.script_fields.is_some() {
use_doc_value_fields_err("script_fields")?;
}
if self.explain.is_some() {
unsupported_err("explain")?;
}
if self.highlight.is_some() {
unsupported_err("highlight")?;
}
if self.version.is_some() {
unsupported_err("version")?;
}
self.doc_value_fields = self
.doc_value_fields
.iter()

2
src/aggregation/mod.rs
View File

@@ -143,6 +143,8 @@ use std::fmt::Display;
#[cfg(test)]
mod agg_tests;
mod agg_bench;
use core::fmt;
pub use agg_limits::AggregationLimits;

2
src/collector/facet_collector.rs
View File

@@ -1,7 +1,7 @@
use std::cmp::Ordering;
use std::collections::{btree_map, BTreeMap, BTreeSet, BinaryHeap};
use std::io;
use std::ops::Bound;
use std::{io, u64, usize};
use crate::collector::{Collector, SegmentCollector};
use crate::fastfield::FacetReader;

3
src/collector/top_collector.rs
View File

@@ -4,8 +4,7 @@ use std::marker::PhantomData;
use serde::{Deserialize, Serialize};
use super::top_score_collector::TopNComputer;
use crate::index::SegmentReader;
use crate::{DocAddress, DocId, SegmentOrdinal};
use crate::{DocAddress, DocId, SegmentOrdinal, SegmentReader};
/// Contains a feature (field, score, etc.) of a document along with the document address.
///

113
src/core/executor.rs
View File

@@ -1,25 +1,19 @@
use std::sync::Arc;
#[cfg(feature = "quickwit")]
use futures_util::{future::Either, FutureExt};
use rayon::{ThreadPool, ThreadPoolBuilder};
use crate::TantivyError;
/// Executor makes it possible to run tasks in single thread or
/// in a thread pool.
#[derive(Clone)]
/// Search executor whether search request are single thread or multithread.
///
/// We don't expose Rayon thread pool directly here for several reasons.
///
/// First dependency hell. It is not a good idea to expose the
/// API of a dependency, knowing it might conflict with a different version
/// used by the client. Second, we may stop using rayon in the future.
pub enum Executor {
/// Single thread variant of an Executor
SingleThread,
/// Thread pool variant of an Executor
ThreadPool(Arc<rayon::ThreadPool>),
}
#[cfg(feature = "quickwit")]
impl From<Arc<rayon::ThreadPool>> for Executor {
fn from(thread_pool: Arc<rayon::ThreadPool>) -> Self {
Executor::ThreadPool(thread_pool)
}
ThreadPool(ThreadPool),
}
impl Executor {
@@ -30,11 +24,11 @@ impl Executor {
/// Creates an Executor that dispatches the tasks in a thread pool.
pub fn multi_thread(num_threads: usize, prefix: &'static str) -> crate::Result<Executor> {
let pool = rayon::ThreadPoolBuilder::new()
let pool = ThreadPoolBuilder::new()
.num_threads(num_threads)
.thread_name(move |num| format!("{prefix}{num}"))
.build()?;
Ok(Executor::ThreadPool(Arc::new(pool)))
Ok(Executor::ThreadPool(pool))
}
/// Perform a map in the thread pool.
@@ -97,36 +91,11 @@ impl Executor {
}
}
}
/// Spawn a task on the pool, returning a future completing on task success.
///
/// If the task panic, returns `Err(())`.
#[cfg(feature = "quickwit")]
pub fn spawn_blocking<T: Send + 'static>(
&self,
cpu_intensive_task: impl FnOnce() -> T + Send + 'static,
) -> impl std::future::Future<Output = Result<T, ()>> {
match self {
Executor::SingleThread => Either::Left(std::future::ready(Ok(cpu_intensive_task()))),
Executor::ThreadPool(pool) => {
let (sender, receiver) = oneshot::channel();
pool.spawn(|| {
if sender.is_closed() {
return;
}
let task_result = cpu_intensive_task();
let _ = sender.send(task_result);
});
let res = receiver.map(|res| res.map_err(|_| ()));
Either::Right(res)
}
}
}
}
#[cfg(test)]
mod tests {
use super::Executor;
#[test]
@@ -178,62 +147,4 @@ mod tests {
assert_eq!(result[i], i * 2);
}
}
#[cfg(feature = "quickwit")]
#[test]
fn test_cancel_cpu_intensive_tasks() {
use std::sync::atomic::{AtomicU64, Ordering};
use std::sync::Arc;
let counter: Arc<AtomicU64> = Default::default();
let other_counter: Arc<AtomicU64> = Default::default();
let mut futures = Vec::new();
let mut other_futures = Vec::new();
let (tx, rx) = crossbeam_channel::bounded::<()>(0);
let rx = Arc::new(rx);
let executor = Executor::multi_thread(3, "search-test").unwrap();
for i in 0..1000 {
let counter_clone: Arc<AtomicU64> = counter.clone();
let other_counter_clone: Arc<AtomicU64> = other_counter.clone();
let rx_clone = rx.clone();
let rx_clone2 = rx.clone();
let fut = executor.spawn_blocking(move || {
counter_clone.fetch_add(1, Ordering::SeqCst);
let () = rx_clone.recv().unwrap();
});
futures.push(fut);
let other_fut = executor.spawn_blocking(move || {
other_counter_clone.fetch_add(1, Ordering::SeqCst);
let () = rx_clone2.recv().unwrap();
});
other_futures.push(other_fut);
}
// We execute 100 futures.
for i in 0..100 {
tx.send(()).unwrap();
}
let counter_val = counter.load(Ordering::SeqCst);
let other_counter_val = other_counter.load(Ordering::SeqCst);
assert!(counter_val >= 30);
assert!(other_counter_val >= 30);
drop(other_futures);
// We execute 100 futures.
for i in 0..100 {
tx.send(()).unwrap();
}
let counter_val2 = counter.load(Ordering::SeqCst);
assert!(counter_val2 >= counter_val + 100 - 6);
let other_counter_val2 = other_counter.load(Ordering::SeqCst);
assert!(other_counter_val2 <= other_counter_val + 6);
}
}

373
src/core/json_utils.rs
View File

@@ -1,10 +1,12 @@
use common::json_path_writer::JSON_PATH_SEGMENT_SEP;
use columnar::MonotonicallyMappableToU64;
use common::{replace_in_place, JsonPathWriter};
use rustc_hash::FxHashMap;
use crate::fastfield::FastValue;
use crate::postings::{IndexingContext, IndexingPosition, PostingsWriter};
use crate::schema::document::{ReferenceValue, ReferenceValueLeaf, Value};
use crate::schema::Type;
use crate::schema::term::JSON_PATH_SEGMENT_SEP;
use crate::schema::{Field, Type, DATE_TIME_PRECISION_INDEXED};
use crate::time::format_description::well_known::Rfc3339;
use crate::time::{OffsetDateTime, UtcOffset};
use crate::tokenizer::TextAnalyzer;
@@ -31,7 +33,7 @@ use crate::{DateTime, DocId, Term};
/// position 1.
/// As a result, with lemmatization, "The Smiths" will match our object.
///
/// Worse, if a same term appears in the second object, a non increasing value would be pushed
/// Worse, if a same term is appears in the second object, a non increasing value would be pushed
/// to the position recorder probably provoking a panic.
///
/// This problem is solved for regular multivalued object by offsetting the position
@@ -50,7 +52,7 @@ use crate::{DateTime, DocId, Term};
/// We can therefore afford working with a map that is not imperfect. It is fine if several
/// path map to the same index position as long as the probability is relatively low.
#[derive(Default)]
pub(crate) struct IndexingPositionsPerPath {
struct IndexingPositionsPerPath {
positions_per_path: FxHashMap<u32, IndexingPosition>,
}
@@ -58,9 +60,6 @@ impl IndexingPositionsPerPath {
fn get_position_from_id(&mut self, id: u32) -> &mut IndexingPosition {
self.positions_per_path.entry(id).or_default()
}
pub fn clear(&mut self) {
self.positions_per_path.clear();
}
}
/// Convert JSON_PATH_SEGMENT_SEP to a dot.
@@ -71,6 +70,36 @@ pub fn json_path_sep_to_dot(path: &mut str) {
}
}
#[allow(clippy::too_many_arguments)]
pub(crate) fn index_json_values<'a, V: Value<'a>>(
doc: DocId,
json_visitors: impl Iterator<Item = crate::Result<V::ObjectIter>>,
text_analyzer: &mut TextAnalyzer,
expand_dots_enabled: bool,
term_buffer: &mut Term,
postings_writer: &mut dyn PostingsWriter,
json_path_writer: &mut JsonPathWriter,
ctx: &mut IndexingContext,
) -> crate::Result<()> {
json_path_writer.clear();
json_path_writer.set_expand_dots(expand_dots_enabled);
let mut positions_per_path: IndexingPositionsPerPath = Default::default();
for json_visitor_res in json_visitors {
let json_visitor = json_visitor_res?;
index_json_object::<V>(
doc,
json_visitor,
text_analyzer,
term_buffer,
json_path_writer,
postings_writer,
ctx,
&mut positions_per_path,
);
}
Ok(())
}
#[allow(clippy::too_many_arguments)]
fn index_json_object<'a, V: Value<'a>>(
doc: DocId,
@@ -99,7 +128,7 @@ fn index_json_object<'a, V: Value<'a>>(
}
#[allow(clippy::too_many_arguments)]
pub(crate) fn index_json_value<'a, V: Value<'a>>(
fn index_json_value<'a, V: Value<'a>>(
doc: DocId,
json_value: V,
text_analyzer: &mut TextAnalyzer,
@@ -139,18 +168,12 @@ pub(crate) fn index_json_value<'a, V: Value<'a>>(
);
}
ReferenceValueLeaf::U64(val) => {
// try to parse to i64, since when querying we will apply the same logic and prefer
// i64 values
set_path_id(
term_buffer,
ctx.path_to_unordered_id
.get_or_allocate_unordered_id(json_path_writer.as_str()),
);
if let Ok(i64_val) = val.try_into() {
term_buffer.append_type_and_fast_value::<i64>(i64_val);
} else {
term_buffer.append_type_and_fast_value(val);
}
term_buffer.append_type_and_fast_value(val);
postings_writer.subscribe(doc, 0u32, term_buffer, ctx);
}
ReferenceValueLeaf::I64(val) => {
@@ -233,42 +256,71 @@ pub(crate) fn index_json_value<'a, V: Value<'a>>(
}
}
/// Tries to infer a JSON type from a string and append it to the term.
///
/// The term must be json + JSON path.
pub fn convert_to_fast_value_and_append_to_json_term(mut term: Term, phrase: &str) -> Option<Term> {
assert_eq!(
term.value()
.as_json_value_bytes()
.expect("expecting a Term with a json type and json path")
.as_serialized()
.len(),
0,
"JSON value bytes should be empty"
);
// Tries to infer a JSON type from a string.
pub fn convert_to_fast_value_and_get_term(
json_term_writer: &mut JsonTermWriter,
phrase: &str,
) -> Option<Term> {
if let Ok(dt) = OffsetDateTime::parse(phrase, &Rfc3339) {
let dt_utc = dt.to_offset(UtcOffset::UTC);
term.append_type_and_fast_value(DateTime::from_utc(dt_utc));
return Some(term);
return Some(set_fastvalue_and_get_term(
json_term_writer,
DateTime::from_utc(dt_utc),
));
}
if let Ok(i64_val) = str::parse::<i64>(phrase) {
term.append_type_and_fast_value(i64_val);
return Some(term);
return Some(set_fastvalue_and_get_term(json_term_writer, i64_val));
}
if let Ok(u64_val) = str::parse::<u64>(phrase) {
term.append_type_and_fast_value(u64_val);
return Some(term);
return Some(set_fastvalue_and_get_term(json_term_writer, u64_val));
}
if let Ok(f64_val) = str::parse::<f64>(phrase) {
term.append_type_and_fast_value(f64_val);
return Some(term);
return Some(set_fastvalue_and_get_term(json_term_writer, f64_val));
}
if let Ok(bool_val) = str::parse::<bool>(phrase) {
term.append_type_and_fast_value(bool_val);
return Some(term);
return Some(set_fastvalue_and_get_term(json_term_writer, bool_val));
}
None
}
// helper function to generate a Term from a json fastvalue
pub(crate) fn set_fastvalue_and_get_term<T: FastValue>(
json_term_writer: &mut JsonTermWriter,
value: T,
) -> Term {
json_term_writer.set_fast_value(value);
json_term_writer.term().clone()
}
// helper function to generate a list of terms with their positions from a textual json value
pub(crate) fn set_string_and_get_terms(
json_term_writer: &mut JsonTermWriter,
value: &str,
text_analyzer: &mut TextAnalyzer,
) -> Vec<(usize, Term)> {
let mut positions_and_terms = Vec::<(usize, Term)>::new();
json_term_writer.close_path_and_set_type(Type::Str);
let term_num_bytes = json_term_writer.term_buffer.len_bytes();
let mut token_stream = text_analyzer.token_stream(value);
token_stream.process(&mut |token| {
json_term_writer
.term_buffer
.truncate_value_bytes(term_num_bytes);
json_term_writer
.term_buffer
.append_bytes(token.text.as_bytes());
positions_and_terms.push((token.position, json_term_writer.term().clone()));
});
positions_and_terms
}
/// Writes a value of a JSON field to a `Term`.
/// The Term format is as follows:
/// `[JSON_TYPE][JSON_PATH][JSON_END_OF_PATH][VALUE_BYTES]`
pub struct JsonTermWriter<'a> {
term_buffer: &'a mut Term,
path_stack: Vec<usize>,
expand_dots_enabled: bool,
}
/// Splits a json path supplied to the query parser in such a way that
/// `.` can be escaped.
@@ -325,48 +377,158 @@ pub(crate) fn encode_column_name(
path.into()
}
impl<'a> JsonTermWriter<'a> {
pub fn from_field_and_json_path(
field: Field,
json_path: &str,
expand_dots_enabled: bool,
term_buffer: &'a mut Term,
) -> Self {
term_buffer.set_field_and_type(field, Type::Json);
let mut json_term_writer = Self::wrap(term_buffer, expand_dots_enabled);
for segment in split_json_path(json_path) {
json_term_writer.push_path_segment(&segment);
}
json_term_writer
}
pub fn wrap(term_buffer: &'a mut Term, expand_dots_enabled: bool) -> Self {
term_buffer.clear_with_type(Type::Json);
let mut path_stack = Vec::with_capacity(10);
path_stack.push(0);
Self {
term_buffer,
path_stack,
expand_dots_enabled,
}
}
fn trim_to_end_of_path(&mut self) {
let end_of_path = *self.path_stack.last().unwrap();
self.term_buffer.truncate_value_bytes(end_of_path);
}
pub fn close_path_and_set_type(&mut self, typ: Type) {
self.trim_to_end_of_path();
self.term_buffer.set_json_path_end();
self.term_buffer.append_bytes(&[typ.to_code()]);
}
// TODO: Remove this function and use JsonPathWriter instead.
pub fn push_path_segment(&mut self, segment: &str) {
// the path stack should never be empty.
self.trim_to_end_of_path();
if self.path_stack.len() > 1 {
self.term_buffer.set_json_path_separator();
}
let appended_segment = self.term_buffer.append_bytes(segment.as_bytes());
if self.expand_dots_enabled {
// We need to replace `.` by JSON_PATH_SEGMENT_SEP.
replace_in_place(b'.', JSON_PATH_SEGMENT_SEP, appended_segment);
}
self.term_buffer.add_json_path_separator();
self.path_stack.push(self.term_buffer.len_bytes());
}
pub fn pop_path_segment(&mut self) {
self.path_stack.pop();
assert!(!self.path_stack.is_empty());
self.trim_to_end_of_path();
}
/// Returns the json path of the term being currently built.
#[cfg(test)]
pub(crate) fn path(&self) -> &[u8] {
let end_of_path = self.path_stack.last().cloned().unwrap_or(1);
&self.term().serialized_value_bytes()[..end_of_path - 1]
}
pub(crate) fn set_fast_value<T: FastValue>(&mut self, val: T) {
self.close_path_and_set_type(T::to_type());
let value = if T::to_type() == Type::Date {
DateTime::from_u64(val.to_u64())
.truncate(DATE_TIME_PRECISION_INDEXED)
.to_u64()
} else {
val.to_u64()
};
self.term_buffer
.append_bytes(value.to_be_bytes().as_slice());
}
pub fn set_str(&mut self, text: &str) {
self.close_path_and_set_type(Type::Str);
self.term_buffer.append_bytes(text.as_bytes());
}
pub fn term(&self) -> &Term {
self.term_buffer
}
}
#[cfg(test)]
mod tests {
use super::split_json_path;
use crate::schema::Field;
use super::{split_json_path, JsonTermWriter};
use crate::schema::{Field, Type};
use crate::Term;
#[test]
fn test_json_writer() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "attributes.color", false);
term.append_type_and_str("red");
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("attributes");
json_writer.push_path_segment("color");
json_writer.set_str("red");
assert_eq!(
format!("{:?}", term),
format!("{:?}", json_writer.term()),
"Term(field=1, type=Json, path=attributes.color, type=Str, \"red\")"
);
let mut term = Term::from_field_json_path(field, "attributes.dimensions.width", false);
term.append_type_and_fast_value(400i64);
json_writer.set_str("blue");
assert_eq!(
format!("{:?}", term),
format!("{:?}", json_writer.term()),
"Term(field=1, type=Json, path=attributes.color, type=Str, \"blue\")"
);
json_writer.pop_path_segment();
json_writer.push_path_segment("dimensions");
json_writer.push_path_segment("width");
json_writer.set_fast_value(400i64);
assert_eq!(
format!("{:?}", json_writer.term()),
"Term(field=1, type=Json, path=attributes.dimensions.width, type=I64, 400)"
);
json_writer.pop_path_segment();
json_writer.push_path_segment("height");
json_writer.set_fast_value(300i64);
assert_eq!(
format!("{:?}", json_writer.term()),
"Term(field=1, type=Json, path=attributes.dimensions.height, type=I64, 300)"
);
}
#[test]
fn test_string_term() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_str("red");
assert_eq!(term.serialized_term(), b"\x00\x00\x00\x01jcolor\x00sred")
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.set_str("red");
assert_eq!(
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00sred"
)
}
#[test]
fn test_i64_term() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(-4i64);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.set_fast_value(-4i64);
assert_eq!(
term.serialized_term(),
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00i\x7f\xff\xff\xff\xff\xff\xff\xfc"
)
}
@@ -374,11 +536,12 @@ mod tests {
#[test]
fn test_u64_term() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(4u64);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.set_fast_value(4u64);
assert_eq!(
term.serialized_term(),
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00u\x00\x00\x00\x00\x00\x00\x00\x04"
)
}
@@ -386,10 +549,12 @@ mod tests {
#[test]
fn test_f64_term() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(4.0f64);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.set_fast_value(4.0f64);
assert_eq!(
term.serialized_term(),
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00f\xc0\x10\x00\x00\x00\x00\x00\x00"
)
}
@@ -397,14 +562,90 @@ mod tests {
#[test]
fn test_bool_term() {
let field = Field::from_field_id(1);
let mut term = Term::from_field_json_path(field, "color", false);
term.append_type_and_fast_value(true);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.set_fast_value(true);
assert_eq!(
term.serialized_term(),
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00o\x00\x00\x00\x00\x00\x00\x00\x01"
)
}
#[test]
fn test_push_after_set_path_segment() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("attribute");
json_writer.set_str("something");
json_writer.push_path_segment("color");
json_writer.set_str("red");
assert_eq!(
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jattribute\x01color\x00sred"
)
}
#[test]
fn test_pop_segment() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
json_writer.push_path_segment("hue");
json_writer.pop_path_segment();
json_writer.set_str("red");
assert_eq!(
json_writer.term().serialized_term(),
b"\x00\x00\x00\x01jcolor\x00sred"
)
}
#[test]
fn test_json_writer_path() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color");
assert_eq!(json_writer.path(), b"color");
json_writer.push_path_segment("hue");
assert_eq!(json_writer.path(), b"color\x01hue");
json_writer.set_str("pink");
assert_eq!(json_writer.path(), b"color\x01hue");
}
#[test]
fn test_json_path_expand_dots_disabled() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, false);
json_writer.push_path_segment("color.hue");
assert_eq!(json_writer.path(), b"color.hue");
}
#[test]
fn test_json_path_expand_dots_enabled() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, true);
json_writer.push_path_segment("color.hue");
assert_eq!(json_writer.path(), b"color\x01hue");
}
#[test]
fn test_json_path_expand_dots_enabled_pop_segment() {
let field = Field::from_field_id(1);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_writer = JsonTermWriter::wrap(&mut term, true);
json_writer.push_path_segment("hello");
assert_eq!(json_writer.path(), b"hello");
json_writer.push_path_segment("color.hue");
assert_eq!(json_writer.path(), b"hello\x01color\x01hue");
json_writer.pop_path_segment();
assert_eq!(json_writer.path(), b"hello");
}
#[test]
fn test_split_json_path_simple() {
let json_path = split_json_path("titi.toto");

33
src/core/searcher.rs
View File

@@ -1,17 +1,16 @@
use std::collections::BTreeMap;
use std::marker::PhantomData;
use std::sync::Arc;
use std::{fmt, io};
use crate::collector::Collector;
use crate::core::Executor;
use crate::index::{SegmentId, SegmentReader};
use crate::index::SegmentReader;
use crate::query::{Bm25StatisticsProvider, EnableScoring, Query};
use crate::schema::document::{DocumentDeserialize, DocumentDeserializeSeed};
use crate::schema::document::DocumentDeserialize;
use crate::schema::{Schema, Term};
use crate::space_usage::SearcherSpaceUsage;
use crate::store::{CacheStats, StoreReader};
use crate::{DocAddress, Index, Opstamp, TrackedObject};
use crate::{DocAddress, Index, Opstamp, SegmentId, TrackedObject};
/// Identifies the searcher generation accessed by a [`Searcher`].
///
@@ -87,17 +86,8 @@ impl Searcher {
/// The searcher uses the segment ordinal to route the
/// request to the right `Segment`.
pub fn doc<D: DocumentDeserialize>(&self, doc_address: DocAddress) -> crate::Result<D> {
self.doc_seed(doc_address, PhantomData)
}
/// A stateful variant of [`doc`][Self::doc].`
pub fn doc_seed<T: DocumentDeserializeSeed>(
&self,
doc_address: DocAddress,
seed: T,
) -> crate::Result<T::Value> {
let store_reader = &self.inner.store_readers[doc_address.segment_ord as usize];
store_reader.get_seed(doc_address.doc_id, seed)
store_reader.get(doc_address.doc_id)
}
/// The cache stats for the underlying store reader.
@@ -119,21 +109,8 @@ impl Searcher {
&self,
doc_address: DocAddress,
) -> crate::Result<D> {
self.doc_async_seed(doc_address, PhantomData).await
}
#[cfg(feature = "quickwit")]
/// A stateful variant of [`doc_async`][Self::doc_async].
pub async fn doc_async_seed<T: DocumentDeserializeSeed>(
&self,
doc_address: DocAddress,
seed: T,
) -> crate::Result<T::Value> {
let executor = self.inner.index.search_executor();
let store_reader = &self.inner.store_readers[doc_address.segment_ord as usize];
store_reader
.get_async_seed(doc_address.doc_id, executor, seed)
.await
store_reader.get_async(doc_address.doc_id).await
}
/// Access the schema associated with the index of this searcher.

37
src/core/tests.rs
View File

@@ -1,14 +1,13 @@
use crate::collector::Count;
use crate::directory::{RamDirectory, WatchCallback};
use crate::index::SegmentId;
use crate::indexer::{LogMergePolicy, NoMergePolicy};
use crate::postings::Postings;
use crate::json_utils::JsonTermWriter;
use crate::query::TermQuery;
use crate::schema::{Field, IndexRecordOption, Schema, INDEXED, STRING, TEXT};
use crate::schema::{Field, IndexRecordOption, Schema, Type, INDEXED, STRING, TEXT};
use crate::tokenizer::TokenizerManager;
use crate::{
Directory, DocSet, Index, IndexBuilder, IndexReader, IndexSettings, IndexWriter, ReloadPolicy,
TantivyDocument, Term,
Directory, DocSet, Index, IndexBuilder, IndexReader, IndexSettings, IndexWriter, Postings,
ReloadPolicy, SegmentId, TantivyDocument, Term,
};
#[test]
@@ -417,12 +416,16 @@ fn test_non_text_json_term_freq() {
let searcher = reader.searcher();
let segment_reader = searcher.segment_reader(0u32);
let inv_idx = segment_reader.inverted_index(field).unwrap();
let mut term = Term::from_field_json_path(field, "tenant_id", false);
term.append_type_and_fast_value(75i64);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
json_term_writer.push_path_segment("tenant_id");
json_term_writer.close_path_and_set_type(Type::U64);
json_term_writer.set_fast_value(75u64);
let postings = inv_idx
.read_postings(&term, IndexRecordOption::WithFreqsAndPositions)
.read_postings(
json_term_writer.term(),
IndexRecordOption::WithFreqsAndPositions,
)
.unwrap()
.unwrap();
assert_eq!(postings.doc(), 0);
@@ -451,12 +454,16 @@ fn test_non_text_json_term_freq_bitpacked() {
let searcher = reader.searcher();
let segment_reader = searcher.segment_reader(0u32);
let inv_idx = segment_reader.inverted_index(field).unwrap();
let mut term = Term::from_field_json_path(field, "tenant_id", false);
term.append_type_and_fast_value(75i64);
let mut term = Term::with_type_and_field(Type::Json, field);
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
json_term_writer.push_path_segment("tenant_id");
json_term_writer.close_path_and_set_type(Type::U64);
json_term_writer.set_fast_value(75u64);
let mut postings = inv_idx
.read_postings(&term, IndexRecordOption::WithFreqsAndPositions)
.read_postings(
json_term_writer.term(),
IndexRecordOption::WithFreqsAndPositions,
)
.unwrap()
.unwrap();
assert_eq!(postings.doc(), 0);

8
src/fastfield/facet_reader.rs
View File

@@ -62,6 +62,7 @@ impl FacetReader {
#[cfg(test)]
mod tests {
use crate::schema::document::Value;
use crate::schema::{Facet, FacetOptions, SchemaBuilder, STORED};
use crate::{DocAddress, Index, IndexWriter, TantivyDocument};
@@ -145,11 +146,8 @@ mod tests {
facet_ords.extend(facet_reader.facet_ords(0u32));
assert_eq!(&facet_ords, &[0u64]);
let doc = searcher.doc::<TantivyDocument>(DocAddress::new(0u32, 0u32))?;
let value: Option<Facet> = doc
.get_first(facet_field)
.and_then(|v| v.as_facet())
.map(|facet| Facet::from_encoded_string(facet.to_string()));
assert_eq!(value, Facet::from_text("/a/b").ok());
let value: Option<&Facet> = doc.get_first(facet_field).and_then(|v| v.as_facet());
assert_eq!(value, Facet::from_text("/a/b").ok().as_ref());
Ok(())
}

9
src/fastfield/mod.rs
View File

@@ -80,7 +80,7 @@ mod tests {
use std::path::Path;
use columnar::StrColumn;
use common::{ByteCount, DateTimePrecision, HasLen, TerminatingWrite};
use common::{ByteCount, HasLen, TerminatingWrite};
use once_cell::sync::Lazy;
use rand::prelude::SliceRandom;
use rand::rngs::StdRng;
@@ -88,15 +88,14 @@ mod tests {
use super::*;
use crate::directory::{Directory, RamDirectory, WritePtr};
use crate::index::SegmentId;
use crate::merge_policy::NoMergePolicy;
use crate::schema::{
DateOptions, Facet, FacetOptions, Field, JsonObjectOptions, Schema, SchemaBuilder,
TantivyDocument, TextOptions, FAST, INDEXED, STORED, STRING, TEXT,
Facet, FacetOptions, Field, JsonObjectOptions, Schema, SchemaBuilder, TantivyDocument,
TextOptions, FAST, INDEXED, STORED, STRING, TEXT,
};
use crate::time::OffsetDateTime;
use crate::tokenizer::{LowerCaser, RawTokenizer, TextAnalyzer, TokenizerManager};
use crate::{Index, IndexWriter, SegmentReader};
use crate::{DateOptions, DateTimePrecision, Index, IndexWriter, SegmentId, SegmentReader};
pub static SCHEMA: Lazy<Schema> = Lazy::new(|| {
let mut schema_builder = Schema::builder();

7
src/fastfield/writer.rs
View File

@@ -1,14 +1,14 @@
use std::io;
use columnar::{ColumnarWriter, NumericalValue};
use common::{DateTimePrecision, JsonPathWriter};
use common::JsonPathWriter;
use tokenizer_api::Token;
use crate::indexer::doc_id_mapping::DocIdMapping;
use crate::schema::document::{Document, ReferenceValue, ReferenceValueLeaf, Value};
use crate::schema::{value_type_to_column_type, Field, FieldType, Schema, Type};
use crate::tokenizer::{TextAnalyzer, TokenizerManager};
use crate::{DocId, TantivyError};
use crate::{DateTimePrecision, DocId, TantivyError};
/// Only index JSON down to a depth of 20.
/// This is mostly to guard us from a stack overflow triggered by malicious input.
@@ -183,7 +183,8 @@ impl FastFieldsWriter {
.record_datetime(doc_id, field_name, truncated_datetime);
}
ReferenceValueLeaf::Facet(val) => {
self.columnar_writer.record_str(doc_id, field_name, val);
self.columnar_writer
.record_str(doc_id, field_name, val.encoded_str());
}
ReferenceValueLeaf::Bytes(val) => {
self.columnar_writer.record_bytes(doc_id, field_name, val);

3
src/functional_test.rs
View File

@@ -1,12 +1,9 @@
#![allow(deprecated)] // Remove with index sorting
use std::collections::HashSet;
use rand::{thread_rng, Rng};
use crate::indexer::index_writer::MEMORY_BUDGET_NUM_BYTES_MIN;
use crate::schema::*;
#[allow(deprecated)]
use crate::{doc, schema, Index, IndexSettings, IndexSortByField, IndexWriter, Order, Searcher};
fn check_index_content(searcher: &Searcher, vals: &[u64]) -> crate::Result<()> {

40
src/index/index.rs
View File

@@ -3,7 +3,7 @@ use std::fmt;
#[cfg(feature = "mmap")]
use std::path::Path;
use std::path::PathBuf;
use std::thread::available_parallelism;
use std::sync::Arc;
use super::segment::Segment;
use super::segment_reader::merge_field_meta_data;
@@ -20,7 +20,7 @@ use crate::indexer::segment_updater::save_metas;
use crate::indexer::{IndexWriter, SingleSegmentIndexWriter};
use crate::reader::{IndexReader, IndexReaderBuilder};
use crate::schema::document::Document;
use crate::schema::{Field, FieldType, Schema, Type};
use crate::schema::{Field, FieldType, Schema};
use crate::tokenizer::{TextAnalyzer, TokenizerManager};
use crate::SegmentReader;
@@ -83,7 +83,7 @@ fn save_new_metas(
///
/// ```
/// use tantivy::schema::*;
/// use tantivy::{Index, IndexSettings};
/// use tantivy::{Index, IndexSettings, IndexSortByField, Order};
///
/// let mut schema_builder = Schema::builder();
/// let id_field = schema_builder.add_text_field("id", STRING);
@@ -96,7 +96,10 @@ fn save_new_metas(
///
/// let schema = schema_builder.build();
/// let settings = IndexSettings{
/// docstore_blocksize: 100_000,
/// sort_by_field: Some(IndexSortByField{
/// field: "number".to_string(),
/// order: Order::Asc
/// }),
/// ..Default::default()
/// };
/// let index = Index::builder().schema(schema).settings(settings).create_in_ram();
@@ -248,15 +251,6 @@ impl IndexBuilder {
sort_by_field.field
)));
}
let supported_field_types = [Type::I64, Type::U64, Type::F64, Type::Date];
let field_type = entry.field_type().value_type();
if !supported_field_types.contains(&field_type) {
return Err(TantivyError::InvalidArgument(format!(
"Unsupported field type in sort_by_field: {:?}. Supported field types: \
{:?} ",
field_type, supported_field_types,
)));
}
}
Ok(())
} else {
@@ -293,7 +287,7 @@ pub struct Index {
directory: ManagedDirectory,
schema: Schema,
settings: IndexSettings,
executor: Executor,
executor: Arc<Executor>,
tokenizers: TokenizerManager,
fast_field_tokenizers: TokenizerManager,
inventory: SegmentMetaInventory,
@@ -318,25 +312,29 @@ impl Index {
///
/// By default the executor is single thread, and simply runs in the calling thread.
pub fn search_executor(&self) -> &Executor {
&self.executor
self.executor.as_ref()
}
/// Replace the default single thread search executor pool
/// by a thread pool with a given number of threads.
pub fn set_multithread_executor(&mut self, num_threads: usize) -> crate::Result<()> {
self.executor = Executor::multi_thread(num_threads, "tantivy-search-")?;
self.executor = Arc::new(Executor::multi_thread(num_threads, "tantivy-search-")?);
Ok(())
}
/// Custom thread pool by a outer thread pool.
pub fn set_executor(&mut self, executor: Executor) {
self.executor = executor;
pub fn set_shared_multithread_executor(
&mut self,
shared_thread_pool: Arc<Executor>,
) -> crate::Result<()> {
self.executor = shared_thread_pool.clone();
Ok(())
}
/// Replace the default single thread search executor pool
/// by a thread pool with as many threads as there are CPUs on the system.
pub fn set_default_multithread_executor(&mut self) -> crate::Result<()> {
let default_num_threads = available_parallelism()?.get();
let default_num_threads = num_cpus::get();
self.set_multithread_executor(default_num_threads)
}
@@ -414,7 +412,7 @@ impl Index {
schema,
tokenizers: TokenizerManager::default(),
fast_field_tokenizers: TokenizerManager::default(),
executor: Executor::single_thread(),
executor: Arc::new(Executor::single_thread()),
inventory,
}
}
@@ -617,7 +615,7 @@ impl Index {
&self,
memory_budget_in_bytes: usize,
) -> crate::Result<IndexWriter<D>> {
let mut num_threads = std::cmp::min(available_parallelism()?.get(), MAX_NUM_THREAD);
let mut num_threads = std::cmp::min(num_cpus::get(), MAX_NUM_THREAD);
let memory_budget_num_bytes_per_thread = memory_budget_in_bytes / num_threads;
if memory_budget_num_bytes_per_thread < MEMORY_BUDGET_NUM_BYTES_MIN {
num_threads = (memory_budget_in_bytes / MEMORY_BUDGET_NUM_BYTES_MIN).max(1);

4
src/index/index_meta.rs
View File

@@ -288,10 +288,6 @@ impl Default for IndexSettings {
/// Presorting documents can greatly improve performance
/// in some scenarios, by applying top n
/// optimizations.
#[deprecated(
since = "0.22.0",
note = "We plan to remove index sorting in `0.23`. If you need index sorting, please comment on the related issue https://github.com/quickwit-oss/tantivy/issues/2352 and explain your use case."
)]
#[derive(Clone, Debug, Serialize, Deserialize, Eq, PartialEq)]
pub struct IndexSortByField {
/// The field to sort the documents by

3
src/index/inverted_index_reader.rs
View File

@@ -1,13 +1,12 @@
use std::io;
use common::json_path_writer::JSON_END_OF_PATH;
use common::BinarySerializable;
use fnv::FnvHashSet;
use crate::directory::FileSlice;
use crate::positions::PositionReader;
use crate::postings::{BlockSegmentPostings, SegmentPostings, TermInfo};
use crate::schema::{IndexRecordOption, Term, Type};
use crate::schema::{IndexRecordOption, Term, Type, JSON_END_OF_PATH};
use crate::termdict::TermDictionary;
/// The inverted index reader is in charge of accessing

2
src/index/mod.rs
View File

@@ -1,3 +1,5 @@
//! # Index Module
//!
//! The `index` module in Tantivy contains core components to read and write indexes.
//!
//! It contains `Index` and `Segment`, where a `Index` consists of one or more `Segment`s.

2
src/index/segment_reader.rs
View File

@@ -406,7 +406,7 @@ impl SegmentReader {
}
/// Returns an iterator that will iterate over the alive document ids
pub fn doc_ids_alive(&self) -> Box<dyn Iterator<Item = DocId> + Send + '_> {
pub fn doc_ids_alive(&self) -> Box<dyn Iterator<Item = DocId> + '_> {
if let Some(alive_bitset) = &self.alive_bitset_opt {
Box::new(alive_bitset.iter_alive())
} else {

3
src/indexer/delete_queue.rs
View File

@@ -246,9 +246,8 @@ impl DeleteCursor {
mod tests {
use super::{DeleteOperation, DeleteQueue};
use crate::index::SegmentReader;
use crate::query::{Explanation, Scorer, Weight};
use crate::{DocId, Score};
use crate::{DocId, Score, SegmentReader};
struct DummyWeight;
impl Weight for DummyWeight {

38
src/indexer/doc_id_mapping.rs
View File

@@ -159,7 +159,7 @@ mod tests_indexsorting {
use crate::indexer::NoMergePolicy;
use crate::query::QueryParser;
use crate::schema::*;
use crate::{DocAddress, Index, IndexBuilder, IndexSettings, IndexSortByField, Order};
use crate::{DocAddress, Index, IndexSettings, IndexSortByField, Order};
fn create_test_index(
index_settings: Option<IndexSettings>,
@@ -306,10 +306,12 @@ mod tests_indexsorting {
let my_string_field = index.schema().get_field("string_field").unwrap();
let searcher = index.reader()?.searcher();
{
assert!(searcher
.doc::<TantivyDocument>(DocAddress::new(0, 0))?
.get_first(my_string_field)
.is_none());
assert_eq!(
searcher
.doc::<TantivyDocument>(DocAddress::new(0, 0))?
.get_first(my_string_field),
None
);
assert_eq!(
searcher
.doc::<TantivyDocument>(DocAddress::new(0, 3))?
@@ -342,7 +344,7 @@ mod tests_indexsorting {
Some("blublub")
);
let doc = searcher.doc::<TantivyDocument>(DocAddress::new(0, 4))?;
assert!(doc.get_first(my_string_field).is_none());
assert_eq!(doc.get_first(my_string_field), None);
}
// sort by field desc
let index = create_test_index(
@@ -555,28 +557,4 @@ mod tests_indexsorting {
&[2000, 8000, 3000]
);
}
#[test]
fn test_text_sort() -> crate::Result<()> {
let mut schema_builder = SchemaBuilder::new();
schema_builder.add_text_field("id", STRING | FAST | STORED);
schema_builder.add_text_field("name", TEXT | STORED);
let resp = IndexBuilder::new()
.schema(schema_builder.build())
.settings(IndexSettings {
sort_by_field: Some(IndexSortByField {
field: "id".to_string(),
order: Order::Asc,
}),
..Default::default()
})
.create_in_ram();
assert!(resp
.unwrap_err()
.to_string()
.contains("Unsupported field type"));
Ok(())
}
}

3
src/indexer/index_writer.rs
View File

@@ -814,6 +814,7 @@ mod tests {
use crate::indexer::index_writer::MEMORY_BUDGET_NUM_BYTES_MIN;
use crate::indexer::NoMergePolicy;
use crate::query::{BooleanQuery, Occur, Query, QueryParser, TermQuery};
use crate::schema::document::Value;
use crate::schema::{
self, Facet, FacetOptions, IndexRecordOption, IpAddrOptions, NumericOptions, Schema,
TextFieldIndexing, TextOptions, FAST, INDEXED, STORED, STRING, TEXT,
@@ -2012,7 +2013,7 @@ mod tests {
let mut bool2 = doc.get_all(multi_bools);
assert_eq!(bool, bool2.next().unwrap().as_bool().unwrap());
assert_ne!(bool, bool2.next().unwrap().as_bool().unwrap());
assert!(bool2.next().is_none())
assert_eq!(None, bool2.next())
}
}
}

4
src/indexer/log_merge_policy.rs
View File

@@ -144,9 +144,9 @@ mod tests {
use once_cell::sync::Lazy;
use super::*;
use crate::index::{SegmentId, SegmentMetaInventory};
use crate::schema;
use crate::index::SegmentMetaInventory;
use crate::schema::INDEXED;
use crate::{schema, SegmentId};
static INVENTORY: Lazy<SegmentMetaInventory> = Lazy::new(SegmentMetaInventory::default);

3
src/indexer/merge_operation.rs
View File

@@ -1,8 +1,7 @@
use std::collections::HashSet;
use std::ops::Deref;
use crate::index::SegmentId;
use crate::{Inventory, Opstamp, TrackedObject};
use crate::{Inventory, Opstamp, SegmentId, TrackedObject};
#[derive(Default)]
pub(crate) struct MergeOperationInventory(Inventory<InnerMergeOperation>);

10
src/indexer/merger.rs
View File

@@ -13,7 +13,7 @@ use crate::docset::{DocSet, TERMINATED};
use crate::error::DataCorruption;
use crate::fastfield::{AliveBitSet, FastFieldNotAvailableError};
use crate::fieldnorm::{FieldNormReader, FieldNormReaders, FieldNormsSerializer, FieldNormsWriter};
use crate::index::{Segment, SegmentComponent, SegmentReader};
use crate::index::{Segment, SegmentReader};
use crate::indexer::doc_id_mapping::{MappingType, SegmentDocIdMapping};
use crate::indexer::SegmentSerializer;
use crate::postings::{InvertedIndexSerializer, Postings, SegmentPostings};
@@ -21,7 +21,8 @@ use crate::schema::{value_type_to_column_type, Field, FieldType, Schema};
use crate::store::StoreWriter;
use crate::termdict::{TermMerger, TermOrdinal};
use crate::{
DocAddress, DocId, IndexSettings, IndexSortByField, InvertedIndexReader, Order, SegmentOrdinal,
DocAddress, DocId, IndexSettings, IndexSortByField, InvertedIndexReader, Order,
SegmentComponent, SegmentOrdinal,
};
/// Segment's max doc must be `< MAX_DOC_LIMIT`.
@@ -793,8 +794,9 @@ mod tests {
BytesFastFieldTestCollector, FastFieldTestCollector, TEST_COLLECTOR_WITH_SCORE,
};
use crate::collector::{Count, FacetCollector};
use crate::index::{Index, SegmentId};
use crate::index::Index;
use crate::query::{AllQuery, BooleanQuery, EnableScoring, Scorer, TermQuery};
use crate::schema::document::Value;
use crate::schema::{
Facet, FacetOptions, IndexRecordOption, NumericOptions, TantivyDocument, Term,
TextFieldIndexing, INDEXED, TEXT,
@@ -802,7 +804,7 @@ mod tests {
use crate::time::OffsetDateTime;
use crate::{
assert_nearly_equals, schema, DateTime, DocAddress, DocId, DocSet, IndexSettings,
IndexSortByField, IndexWriter, Order, Searcher,
IndexSortByField, IndexWriter, Order, Searcher, SegmentId,
};
#[test]

6
src/indexer/merger_sorted_index_test.rs
View File

@@ -3,15 +3,15 @@ mod tests {
use crate::collector::TopDocs;
use crate::fastfield::AliveBitSet;
use crate::index::Index;
use crate::postings::Postings;
use crate::query::QueryParser;
use crate::schema::document::Value;
use crate::schema::{
self, BytesOptions, Facet, FacetOptions, IndexRecordOption, NumericOptions,
TextFieldIndexing, TextOptions,
};
use crate::{
DocAddress, DocSet, IndexSettings, IndexSortByField, IndexWriter, Order, TantivyDocument,
Term,
DocAddress, DocSet, IndexSettings, IndexSortByField, IndexWriter, Order, Postings,
TantivyDocument, Term,
};
fn create_test_index_posting_list_issue(index_settings: Option<IndexSettings>) -> Index {

78
src/indexer/mod.rs
View File

@@ -156,24 +156,16 @@ mod tests_mmap {
}
#[test]
fn test_json_field_1byte() {
// Test when field name contains a '1' byte, which has special meaning in tantivy.
// The 1 byte can be addressed as '1' byte or '.'.
// Test when field name contains a 1 byte, which has special meaning in tantivy.
let field_name_in = "\u{0001}";
let field_name_out = "\u{0001}";
test_json_field_name(field_name_in, field_name_out);
// Test when field name contains a '1' byte, which has special meaning in tantivy.
// Test when field name contains a 1 byte, which has special meaning in tantivy.
let field_name_in = "\u{0001}";
let field_name_out = ".";
test_json_field_name(field_name_in, field_name_out);
}
#[test]
fn test_json_field_dot() {
// Test when field name contains a '.'
let field_name_in = ".";
let field_name_out = ".";
test_json_field_name(field_name_in, field_name_out);
}
fn test_json_field_name(field_name_in: &str, field_name_out: &str) {
let mut schema_builder = Schema::builder();
@@ -182,7 +174,7 @@ mod tests_mmap {
let index = Index::create_in_ram(schema_builder.build());
let mut index_writer = index.writer_for_tests().unwrap();
index_writer
.add_document(doc!(field=>json!({format!("{field_name_in}"): "test1", format!("num{field_name_in}"): 10})))
.add_document(doc!(field=>json!({format!("{field_name_in}"): "test1"})))
.unwrap();
index_writer
.add_document(doc!(field=>json!({format!("a{field_name_in}"): "test2"})))
@@ -213,10 +205,10 @@ mod tests_mmap {
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let parse_query = QueryParser::for_index(&index, Vec::new());
let test_query = |query_str: &str| {
let query = parse_query.parse_query(query_str).unwrap();
let test_query = |field_name: &str| {
let query = parse_query.parse_query(field_name).unwrap();
let num_docs = searcher.search(&query, &Count).unwrap();
assert_eq!(num_docs, 1, "{}", query_str);
assert_eq!(num_docs, 1);
};
test_query(format!("json.{field_name_out}:test1").as_str());
test_query(format!("json.a{field_name_out}:test2").as_str());
@@ -260,64 +252,6 @@ mod tests_mmap {
"test6",
);
test_agg(format!("json.{field_name_out}a").as_str(), "test7");
// `.` is stored as `\u{0001}` internally in tantivy
let field_name_out_internal = if field_name_out == "." {
"\u{0001}"
} else {
field_name_out
};
let mut fields = reader.searcher().segment_readers()[0]
.inverted_index(field)
.unwrap()
.list_encoded_fields()
.unwrap();
assert_eq!(fields.len(), 8);
fields.sort();
let mut expected_fields = vec![
(format!("a{field_name_out_internal}"), Type::Str),
(format!("a{field_name_out_internal}a"), Type::Str),
(
format!("a{field_name_out_internal}a{field_name_out_internal}"),
Type::Str,
),
(
format!("a{field_name_out_internal}\u{1}ab{field_name_out_internal}"),
Type::Str,
),
(
format!("a{field_name_out_internal}\u{1}a{field_name_out_internal}"),
Type::Str,
),
(format!("{field_name_out_internal}a"), Type::Str),
(format!("{field_name_out_internal}"), Type::Str),
(format!("num{field_name_out_internal}"), Type::I64),
];
expected_fields.sort();
assert_eq!(fields, expected_fields);
// Check columnar reader
let mut columns = reader.searcher().segment_readers()[0]
.fast_fields()
.columnar()
.list_columns()
.unwrap()
.into_iter()
.map(|(name, _)| name)
.collect::<Vec<_>>();
let mut expected_columns = vec![
format!("json\u{1}{field_name_out_internal}"),
format!("json\u{1}{field_name_out_internal}a"),
format!("json\u{1}a{field_name_out_internal}"),
format!("json\u{1}a{field_name_out_internal}a"),
format!("json\u{1}a{field_name_out_internal}a{field_name_out_internal}"),
format!("json\u{1}a{field_name_out_internal}\u{1}ab{field_name_out_internal}"),
format!("json\u{1}a{field_name_out_internal}\u{1}a{field_name_out_internal}"),
format!("json\u{1}num{field_name_out_internal}"),
];
columns.sort();
expected_columns.sort();
assert_eq!(columns, expected_columns);
}
#[test]

284
src/indexer/segment_writer.rs
View File

@@ -5,20 +5,20 @@ use tokenizer_api::BoxTokenStream;
use super::doc_id_mapping::{get_doc_id_mapping_from_field, DocIdMapping};
use super::operation::AddOperation;
use crate::core::json_utils::index_json_values;
use crate::fastfield::FastFieldsWriter;
use crate::fieldnorm::{FieldNormReaders, FieldNormsWriter};
use crate::index::{Segment, SegmentComponent};
use crate::index::Segment;
use crate::indexer::segment_serializer::SegmentSerializer;
use crate::json_utils::{index_json_value, IndexingPositionsPerPath};
use crate::postings::{
compute_table_memory_size, serialize_postings, IndexingContext, IndexingPosition,
PerFieldPostingsWriter, PostingsWriter,
};
use crate::schema::document::{Document, Value};
use crate::schema::document::{Document, ReferenceValue, Value};
use crate::schema::{FieldEntry, FieldType, Schema, Term, DATE_TIME_PRECISION_INDEXED};
use crate::store::{StoreReader, StoreWriter};
use crate::tokenizer::{FacetTokenizer, PreTokenizedStream, TextAnalyzer, Tokenizer};
use crate::{DocId, Opstamp, TantivyError};
use crate::{DocId, Opstamp, SegmentComponent, TantivyError};
/// Computes the initial size of the hash table.
///
@@ -68,7 +68,6 @@ pub struct SegmentWriter {
pub(crate) fast_field_writers: FastFieldsWriter,
pub(crate) fieldnorms_writer: FieldNormsWriter,
pub(crate) json_path_writer: JsonPathWriter,
pub(crate) json_positions_per_path: IndexingPositionsPerPath,
pub(crate) doc_opstamps: Vec<Opstamp>,
per_field_text_analyzers: Vec<TextAnalyzer>,
term_buffer: Term,
@@ -120,7 +119,6 @@ impl SegmentWriter {
per_field_postings_writers,
fieldnorms_writer: FieldNormsWriter::for_schema(&schema),
json_path_writer: JsonPathWriter::default(),
json_positions_per_path: IndexingPositionsPerPath::default(),
segment_serializer,
fast_field_writers: FastFieldsWriter::from_schema_and_tokenizer_manager(
&schema,
@@ -206,7 +204,8 @@ impl SegmentWriter {
// Used to help with linting and type checking.
let value = value_access as D::Value<'_>;
let facet_str = value.as_facet().ok_or_else(make_schema_error)?;
let facet = value.as_facet().ok_or_else(make_schema_error)?;
let facet_str = facet.encoded_str();
let mut facet_tokenizer = facet_tokenizer.token_stream(facet_str);
let mut indexing_position = IndexingPosition::default();
postings_writer.index_text(
@@ -229,7 +228,7 @@ impl SegmentWriter {
&mut self.per_field_text_analyzers[field.field_id() as usize];
text_analyzer.token_stream(text)
} else if let Some(tok_str) = value.as_pre_tokenized_text() {
BoxTokenStream::new(PreTokenizedStream::from(*tok_str.clone()))
BoxTokenStream::new(PreTokenizedStream::from(tok_str.clone()))
} else {
continue;
};
@@ -343,24 +342,26 @@ impl SegmentWriter {
FieldType::JsonObject(json_options) => {
let text_analyzer =
&mut self.per_field_text_analyzers[field.field_id() as usize];
let json_values_it = values.map(|value_access| {
// Used to help with linting and type checking.
let value_access = value_access as D::Value<'_>;
let value = value_access.as_value();
self.json_positions_per_path.clear();
self.json_path_writer
.set_expand_dots(json_options.is_expand_dots_enabled());
for json_value in values {
self.json_path_writer.clear();
index_json_value(
doc_id,
json_value,
text_analyzer,
term_buffer,
&mut self.json_path_writer,
postings_writer,
ctx,
&mut self.json_positions_per_path,
);
}
match value {
ReferenceValue::Object(object_iter) => Ok(object_iter),
_ => Err(make_schema_error()),
}
});
index_json_values::<D::Value<'_>>(
doc_id,
json_values_it,
text_analyzer,
json_options.is_expand_dots_enabled(),
term_buffer,
postings_writer,
&mut self.json_path_writer,
ctx,
)?;
}
FieldType::IpAddr(_) => {
let mut num_vals = 0;
@@ -495,21 +496,22 @@ mod tests {
use tempfile::TempDir;
use crate::collector::{Count, TopDocs};
use crate::core::json_utils::JsonTermWriter;
use crate::directory::RamDirectory;
use crate::fastfield::FastValue;
use crate::postings::{Postings, TermInfo};
use crate::postings::TermInfo;
use crate::query::{PhraseQuery, QueryParser};
use crate::schema::document::Value;
use crate::schema::{
Document, IndexRecordOption, OwnedValue, Schema, TextFieldIndexing, TextOptions, STORED,
STRING, TEXT,
Document, IndexRecordOption, Schema, TextFieldIndexing, TextOptions, Type, STORED, STRING,
TEXT,
};
use crate::store::{Compressor, StoreReader, StoreWriter};
use crate::time::format_description::well_known::Rfc3339;
use crate::time::OffsetDateTime;
use crate::tokenizer::{PreTokenizedString, Token};
use crate::{
DateTime, Directory, DocAddress, DocSet, Index, IndexWriter, TantivyDocument, Term,
TERMINATED,
DateTime, Directory, DocAddress, DocSet, Index, IndexWriter, Postings, TantivyDocument,
Term, TERMINATED,
};
#[test]
@@ -554,12 +556,9 @@ mod tests {
let reader = StoreReader::open(directory.open_read(path).unwrap(), 0).unwrap();
let doc = reader.get::<TantivyDocument>(0).unwrap();
assert_eq!(doc.field_values().count(), 2);
assert_eq!(doc.get_all(text_field).next().unwrap().as_str(), Some("A"));
assert_eq!(
doc.get_all(text_field).nth(1).unwrap().as_str(),
Some("title")
);
assert_eq!(doc.field_values().len(), 2);
assert_eq!(doc.field_values()[0].value().as_str(), Some("A"));
assert_eq!(doc.field_values()[1].value().as_str(), Some("title"));
}
#[test]
fn test_simple_json_indexing() {
@@ -599,51 +598,12 @@ mod tests {
assert_eq!(score_docs.len(), 2);
}
#[test]
fn test_flat_json_indexing() {
// A JSON Object that contains mixed values on the first level
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", STORED | STRING);
let schema = schema_builder.build();
let index = Index::create_in_ram(schema.clone());
let mut writer = index.writer_for_tests().unwrap();
// Text, i64, u64
writer.add_document(doc!(json_field=>"b")).unwrap();
writer
.add_document(doc!(json_field=>OwnedValue::I64(10i64)))
.unwrap();
writer
.add_document(doc!(json_field=>OwnedValue::U64(55u64)))
.unwrap();
writer
.add_document(doc!(json_field=>json!({"my_field": "a"})))
.unwrap();
writer.commit().unwrap();
let search_and_expect = |query| {
let query_parser = QueryParser::for_index(&index, vec![json_field]);
let text_query = query_parser.parse_query(query).unwrap();
let score_docs: Vec<(_, DocAddress)> = index
.reader()
.unwrap()
.searcher()
.search(&text_query, &TopDocs::with_limit(4))
.unwrap();
assert_eq!(score_docs.len(), 1);
};
search_and_expect("my_field:a");
search_and_expect("b");
search_and_expect("10");
search_and_expect("55");
}
#[test]
fn test_json_indexing() {
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", STORED | TEXT);
let schema = schema_builder.build();
let json_val: serde_json::Value = serde_json::from_str(
let json_val: serde_json::Map<String, serde_json::Value> = serde_json::from_str(
r#"{
"toto": "titi",
"float": -0.2,
@@ -671,125 +631,129 @@ mod tests {
doc_id: 0u32,
})
.unwrap();
let serdeser_json_val = serde_json::from_str::<serde_json::Value>(&doc.to_json(&schema))
let serdeser_json_val = serde_json::from_str::<serde_json::Map<String, serde_json::Value>>(
&doc.to_json(&schema),
)
.unwrap()
.get("json")
.unwrap()[0]
.as_object()
.unwrap()
.get("json")
.unwrap()[0]
.clone();
assert_eq!(json_val, serdeser_json_val);
let segment_reader = searcher.segment_reader(0u32);
let inv_idx = segment_reader.inverted_index(json_field).unwrap();
let term_dict = inv_idx.terms();
let mut term = Term::with_type_and_field(Type::Json, json_field);
let mut term_stream = term_dict.stream().unwrap();
let term_from_path =
|path: &str| -> Term { Term::from_field_json_path(json_field, path, false) };
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
fn set_fast_val<T: FastValue>(val: T, mut term: Term) -> Term {
term.append_type_and_fast_value(val);
term
}
fn set_str(val: &str, mut term: Term) -> Term {
term.append_type_and_str(val);
term
}
let term = term_from_path("bool");
json_term_writer.push_path_segment("bool");
json_term_writer.set_fast_value(true);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(true, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
let term = term_from_path("complexobject.field\\.with\\.dot");
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("complexobject");
json_term_writer.push_path_segment("field.with.dot");
json_term_writer.set_fast_value(1i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(1i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
// Date
let term = term_from_path("date");
json_term_writer.pop_path_segment();
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("date");
json_term_writer.set_fast_value(DateTime::from_utc(
OffsetDateTime::parse("1985-04-12T23:20:50.52Z", &Rfc3339).unwrap(),
));
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(
DateTime::from_utc(
OffsetDateTime::parse("1985-04-12T23:20:50.52Z", &Rfc3339).unwrap(),
),
term
)
.serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
// Float
let term = term_from_path("float");
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("float");
json_term_writer.set_fast_value(-0.2f64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(-0.2f64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
// Number In Array
let term = term_from_path("my_arr");
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("my_arr");
json_term_writer.set_fast_value(2i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(2i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
let term = term_from_path("my_arr");
json_term_writer.set_fast_value(3i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(3i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
let term = term_from_path("my_arr");
json_term_writer.set_fast_value(4i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(4i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
// El in Array
let term = term_from_path("my_arr.my_key");
json_term_writer.push_path_segment("my_key");
json_term_writer.set_str("tokens");
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_str("tokens", term).serialized_value_bytes()
);
let term = term_from_path("my_arr.my_key");
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_str("two", term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
// Signed
let term = term_from_path("signed");
json_term_writer.set_str("two");
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(-2i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
let term = term_from_path("toto");
json_term_writer.pop_path_segment();
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("signed");
json_term_writer.set_fast_value(-2i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_str("titi", term).serialized_value_bytes()
);
// Unsigned
let term = term_from_path("unsigned");
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
set_fast_val(1i64, term).serialized_value_bytes()
json_term_writer.term().serialized_value_bytes()
);
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("toto");
json_term_writer.set_str("titi");
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
json_term_writer.term().serialized_value_bytes()
);
json_term_writer.pop_path_segment();
json_term_writer.push_path_segment("unsigned");
json_term_writer.set_fast_value(1i64);
assert!(term_stream.advance());
assert_eq!(
term_stream.key(),
json_term_writer.term().serialized_value_bytes()
);
assert!(!term_stream.advance());
}
@@ -810,9 +774,14 @@ mod tests {
let searcher = reader.searcher();
let segment_reader = searcher.segment_reader(0u32);
let inv_index = segment_reader.inverted_index(json_field).unwrap();
let mut term = Term::from_field_json_path(json_field, "mykey", false);
term.append_type_and_str("token");
let term_info = inv_index.get_term_info(&term).unwrap().unwrap();
let mut term = Term::with_type_and_field(Type::Json, json_field);
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
json_term_writer.push_path_segment("mykey");
json_term_writer.set_str("token");
let term_info = inv_index
.get_term_info(json_term_writer.term())
.unwrap()
.unwrap();
assert_eq!(
term_info,
TermInfo {
@@ -838,7 +807,7 @@ mod tests {
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", STRING);
let schema = schema_builder.build();
let json_val: serde_json::Value =
let json_val: serde_json::Map<String, serde_json::Value> =
serde_json::from_str(r#"{"mykey": "two tokens"}"#).unwrap();
let doc = doc!(json_field=>json_val);
let index = Index::create_in_ram(schema);
@@ -849,9 +818,14 @@ mod tests {
let searcher = reader.searcher();
let segment_reader = searcher.segment_reader(0u32);
let inv_index = segment_reader.inverted_index(json_field).unwrap();
let mut term = Term::from_field_json_path(json_field, "mykey", false);
term.append_type_and_str("two tokens");
let term_info = inv_index.get_term_info(&term).unwrap().unwrap();
let mut term = Term::with_type_and_field(Type::Json, json_field);
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
json_term_writer.push_path_segment("mykey");
json_term_writer.set_str("two tokens");
let term_info = inv_index
.get_term_info(json_term_writer.term())
.unwrap()
.unwrap();
assert_eq!(
term_info,
TermInfo {
@@ -878,7 +852,7 @@ mod tests {
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", TEXT);
let schema = schema_builder.build();
let json_val: serde_json::Value = serde_json::from_str(
let json_val: serde_json::Map<String, serde_json::Value> = serde_json::from_str(
r#"{"mykey": [{"field": "hello happy tax payer"}, {"field": "nothello"}]}"#,
)
.unwrap();
@@ -889,18 +863,16 @@ mod tests {
writer.commit().unwrap();
let reader = index.reader().unwrap();
let searcher = reader.searcher();
let term = Term::from_field_json_path(json_field, "mykey.field", false);
let mut hello_term = term.clone();
hello_term.append_type_and_str("hello");
let mut nothello_term = term.clone();
nothello_term.append_type_and_str("nothello");
let mut happy_term = term.clone();
happy_term.append_type_and_str("happy");
let mut term = Term::with_type_and_field(Type::Json, json_field);
let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
json_term_writer.push_path_segment("mykey");
json_term_writer.push_path_segment("field");
json_term_writer.set_str("hello");
let hello_term = json_term_writer.term().clone();
json_term_writer.set_str("nothello");
let nothello_term = json_term_writer.term().clone();
json_term_writer.set_str("happy");
let happy_term = json_term_writer.term().clone();
let phrase_query = PhraseQuery::new(vec![hello_term, happy_term.clone()]);
assert_eq!(searcher.search(&phrase_query, &Count).unwrap(), 1);
let phrase_query = PhraseQuery::new(vec![nothello_term, happy_term]);

50
src/lib.rs
View File

@@ -178,7 +178,6 @@ pub use crate::future_result::FutureResult;
pub type Result<T> = std::result::Result<T, TantivyError>;
mod core;
#[allow(deprecated)] // Remove with index sorting
pub mod indexer;
#[allow(unused_doc_comments)]
@@ -190,7 +189,6 @@ pub mod collector;
pub mod directory;
pub mod fastfield;
pub mod fieldnorm;
#[allow(deprecated)] // Remove with index sorting
pub mod index;
pub mod positions;
pub mod postings;
@@ -216,17 +214,29 @@ use once_cell::sync::Lazy;
use serde::{Deserialize, Serialize};
pub use self::docset::{DocSet, COLLECT_BLOCK_BUFFER_LEN, TERMINATED};
#[deprecated(
since = "0.22.0",
note = "Will be removed in tantivy 0.23. Use export from snippet module instead"
)]
pub use self::snippet::{Snippet, SnippetGenerator};
#[doc(hidden)]
pub use crate::core::json_utils;
pub use crate::core::{Executor, Searcher, SearcherGeneration};
pub use crate::directory::Directory;
#[allow(deprecated)] // Remove with index sorting
pub use crate::index::{
Index, IndexBuilder, IndexMeta, IndexSettings, IndexSortByField, InvertedIndexReader, Order,
Segment, SegmentMeta, SegmentReader,
Segment, SegmentComponent, SegmentId, SegmentMeta, SegmentReader,
};
#[deprecated(
since = "0.22.0",
note = "Will be removed in tantivy 0.23. Use export from indexer module instead"
)]
pub use crate::indexer::PreparedCommit;
pub use crate::indexer::{IndexWriter, SingleSegmentIndexWriter};
pub use crate::schema::{Document, TantivyDocument, Term};
pub use crate::postings::Postings;
#[allow(deprecated)]
pub use crate::schema::DatePrecision;
pub use crate::schema::{DateOptions, DateTimePrecision, Document, TantivyDocument, Term};
/// Index format version.
const INDEX_FORMAT_VERSION: u32 = 6;
@@ -244,7 +254,7 @@ pub struct Version {
impl fmt::Debug for Version {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
fmt::Display::fmt(self, f)
write!(f, "{}", self.to_string())
}
}
@@ -255,10 +265,9 @@ static VERSION: Lazy<Version> = Lazy::new(|| Version {
index_format_version: INDEX_FORMAT_VERSION,
});
impl fmt::Display for Version {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(
f,
impl ToString for Version {
fn to_string(&self) -> String {
format!(
"tantivy v{}.{}.{}, index_format v{}",
self.major, self.minor, self.patch, self.index_format_version
)
@@ -381,10 +390,9 @@ pub mod tests {
use crate::docset::{DocSet, TERMINATED};
use crate::index::SegmentReader;
use crate::merge_policy::NoMergePolicy;
use crate::postings::Postings;
use crate::query::BooleanQuery;
use crate::schema::*;
use crate::{DateTime, DocAddress, Index, IndexWriter, ReloadPolicy};
use crate::{DateTime, DocAddress, Index, IndexWriter, Postings, ReloadPolicy};
pub fn fixed_size_test<O: BinarySerializable + FixedSize + Default>() {
let mut buffer = Vec::new();
@@ -436,6 +444,7 @@ pub mod tests {
}
#[test]
#[cfg(not(feature = "lz4"))]
fn test_version_string() {
use regex::Regex;
let regex_ptn = Regex::new(
@@ -935,7 +944,7 @@ pub mod tests {
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", STORED | TEXT);
let schema = schema_builder.build();
let json_val: serde_json::Value = serde_json::from_str(
let json_val: serde_json::Map<String, serde_json::Value> = serde_json::from_str(
r#"{
"signed": 2,
"float": 2.0,
@@ -1025,16 +1034,13 @@ pub mod tests {
text_field => "some other value",
other_text_field => "short");
assert_eq!(document.len(), 3);
let values: Vec<OwnedValue> = document.get_all(text_field).map(OwnedValue::from).collect();
let values: Vec<&OwnedValue> = document.get_all(text_field).collect();
assert_eq!(values.len(), 2);
assert_eq!(values[0].as_ref().as_str(), Some("tantivy"));
assert_eq!(values[1].as_ref().as_str(), Some("some other value"));
let values: Vec<OwnedValue> = document
.get_all(other_text_field)
.map(OwnedValue::from)
.collect();
assert_eq!(values[0].as_str(), Some("tantivy"));
assert_eq!(values[1].as_str(), Some("some other value"));
let values: Vec<&OwnedValue> = document.get_all(other_text_field).collect();
assert_eq!(values.len(), 1);
assert_eq!(values[0].as_ref().as_str(), Some("short"));
assert_eq!(values[0].as_str(), Some("short"));
}
#[test]
@@ -1101,9 +1107,9 @@ pub mod tests {
#[test]
fn test_update_via_delete_insert() -> crate::Result<()> {
use crate::collector::Count;
use crate::index::SegmentId;
use crate::indexer::NoMergePolicy;
use crate::query::AllQuery;
use crate::SegmentId;
const DOC_COUNT: u64 = 2u64;

3
src/macros.rs
View File

@@ -41,7 +41,6 @@
/// );
/// # }
/// ```
#[macro_export]
macro_rules! doc(
() => {
@@ -53,7 +52,7 @@ macro_rules! doc(
{
let mut document = $crate::TantivyDocument::default();
$(
document.add_field_value($field, &$value);
document.add_field_value($field, $value);
)*
document
}

3
src/postings/json_postings_writer.rs
View File

@@ -1,6 +1,5 @@
use std::io;
use common::json_path_writer::JSON_END_OF_PATH;
use stacker::Addr;
use crate::indexer::doc_id_mapping::DocIdMapping;
@@ -8,7 +7,7 @@ use crate::indexer::path_to_unordered_id::OrderedPathId;
use crate::postings::postings_writer::SpecializedPostingsWriter;
use crate::postings::recorder::{BufferLender, DocIdRecorder, Recorder};
use crate::postings::{FieldSerializer, IndexingContext, IndexingPosition, PostingsWriter};
use crate::schema::{Field, Type};
use crate::schema::{Field, Type, JSON_END_OF_PATH};
use crate::tokenizer::TokenStream;
use crate::{DocId, Term};

2
src/postings/serializer.rs
View File

@@ -56,7 +56,7 @@ pub struct InvertedIndexSerializer {
impl InvertedIndexSerializer {
/// Open a new `InvertedIndexSerializer` for the given segment
pub fn open(segment: &mut Segment) -> crate::Result<InvertedIndexSerializer> {
use crate::index::SegmentComponent::{Positions, Postings, Terms};
use crate::SegmentComponent::{Positions, Postings, Terms};
let inv_index_serializer = InvertedIndexSerializer {
terms_write: CompositeWrite::wrap(segment.open_write(Terms)?),
postings_write: CompositeWrite::wrap(segment.open_write(Postings)?),

3
src/query/empty_query.rs
View File

@@ -1,9 +1,8 @@
use super::Scorer;
use crate::docset::TERMINATED;
use crate::index::SegmentReader;
use crate::query::explanation::does_not_match;
use crate::query::{EnableScoring, Explanation, Query, Weight};
use crate::{DocId, DocSet, Score, Searcher};
use crate::{DocId, DocSet, Score, Searcher, SegmentReader};
/// `EmptyQuery` is a dummy `Query` in which no document matches.
///

4
src/query/more_like_this/more_like_this.rs
View File

@@ -180,7 +180,7 @@ impl MoreLikeThis {
let facets: Vec<&str> = values
.iter()
.map(|value| {
value.as_facet().ok_or_else(|| {
value.as_facet().map(|f| f.encoded_str()).ok_or_else(|| {
TantivyError::InvalidArgument("invalid field value".to_string())
})
})
@@ -220,7 +220,7 @@ impl MoreLikeThis {
let mut token_stream = tokenizer.token_stream(text);
token_stream.process(sink);
} else if let Some(tok_str) = value.as_pre_tokenized_text() {
let mut token_stream = PreTokenizedStream::from(*tok_str.clone());
let mut token_stream = PreTokenizedStream::from(tok_str.clone());
token_stream.process(sink);
}
}

37
src/query/query_parser/query_parser.rs
View File

@@ -10,8 +10,10 @@ use query_grammar::{UserInputAst, UserInputBound, UserInputLeaf, UserInputLitera
use rustc_hash::FxHashMap;
use super::logical_ast::*;
use crate::core::json_utils::{
convert_to_fast_value_and_get_term, set_string_and_get_terms, JsonTermWriter,
};
use crate::index::Index;
use crate::json_utils::convert_to_fast_value_and_append_to_json_term;
use crate::query::range_query::{is_type_valid_for_fastfield_range_query, RangeQuery};
use crate::query::{
AllQuery, BooleanQuery, BoostQuery, EmptyQuery, FuzzyTermQuery, Occur, PhrasePrefixQuery,
@@ -963,27 +965,20 @@ fn generate_literals_for_json_object(
})?;
let index_record_option = text_options.index_option();
let mut logical_literals = Vec::new();
let get_term_with_path =
|| Term::from_field_json_path(field, json_path, json_options.is_expand_dots_enabled());
// Try to convert the phrase to a fast value
if let Some(term) = convert_to_fast_value_and_append_to_json_term(get_term_with_path(), phrase)
{
let mut term = Term::with_capacity(100);
let mut json_term_writer = JsonTermWriter::from_field_and_json_path(
field,
json_path,
json_options.is_expand_dots_enabled(),
&mut term,
);
if let Some(term) = convert_to_fast_value_and_get_term(&mut json_term_writer, phrase) {
logical_literals.push(LogicalLiteral::Term(term));
}
// Try to tokenize the phrase and create Terms.
let mut positions_and_terms = Vec::<(usize, Term)>::new();
let mut token_stream = text_analyzer.token_stream(phrase);
token_stream.process(&mut |token| {
let mut term = get_term_with_path();
term.append_type_and_str(&token.text);
positions_and_terms.push((token.position, term.clone()));
});
if positions_and_terms.len() <= 1 {
for (_, term) in positions_and_terms {
let terms = set_string_and_get_terms(&mut json_term_writer, phrase, &mut text_analyzer);
drop(json_term_writer);
if terms.len() <= 1 {
for (_, term) in terms {
logical_literals.push(LogicalLiteral::Term(term));
}
return Ok(logical_literals);
@@ -994,7 +989,7 @@ fn generate_literals_for_json_object(
));
}
logical_literals.push(LogicalLiteral::Phrase {
terms: positions_and_terms,
terms,
slop: 0,
prefix: false,
});

2
src/query/range_query/fast_field_range_query.rs
View File

@@ -174,7 +174,7 @@ impl<T: Send + Sync + PartialOrd + Copy + Debug + 'static> DocSet for RangeDocSe
}
fn size_hint(&self) -> u32 {
self.column.num_docs()
0 // heuristic possible by checking number of hits when fetching a block
}
}

4
src/query/term_query/term_scorer.rs
View File

@@ -127,7 +127,6 @@ impl Scorer for TermScorer {
mod tests {
use proptest::prelude::*;
use crate::index::SegmentId;
use crate::indexer::index_writer::MEMORY_BUDGET_NUM_BYTES_MIN;
use crate::merge_policy::NoMergePolicy;
use crate::postings::compression::COMPRESSION_BLOCK_SIZE;
@@ -135,7 +134,8 @@ mod tests {
use crate::query::{Bm25Weight, EnableScoring, Scorer, TermQuery};
use crate::schema::{IndexRecordOption, Schema, TEXT};
use crate::{
assert_nearly_equals, DocId, DocSet, Index, IndexWriter, Score, Searcher, Term, TERMINATED,
assert_nearly_equals, DocId, DocSet, Index, IndexWriter, Score, Searcher, SegmentId, Term,
TERMINATED,
};
#[test]

3
src/reader/warming.rs
View File

@@ -179,10 +179,9 @@ mod tests {
use super::Warmer;
use crate::core::searcher::SearcherGeneration;
use crate::directory::RamDirectory;
use crate::index::SegmentId;
use crate::indexer::index_writer::MEMORY_BUDGET_NUM_BYTES_MIN;
use crate::schema::{Schema, INDEXED};
use crate::{Index, IndexSettings, ReloadPolicy, Searcher};
use crate::{Index, IndexSettings, ReloadPolicy, Searcher, SegmentId};
#[derive(Default)]
struct TestWarmer {

2
src/schema/date_time_options.rs
View File

@@ -1,5 +1,7 @@
use std::ops::BitOr;
#[allow(deprecated)]
pub use common::DatePrecision;
pub use common::DateTimePrecision;
use serde::{Deserialize, Serialize};

54
src/schema/document/de.rs
View File

@@ -69,28 +69,6 @@ pub trait DocumentDeserialize: Sized {
where D: DocumentDeserializer<'de>;
}
/// A stateful extension of [`DocumentDeserialize`].
pub trait DocumentDeserializeSeed: Sized {
/// The type produced by using this seed.
type Value;
/// Attempts to deserialize `Self::Value` from the given `seed` and `deserializer`.
fn deserialize<'de, D>(self, deserializer: D) -> Result<Self::Value, DeserializeError>
where D: DocumentDeserializer<'de>;
}
impl<T> DocumentDeserializeSeed for PhantomData<T>
where T: DocumentDeserialize
{
/// The type produced by using this seed.
type Value = T;
fn deserialize<'de, D>(self, deserializer: D) -> Result<Self::Value, DeserializeError>
where D: DocumentDeserializer<'de> {
<T as DocumentDeserialize>::deserialize(deserializer)
}
}
/// A deserializer that can walk through each entry in the document.
pub trait DocumentDeserializer<'de> {
/// A indicator as to how many values are in the document.
@@ -182,7 +160,7 @@ pub enum ValueType {
/// A dynamic object value.
Object,
/// A JSON object value. Deprecated.
#[deprecated(note = "We keep this for backwards compatibility, use Object instead")]
#[deprecated]
JSONObject,
}
@@ -895,7 +873,7 @@ mod tests {
);
let facet = Facet::from_text("/hello/world").unwrap();
let result = serialize_value(ReferenceValueLeaf::Facet(facet.encoded_str()).into());
let result = serialize_value(ReferenceValueLeaf::Facet(&facet).into());
let value = deserialize_value(result);
assert_eq!(value, crate::schema::OwnedValue::Facet(facet));
@@ -903,8 +881,7 @@ mod tests {
text: "hello, world".to_string(),
tokens: vec![Token::default(), Token::default()],
};
let result =
serialize_value(ReferenceValueLeaf::PreTokStr(pre_tok_str.clone().into()).into());
let result = serialize_value(ReferenceValueLeaf::PreTokStr(&pre_tok_str).into());
let value = deserialize_value(result);
assert_eq!(value, crate::schema::OwnedValue::PreTokStr(pre_tok_str));
}
@@ -983,19 +960,13 @@ mod tests {
"my-third-key".to_string(),
crate::schema::OwnedValue::F64(123.0),
);
assert_eq!(
value,
crate::schema::OwnedValue::Object(expected_object.into_iter().collect())
);
assert_eq!(value, crate::schema::OwnedValue::Object(expected_object));
let object = serde_json::Map::new();
let result = serialize_value(ReferenceValue::Object(JsonObjectIter(object.iter())));
let value = deserialize_value(result);
let expected_object = BTreeMap::new();
assert_eq!(
value,
crate::schema::OwnedValue::Object(expected_object.into_iter().collect())
);
assert_eq!(value, crate::schema::OwnedValue::Object(expected_object));
let mut object = serde_json::Map::new();
object.insert("my-first-key".into(), serde_json::Value::Null);
@@ -1007,10 +978,7 @@ mod tests {
expected_object.insert("my-first-key".to_string(), crate::schema::OwnedValue::Null);
expected_object.insert("my-second-key".to_string(), crate::schema::OwnedValue::Null);
expected_object.insert("my-third-key".to_string(), crate::schema::OwnedValue::Null);
assert_eq!(
value,
crate::schema::OwnedValue::Object(expected_object.into_iter().collect())
);
assert_eq!(value, crate::schema::OwnedValue::Object(expected_object));
}
#[test]
@@ -1087,10 +1055,7 @@ mod tests {
.collect(),
),
);
assert_eq!(
value,
crate::schema::OwnedValue::Object(expected_object.into_iter().collect())
);
assert_eq!(value, crate::schema::OwnedValue::Object(expected_object));
// Some more extreme nesting that might behave weirdly
let mut object = serde_json::Map::new();
@@ -1112,9 +1077,6 @@ mod tests {
OwnedValue::Array(vec![OwnedValue::Null]),
])]),
);
assert_eq!(
value,
OwnedValue::Object(expected_object.into_iter().collect())
);
assert_eq!(value, OwnedValue::Object(expected_object));
}
}

713
src/schema/document/default_document.rs
View File

@@ -1,64 +1,124 @@
use std::collections::{BTreeMap, HashMap, HashSet};
use std::io::{self, Read, Write};
use std::net::Ipv6Addr;
use columnar::MonotonicallyMappableToU128;
use common::{read_u32_vint_no_advance, serialize_vint_u32, BinarySerializable, DateTime, VInt};
use common::{BinarySerializable, DateTime, VInt};
use serde_json::Map;
pub use CompactDoc as TantivyDocument;
use super::{ReferenceValue, ReferenceValueLeaf, Value};
use crate::schema::document::se::BinaryValueSerializer;
use crate::schema::document::{
DeserializeError, Document, DocumentDeserialize, DocumentDeserializer,
BinaryDocumentDeserializer, BinaryDocumentSerializer, DeserializeError, Document, DocumentDeserialize, DocumentDeserializer, ReferenceValue, ReferenceValueLeaf
};
use crate::schema::field_type::ValueParsingError;
use crate::schema::{Facet, Field, NamedFieldDocument, OwnedValue, Schema};
use crate::schema::field_value::FieldValueIter;
use crate::schema::{Facet, Field, FieldValue, NamedFieldDocument, OwnedValue, Schema, Value};
use crate::tokenizer::PreTokenizedString;
#[repr(packed)]
#[derive(Debug, Clone)]
/// A field value pair in the compact tantivy document
struct FieldValueAddr {
pub field: u16,
pub value_addr: ValueAddr,
/// TantivyDocument provides a default implementation of the `Document` trait.
/// It is the object that can be indexed and then searched for.
///
/// Documents are fundamentally a collection of unordered couples `(field, value)`.
/// In this list, one field may appear more than once.
#[derive(Clone, Debug, serde::Serialize, serde::Deserialize, Default)]
pub struct TantivyDocument {
field_values: Vec<FieldValue>,
}
#[derive(Debug, Clone)]
/// The default document in tantivy. It encodes data in a compact form.
pub struct CompactDoc {
/// `node_data` is a vec of bytes, where each value is serialized into bytes and stored. It
/// includes all the data of the document and also metadata like where the nodes are located
/// in an object or array.
pub node_data: Vec<u8>,
/// The root (Field, Value) pairs
field_values: Vec<FieldValueAddr>,
}
impl Document for TantivyDocument {
type Value<'a> = &'a OwnedValue;
type FieldsValuesIter<'a> = FieldValueIter<'a>;
impl Default for CompactDoc {
fn default() -> Self {
Self::new()
fn iter_fields_and_values(&self) -> Self::FieldsValuesIter<'_> {
FieldValueIter(self.field_values.iter())
}
}
impl CompactDoc {
/// Creates a new, empty document object
/// The reserved capacity is for the total serialized data
pub fn with_capacity(bytes: usize) -> CompactDoc {
CompactDoc {
node_data: Vec::with_capacity(bytes),
field_values: Vec::with_capacity(4),
impl DocumentDeserialize for TantivyDocument {
fn deserialize<'de, D>(mut deserializer: D) -> Result<Self, DeserializeError>
where D: DocumentDeserializer<'de> {
let mut field_values = Vec::with_capacity(deserializer.size_hint());
while let Some((field, value)) = deserializer.next_field()? {
field_values.push(FieldValue::new(field, value));
}
Ok(Self { field_values })
}
}
impl From<Vec<FieldValue>> for TantivyDocument {
fn from(field_values: Vec<FieldValue>) -> Self {
Self { field_values }
}
}
impl PartialEq for TantivyDocument {
fn eq(&self, other: &Self) -> bool {
// super slow, but only here for tests
let convert_to_comparable_map = |field_values: &[FieldValue]| {
let mut field_value_set: HashMap<Field, HashSet<String>> = Default::default();
for field_value in field_values.iter() {
let value = serde_json::to_string(field_value.value()).unwrap();
field_value_set
.entry(field_value.field())
.or_default()
.insert(value);
}
field_value_set
};
let self_field_values: HashMap<Field, HashSet<String>> =
convert_to_comparable_map(&self.field_values);
let other_field_values: HashMap<Field, HashSet<String>> =
convert_to_comparable_map(&other.field_values);
self_field_values.eq(&other_field_values)
}
}
impl Eq for TantivyDocument {}
impl IntoIterator for TantivyDocument {
type Item = FieldValue;
type IntoIter = std::vec::IntoIter<FieldValue>;
fn into_iter(self) -> Self::IntoIter {
self.field_values.into_iter()
}
}
impl TantivyDocument {
pub fn to_bytes(&self, buffer: &mut Vec<u8>) -> std::io::Result<()> {
buffer.clear();
let num_field_values = self.field_values.len();
VInt(num_field_values as u64).serialize(buffer)?;
for (field, value_access) in self.iter_fields_and_values() {
field.serialize(buffer)?;
let mut serializer = BinaryValueSerializer::new(buffer);
match value_access.as_value() {
ReferenceValue::Leaf(ReferenceValueLeaf::PreTokStr(pre_tokenized_text)) => {
serializer.serialize_value(ReferenceValue::Leaf::<&'_ OwnedValue>(
ReferenceValueLeaf::Str(&pre_tokenized_text.text),
))?;
}
_ => {
serializer.serialize_value(value_access.as_value())?;
}
}
}
Ok(())
}
pub fn from_bytes(mut payload: &[u8]) -> Self {
let deserializer = BinaryDocumentDeserializer::from_reader(&mut payload).unwrap();
Self::deserialize(deserializer).unwrap()
}
/// Creates a new, empty document object
pub fn new() -> CompactDoc {
CompactDoc::with_capacity(1024)
}
/// Skrinks the capacity of the document to fit the data
pub fn shrink_to_fit(&mut self) {
self.node_data.shrink_to_fit();
self.field_values.shrink_to_fit();
pub fn new() -> TantivyDocument {
TantivyDocument::default()
}
/// Returns the length of the document.
@@ -70,116 +130,83 @@ impl CompactDoc {
pub fn add_facet<F>(&mut self, field: Field, path: F)
where Facet: From<F> {
let facet = Facet::from(path);
self.add_leaf_field_value(field, ReferenceValueLeaf::Facet(facet.encoded_str()));
let value = OwnedValue::Facet(facet);
self.add_field_value(field, value);
}
/// Add a text field.
pub fn add_text<S: AsRef<str>>(&mut self, field: Field, text: S) {
self.add_leaf_field_value(field, ReferenceValueLeaf::Str(text.as_ref()));
pub fn add_text<S: ToString>(&mut self, field: Field, text: S) {
let value = OwnedValue::Str(text.to_string());
self.add_field_value(field, value);
}
/// Add a pre-tokenized text field.
pub fn add_pre_tokenized_text(&mut self, field: Field, pre_tokenized_text: PreTokenizedString) {
self.add_leaf_field_value(field, pre_tokenized_text);
self.add_field_value(field, pre_tokenized_text);
}
/// Add a u64 field
pub fn add_u64(&mut self, field: Field, value: u64) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a IP address field. Internally only Ipv6Addr is used.
pub fn add_ip_addr(&mut self, field: Field, value: Ipv6Addr) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a i64 field
pub fn add_i64(&mut self, field: Field, value: i64) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a f64 field
pub fn add_f64(&mut self, field: Field, value: f64) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a bool field
pub fn add_bool(&mut self, field: Field, value: bool) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a date field with unspecified time zone offset
pub fn add_date(&mut self, field: Field, value: DateTime) {
self.add_leaf_field_value(field, value);
self.add_field_value(field, value);
}
/// Add a bytes field
pub fn add_bytes(&mut self, field: Field, value: &[u8]) {
self.add_leaf_field_value(field, value);
pub fn add_bytes<T: Into<Vec<u8>>>(&mut self, field: Field, value: T) {
self.add_field_value(field, value.into());
}
/// Add a dynamic object field
pub fn add_object(&mut self, field: Field, object: BTreeMap<String, OwnedValue>) {
self.add_field_value(field, &OwnedValue::from(object));
self.add_field_value(field, object);
}
/// Add a (field, value) to the document.
///
/// `OwnedValue` implements Value, which should be easiest to use, but is not the most
/// performant.
pub fn add_field_value<'a, V: Value<'a>>(&mut self, field: Field, value: V) {
let field_value = FieldValueAddr {
field: field
.field_id()
.try_into()
.expect("support only up to u16::MAX field ids"),
value_addr: self.add_value(value),
};
self.field_values.push(field_value);
}
/// Add a (field, leaf value) to the document.
/// Leaf values don't have nested values.
pub fn add_leaf_field_value<'a, T: Into<ReferenceValueLeaf<'a>>>(
&mut self,
field: Field,
typed_val: T,
) {
pub fn add_field_value<T: Into<OwnedValue>>(&mut self, field: Field, typed_val: T) {
let value = typed_val.into();
let field_value = FieldValueAddr {
field: field
.field_id()
.try_into()
.expect("support only up to u16::MAX field ids"),
value_addr: self.add_value_leaf(value),
};
let field_value = FieldValue { field, value };
self.field_values.push(field_value);
}
/// field_values accessor
pub fn field_values(
&self,
) -> impl Iterator<Item = (Field, ReferenceValue<'_, CompactDocValue<'_>>)> {
self.field_values.iter().map(|field_val| {
let field = Field::from_field_id(field_val.field as u32);
let val = self.extract_value(field_val.value_addr).unwrap();
(field, val)
})
pub fn field_values(&self) -> &[FieldValue] {
&self.field_values
}
/// Returns all of the `ReferenceValue`s associated the given field
pub fn get_all(
&self,
field: Field,
) -> impl Iterator<Item = ReferenceValue<'_, CompactDocValue<'_>>> + '_ {
/// Returns all of the `FieldValue`s associated the given field
pub fn get_all(&self, field: Field) -> impl Iterator<Item = &OwnedValue> {
self.field_values
.iter()
.filter(move |field_value| Field::from_field_id(field_value.field as u32) == field)
.map(|val| self.extract_value(val.value_addr).unwrap())
.filter(move |field_value| field_value.field() == field)
.map(FieldValue::value)
}
/// Returns the first `ReferenceValue` associated the given field
pub fn get_first(&self, field: Field) -> Option<ReferenceValue<'_, CompactDocValue<'_>>> {
/// Returns the first `FieldValue` associated the given field
pub fn get_first(&self, field: Field) -> Option<&OwnedValue> {
self.get_all(field).next()
}
@@ -187,12 +214,12 @@ impl CompactDoc {
pub fn convert_named_doc(
schema: &Schema,
named_doc: NamedFieldDocument,
) -> Result<Self, DocParsingError> {
let mut document = Self::new();
) -> Result<TantivyDocument, DocParsingError> {
let mut document = TantivyDocument::new();
for (field_name, values) in named_doc.0 {
if let Ok(field) = schema.get_field(&field_name) {
for value in values {
document.add_field_value(field, &value);
document.add_field_value(field, value);
}
}
}
@@ -200,7 +227,7 @@ impl CompactDoc {
}
/// Build a document object from a json-object.
pub fn parse_json(schema: &Schema, doc_json: &str) -> Result<Self, DocParsingError> {
pub fn parse_json(schema: &Schema, doc_json: &str) -> Result<TantivyDocument, DocParsingError> {
let json_obj: Map<String, serde_json::Value> =
serde_json::from_str(doc_json).map_err(|_| DocParsingError::invalid_json(doc_json))?;
Self::from_json_object(schema, json_obj)
@@ -210,8 +237,8 @@ impl CompactDoc {
pub fn from_json_object(
schema: &Schema,
json_obj: Map<String, serde_json::Value>,
) -> Result<Self, DocParsingError> {
let mut doc = Self::default();
) -> Result<TantivyDocument, DocParsingError> {
let mut doc = TantivyDocument::default();
for (field_name, json_value) in json_obj {
if let Ok(field) = schema.get_field(&field_name) {
let field_entry = schema.get_field_entry(field);
@@ -222,457 +249,20 @@ impl CompactDoc {
let value = field_type
.value_from_json(json_item)
.map_err(|e| DocParsingError::ValueError(field_name.clone(), e))?;
doc.add_field_value(field, &value);
doc.add_field_value(field, value);
}
}
_ => {
let value = field_type
.value_from_json(json_value)
.map_err(|e| DocParsingError::ValueError(field_name.clone(), e))?;
doc.add_field_value(field, &value);
doc.add_field_value(field, value);
}
}
}
}
Ok(doc)
}
fn add_value_leaf(&mut self, leaf: ReferenceValueLeaf) -> ValueAddr {
let type_id = ValueType::from(&leaf);
// Write into `node_data` and return u32 position as its address
// Null and bool are inlined into the address
let val_addr = match leaf {
ReferenceValueLeaf::Null => 0,
ReferenceValueLeaf::Str(bytes) => {
write_bytes_into(&mut self.node_data, bytes.as_bytes())
}
ReferenceValueLeaf::Facet(bytes) => {
write_bytes_into(&mut self.node_data, bytes.as_bytes())
}
ReferenceValueLeaf::Bytes(bytes) => write_bytes_into(&mut self.node_data, bytes),
ReferenceValueLeaf::U64(num) => write_into(&mut self.node_data, num),
ReferenceValueLeaf::I64(num) => write_into(&mut self.node_data, num),
ReferenceValueLeaf::F64(num) => write_into(&mut self.node_data, num),
ReferenceValueLeaf::Bool(b) => b as u32,
ReferenceValueLeaf::Date(date) => {
write_into(&mut self.node_data, date.into_timestamp_nanos())
}
ReferenceValueLeaf::IpAddr(num) => write_into(&mut self.node_data, num.to_u128()),
ReferenceValueLeaf::PreTokStr(pre_tok) => write_into(&mut self.node_data, *pre_tok),
};
ValueAddr { type_id, val_addr }
}
/// Adds a value and returns in address into the
fn add_value<'a, V: Value<'a>>(&mut self, value: V) -> ValueAddr {
let value = value.as_value();
let type_id = ValueType::from(&value);
match value {
ReferenceValue::Leaf(leaf) => self.add_value_leaf(leaf),
ReferenceValue::Array(elements) => {
// addresses of the elements in node_data
// Reusing a vec would be nicer, but it's not easy because of the recursion
// A global vec would work if every writer get it's discriminator
let mut addresses = Vec::new();
for elem in elements {
let value_addr = self.add_value(elem);
write_into(&mut addresses, value_addr);
}
ValueAddr {
type_id,
val_addr: write_bytes_into(&mut self.node_data, &addresses),
}
}
ReferenceValue::Object(entries) => {
// addresses of the elements in node_data
let mut addresses = Vec::new();
for (key, value) in entries {
let key_addr = self.add_value_leaf(ReferenceValueLeaf::Str(key));
let value_addr = self.add_value(value);
write_into(&mut addresses, key_addr);
write_into(&mut addresses, value_addr);
}
ValueAddr {
type_id,
val_addr: write_bytes_into(&mut self.node_data, &addresses),
}
}
}
}
fn extract_value(
&self,
ref_value: ValueAddr,
) -> io::Result<ReferenceValue<'_, CompactDocValue<'_>>> {
match ref_value.type_id {
ValueType::Null => Ok(ReferenceValueLeaf::Null.into()),
ValueType::Str => {
let str_ref = self.extract_str(ref_value.val_addr);
Ok(ReferenceValueLeaf::Str(str_ref).into())
}
ValueType::Facet => {
let str_ref = self.extract_str(ref_value.val_addr);
Ok(ReferenceValueLeaf::Facet(str_ref).into())
}
ValueType::Bytes => {
let data = self.extract_bytes(ref_value.val_addr);
Ok(ReferenceValueLeaf::Bytes(data).into())
}
ValueType::U64 => self
.read_from::<u64>(ref_value.val_addr)
.map(ReferenceValueLeaf::U64)
.map(Into::into),
ValueType::I64 => self
.read_from::<i64>(ref_value.val_addr)
.map(ReferenceValueLeaf::I64)
.map(Into::into),
ValueType::F64 => self
.read_from::<f64>(ref_value.val_addr)
.map(ReferenceValueLeaf::F64)
.map(Into::into),
ValueType::Bool => Ok(ReferenceValueLeaf::Bool(ref_value.val_addr != 0).into()),
ValueType::Date => self
.read_from::<i64>(ref_value.val_addr)
.map(|ts| ReferenceValueLeaf::Date(DateTime::from_timestamp_nanos(ts)))
.map(Into::into),
ValueType::IpAddr => self
.read_from::<u128>(ref_value.val_addr)
.map(|num| ReferenceValueLeaf::IpAddr(Ipv6Addr::from_u128(num)))
.map(Into::into),
ValueType::PreTokStr => self
.read_from::<PreTokenizedString>(ref_value.val_addr)
.map(Into::into)
.map(ReferenceValueLeaf::PreTokStr)
.map(Into::into),
ValueType::Object => Ok(ReferenceValue::Object(CompactDocObjectIter::new(
self,
ref_value.val_addr,
)?)),
ValueType::Array => Ok(ReferenceValue::Array(CompactDocArrayIter::new(
self,
ref_value.val_addr,
)?)),
}
}
/// get &[u8] reference from node_data
fn extract_bytes(&self, addr: Addr) -> &[u8] {
binary_deserialize_bytes(self.get_slice(addr))
}
/// get &str reference from node_data
fn extract_str(&self, addr: Addr) -> &str {
let data = self.extract_bytes(addr);
// Utf-8 checks would have a noticeable performance overhead here
unsafe { std::str::from_utf8_unchecked(data) }
}
/// deserialized owned value from node_data
fn read_from<T: BinarySerializable>(&self, addr: Addr) -> io::Result<T> {
let data_slice = &self.node_data[addr as usize..];
let mut cursor = std::io::Cursor::new(data_slice);
T::deserialize(&mut cursor)
}
/// get slice from address. The returned slice is open ended
fn get_slice(&self, addr: Addr) -> &[u8] {
&self.node_data[addr as usize..]
}
}
/// BinarySerializable alternative to read references
fn binary_deserialize_bytes(data: &[u8]) -> &[u8] {
let (len, bytes_read) = read_u32_vint_no_advance(data);
&data[bytes_read..bytes_read + len as usize]
}
/// Write bytes and return the position of the written data.
///
/// BinarySerializable alternative to write references
fn write_bytes_into(vec: &mut Vec<u8>, data: &[u8]) -> u32 {
let pos = vec.len() as u32;
let mut buf = [0u8; 8];
let len_vint_bytes = serialize_vint_u32(data.len() as u32, &mut buf);
vec.extend_from_slice(len_vint_bytes);
vec.extend_from_slice(data);
pos
}
/// Serialize and return the position
fn write_into<T: BinarySerializable>(vec: &mut Vec<u8>, value: T) -> u32 {
let pos = vec.len() as u32;
value.serialize(vec).unwrap();
pos
}
impl PartialEq for CompactDoc {
fn eq(&self, other: &Self) -> bool {
// super slow, but only here for tests
let convert_to_comparable_map = |doc: &CompactDoc| {
let mut field_value_set: HashMap<Field, HashSet<String>> = Default::default();
for field_value in doc.field_values.iter() {
let value: OwnedValue = doc.extract_value(field_value.value_addr).unwrap().into();
let value = serde_json::to_string(&value).unwrap();
field_value_set
.entry(Field::from_field_id(field_value.field as u32))
.or_default()
.insert(value);
}
field_value_set
};
let self_field_values: HashMap<Field, HashSet<String>> = convert_to_comparable_map(self);
let other_field_values: HashMap<Field, HashSet<String>> = convert_to_comparable_map(other);
self_field_values.eq(&other_field_values)
}
}
impl Eq for CompactDoc {}
impl DocumentDeserialize for CompactDoc {
fn deserialize<'de, D>(mut deserializer: D) -> Result<Self, DeserializeError>
where D: DocumentDeserializer<'de> {
let mut doc = CompactDoc::default();
// TODO: Deserializing into OwnedValue is wasteful. The deserializer should be able to work
// on slices and referenced data.
while let Some((field, value)) = deserializer.next_field::<OwnedValue>()? {
doc.add_field_value(field, &value);
}
Ok(doc)
}
}
/// A value of Compact Doc needs a reference to the container to extract its payload
#[derive(Debug, Clone, Copy)]
pub struct CompactDocValue<'a> {
container: &'a CompactDoc,
value: ValueAddr,
}
impl<'a> Value<'a> for CompactDocValue<'a> {
type ArrayIter = CompactDocArrayIter<'a>;
type ObjectIter = CompactDocObjectIter<'a>;
fn as_value(&self) -> ReferenceValue<'a, Self> {
self.container.extract_value(self.value).unwrap()
}
}
/// The address in the vec
type Addr = u32;
#[derive(Clone, Copy, Default)]
#[repr(packed)]
/// The value type and the address to its payload in the container.
struct ValueAddr {
type_id: ValueType,
/// This is the address to the value in the vec, except for bool and null, which are inlined
val_addr: Addr,
}
impl BinarySerializable for ValueAddr {
fn serialize<W: Write + ?Sized>(&self, writer: &mut W) -> io::Result<()> {
self.type_id.serialize(writer)?;
VInt(self.val_addr as u64).serialize(writer)
}
fn deserialize<R: Read>(reader: &mut R) -> io::Result<Self> {
let type_id = ValueType::deserialize(reader)?;
let val_addr = VInt::deserialize(reader)?.0 as u32;
Ok(ValueAddr { type_id, val_addr })
}
}
impl std::fmt::Debug for ValueAddr {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
let val_addr = self.val_addr;
f.write_fmt(format_args!("{:?} at {:?}", self.type_id, val_addr))
}
}
/// A enum representing a value for tantivy to index.
///
/// Any changes need to be reflected in `BinarySerializable` for `ValueType`
///
/// We can't use [schema::Type] or [columnar::ColumnType] here, because they are missing
/// some items like Array and PreTokStr.
#[derive(Default, Clone, Copy, Debug, PartialEq)]
#[repr(u8)]
pub enum ValueType {
/// A null value.
#[default]
Null = 0,
/// The str type is used for any text information.
Str = 1,
/// Unsigned 64-bits Integer `u64`
U64 = 2,
/// Signed 64-bits Integer `i64`
I64 = 3,
/// 64-bits Float `f64`
F64 = 4,
/// Date/time with nanoseconds precision
Date = 5,
/// Facet
Facet = 6,
/// Arbitrarily sized byte array
Bytes = 7,
/// IpV6 Address. Internally there is no IpV4, it needs to be converted to `Ipv6Addr`.
IpAddr = 8,
/// Bool value
Bool = 9,
/// Pre-tokenized str type,
PreTokStr = 10,
/// Object
Object = 11,
/// Pre-tokenized str type,
Array = 12,
}
impl BinarySerializable for ValueType {
fn serialize<W: Write + ?Sized>(&self, writer: &mut W) -> io::Result<()> {
(*self as u8).serialize(writer)?;
Ok(())
}
fn deserialize<R: Read>(reader: &mut R) -> io::Result<Self> {
let num = u8::deserialize(reader)?;
let type_id = if (0..=12).contains(&num) {
unsafe { std::mem::transmute(num) }
} else {
return Err(io::Error::new(
io::ErrorKind::InvalidData,
format!("Invalid value type id: {}", num),
));
};
Ok(type_id)
}
}
impl<'a, V: Value<'a>> From<&ReferenceValue<'a, V>> for ValueType {
fn from(value: &ReferenceValue<'a, V>) -> Self {
match value {
ReferenceValue::Leaf(leaf) => leaf.into(),
ReferenceValue::Array(_) => ValueType::Array,
ReferenceValue::Object(_) => ValueType::Object,
}
}
}
impl<'a> From<&ReferenceValueLeaf<'a>> for ValueType {
fn from(value: &ReferenceValueLeaf<'a>) -> Self {
match value {
ReferenceValueLeaf::Null => ValueType::Null,
ReferenceValueLeaf::Str(_) => ValueType::Str,
ReferenceValueLeaf::U64(_) => ValueType::U64,
ReferenceValueLeaf::I64(_) => ValueType::I64,
ReferenceValueLeaf::F64(_) => ValueType::F64,
ReferenceValueLeaf::Bool(_) => ValueType::Bool,
ReferenceValueLeaf::Date(_) => ValueType::Date,
ReferenceValueLeaf::IpAddr(_) => ValueType::IpAddr,
ReferenceValueLeaf::PreTokStr(_) => ValueType::PreTokStr,
ReferenceValueLeaf::Facet(_) => ValueType::Facet,
ReferenceValueLeaf::Bytes(_) => ValueType::Bytes,
}
}
}
#[derive(Debug, Clone)]
/// The Iterator for the object values in the compact document
pub struct CompactDocObjectIter<'a> {
container: &'a CompactDoc,
node_addresses_slice: &'a [u8],
}
impl<'a> CompactDocObjectIter<'a> {
fn new(container: &'a CompactDoc, addr: Addr) -> io::Result<Self> {
// Objects are `&[ValueAddr]` serialized into bytes
let node_addresses_slice = container.extract_bytes(addr);
Ok(Self {
container,
node_addresses_slice,
})
}
}
impl<'a> Iterator for CompactDocObjectIter<'a> {
type Item = (&'a str, CompactDocValue<'a>);
fn next(&mut self) -> Option<Self::Item> {
if self.node_addresses_slice.is_empty() {
return None;
}
let key_addr = ValueAddr::deserialize(&mut self.node_addresses_slice).ok()?;
let key = self.container.extract_str(key_addr.val_addr);
let value = ValueAddr::deserialize(&mut self.node_addresses_slice).ok()?;
let value = CompactDocValue {
container: self.container,
value,
};
Some((key, value))
}
}
#[derive(Debug, Clone)]
/// The Iterator for the array values in the compact document
pub struct CompactDocArrayIter<'a> {
container: &'a CompactDoc,
node_addresses_slice: &'a [u8],
}
impl<'a> CompactDocArrayIter<'a> {
fn new(container: &'a CompactDoc, addr: Addr) -> io::Result<Self> {
// Arrays are &[ValueAddr] serialized into bytes
let node_addresses_slice = container.extract_bytes(addr);
Ok(Self {
container,
node_addresses_slice,
})
}
}
impl<'a> Iterator for CompactDocArrayIter<'a> {
type Item = CompactDocValue<'a>;
fn next(&mut self) -> Option<Self::Item> {
if self.node_addresses_slice.is_empty() {
return None;
}
let value = ValueAddr::deserialize(&mut self.node_addresses_slice).ok()?;
let value = CompactDocValue {
container: self.container,
value,
};
Some(value)
}
}
impl Document for CompactDoc {
type Value<'a> = CompactDocValue<'a>;
type FieldsValuesIter<'a> = FieldValueIterRef<'a>;
fn iter_fields_and_values(&self) -> Self::FieldsValuesIter<'_> {
FieldValueIterRef {
slice: self.field_values.iter(),
container: self,
}
}
}
/// A helper wrapper for creating an iterator over the field values
pub struct FieldValueIterRef<'a> {
slice: std::slice::Iter<'a, FieldValueAddr>,
container: &'a CompactDoc,
}
impl<'a> Iterator for FieldValueIterRef<'a> {
type Item = (Field, CompactDocValue<'a>);
fn next(&mut self) -> Option<Self::Item> {
self.slice.next().map(|field_value| {
(
Field::from_field_id(field_value.field as u32),
CompactDocValue::<'a> {
container: self.container,
value: field_value.value_addr,
},
)
})
}
}
/// Error that may happen when deserializing
@@ -705,40 +295,7 @@ mod tests {
let text_field = schema_builder.add_text_field("title", TEXT);
let mut doc = TantivyDocument::default();
doc.add_text(text_field, "My title");
assert_eq!(doc.field_values().count(), 1);
let schema = schema_builder.build();
let _val = doc.get_first(text_field).unwrap();
let _json = doc.to_named_doc(&schema);
}
#[test]
fn test_json_value() {
let json_str = r#"{
"toto": "titi",
"float": -0.2,
"bool": true,
"unsigned": 1,
"signed": -2,
"complexobject": {
"field.with.dot": 1
},
"date": "1985-04-12T23:20:50.52Z",
"my_arr": [2, 3, {"my_key": "two tokens"}, 4, {"nested_array": [2, 5, 6, [7, 8, {"a": [{"d": {"e":[99]}}, 9000]}, 9, 10], [5, 5]]}]
}"#;
let json_val: std::collections::BTreeMap<String, OwnedValue> =
serde_json::from_str(json_str).unwrap();
let mut schema_builder = Schema::builder();
let json_field = schema_builder.add_json_field("json", TEXT);
let mut doc = TantivyDocument::default();
doc.add_object(json_field, json_val);
let schema = schema_builder.build();
let json = doc.to_json(&schema);
let actual_json: serde_json::Value = serde_json::from_str(&json).unwrap();
let expected_json: serde_json::Value = serde_json::from_str(json_str).unwrap();
assert_eq!(actual_json["json"][0], expected_json);
assert_eq!(doc.field_values().len(), 1);
}
// TODO: Should this be re-added with the serialize method

152
src/schema/document/existing_type_impls.rs
View File

@@ -5,39 +5,21 @@
//! and don't care about some of the more specialised types or only want to customise
//! part of the document structure.
use std::collections::{btree_map, hash_map, BTreeMap, HashMap};
use std::iter::Empty;
use std::net::Ipv6Addr;
use common::DateTime;
use serde_json::Number;
use time::format_description::well_known::Rfc3339;
use time::OffsetDateTime;
use super::facet::Facet;
use super::ReferenceValueLeaf;
use crate::schema::document::{
ArrayAccess, DeserializeError, Document, DocumentDeserialize, DocumentDeserializer,
ObjectAccess, ReferenceValue, Value, ValueDeserialize, ValueDeserializer, ValueVisitor,
};
use crate::schema::Field;
use crate::tokenizer::PreTokenizedString;
// Serde compatibility support.
pub fn can_be_rfc3339_date_time(text: &str) -> bool {
if let Some(&first_byte) = text.as_bytes().first() {
if first_byte.is_ascii_digit() {
return true;
}
}
false
}
impl<'a> Value<'a> for &'a serde_json::Value {
type ArrayIter = std::slice::Iter<'a, serde_json::Value>;
type ObjectIter = JsonObjectIter<'a>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
match self {
serde_json::Value::Null => ReferenceValueLeaf::Null.into(),
@@ -53,19 +35,7 @@ impl<'a> Value<'a> for &'a serde_json::Value {
panic!("Unsupported serde_json number {number}");
}
}
serde_json::Value::String(text) => {
if can_be_rfc3339_date_time(text) {
match OffsetDateTime::parse(text, &Rfc3339) {
Ok(dt) => {
let dt_utc = dt.to_offset(time::UtcOffset::UTC);
ReferenceValueLeaf::Date(DateTime::from_utc(dt_utc)).into()
}
Err(_) => ReferenceValueLeaf::Str(text).into(),
}
} else {
ReferenceValueLeaf::Str(text).into()
}
}
serde_json::Value::String(val) => ReferenceValueLeaf::Str(val).into(),
serde_json::Value::Array(elements) => ReferenceValue::Array(elements.iter()),
serde_json::Value::Object(object) => {
ReferenceValue::Object(JsonObjectIter(object.iter()))
@@ -74,126 +44,6 @@ impl<'a> Value<'a> for &'a serde_json::Value {
}
}
impl<'a> Value<'a> for &'a String {
type ArrayIter = Empty<&'a String>;
type ObjectIter = Empty<(&'a str, &'a String)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Str(self))
}
}
impl<'a> Value<'a> for &'a Facet {
type ArrayIter = Empty<&'a Facet>;
type ObjectIter = Empty<(&'a str, &'a Facet)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Facet(self.encoded_str()))
}
}
impl<'a> Value<'a> for &'a u64 {
type ArrayIter = Empty<&'a u64>;
type ObjectIter = Empty<(&'a str, &'a u64)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::U64(**self))
}
}
impl<'a> Value<'a> for &'a i64 {
type ArrayIter = Empty<&'a i64>;
type ObjectIter = Empty<(&'a str, &'a i64)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::I64(**self))
}
}
impl<'a> Value<'a> for &'a f64 {
type ArrayIter = Empty<&'a f64>;
type ObjectIter = Empty<(&'a str, &'a f64)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::F64(**self))
}
}
impl<'a> Value<'a> for &'a bool {
type ArrayIter = Empty<&'a bool>;
type ObjectIter = Empty<(&'a str, &'a bool)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Bool(**self))
}
}
impl<'a> Value<'a> for &'a str {
type ArrayIter = Empty<&'a str>;
type ObjectIter = Empty<(&'a str, &'a str)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Str(self))
}
}
impl<'a> Value<'a> for &'a &'a str {
type ArrayIter = Empty<&'a &'a str>;
type ObjectIter = Empty<(&'a str, &'a &'a str)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Str(self))
}
}
impl<'a> Value<'a> for &'a [u8] {
type ArrayIter = Empty<&'a [u8]>;
type ObjectIter = Empty<(&'a str, &'a [u8])>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Bytes(self))
}
}
impl<'a> Value<'a> for &'a &'a [u8] {
type ArrayIter = Empty<&'a &'a [u8]>;
type ObjectIter = Empty<(&'a str, &'a &'a [u8])>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Bytes(self))
}
}
impl<'a> Value<'a> for &'a Vec<u8> {
type ArrayIter = Empty<&'a Vec<u8>>;
type ObjectIter = Empty<(&'a str, &'a Vec<u8>)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Bytes(self))
}
}
impl<'a> Value<'a> for &'a DateTime {
type ArrayIter = Empty<&'a DateTime>;
type ObjectIter = Empty<(&'a str, &'a DateTime)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::Date(**self))
}
}
impl<'a> Value<'a> for &'a Ipv6Addr {
type ArrayIter = Empty<&'a Ipv6Addr>;
type ObjectIter = Empty<(&'a str, &'a Ipv6Addr)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::IpAddr(**self))
}
}
impl<'a> Value<'a> for &'a PreTokenizedString {
type ArrayIter = Empty<&'a PreTokenizedString>;
type ObjectIter = Empty<(&'a str, &'a PreTokenizedString)>;
#[inline]
fn as_value(&self) -> ReferenceValue<'a, Self> {
ReferenceValue::Leaf(ReferenceValueLeaf::PreTokStr(Box::new((*self).clone())))
}
}
impl ValueDeserialize for serde_json::Value {
fn deserialize<'de, D>(deserializer: D) -> Result<Self, DeserializeError>
where D: ValueDeserializer<'de> {

37
src/schema/document/mod.rs
View File

@@ -5,24 +5,22 @@
//! - [Value] which provides tantivy with a way to access the document's values in a common way
//! without performing any additional allocations.
//! - [DocumentDeserialize] which implements the necessary code to deserialize the document from the
//! doc store. If you are fine with fetching [TantivyDocument] from the doc store, you can skip
//! implementing this trait for your type.
//! doc store.
//!
//! Tantivy provides a few out-of-box implementations of these core traits to provide
//! some simple usage if you don't want to implement these traits on a custom type yourself.
//!
//! # Out-of-box document implementations
//! - [TantivyDocument] the old document type used by Tantivy before the trait based approach was
//! - [Document] the old document type used by Tantivy before the trait based approach was
//! implemented. This type is still valid and provides all of the original behaviour you might
//! expect.
//! - `BTreeMap<Field, OwnedValue>` a mapping of field_ids to their relevant schema value using a
//! - `BTreeMap<Field, Value>` a mapping of field_ids to their relevant schema value using a
//! BTreeMap.
//! - `HashMap<Field, OwnedValue>` a mapping of field_ids to their relevant schema value using a
//! HashMap.
//! - `HashMap<Field, Value>` a mapping of field_ids to their relevant schema value using a HashMap.
//!
//! # Implementing your custom documents
//! Often in larger projects or higher performance applications you want to avoid the extra overhead
//! of converting your own types to the [TantivyDocument] type, this can often save you a
//! of converting your own types to the Tantivy [Document] type, this can often save you a
//! significant amount of time when indexing by avoiding the additional allocations.
//!
//! ### Important Note
@@ -48,7 +46,6 @@
//!
//! impl Document for MyCustomDocument {
//! // The value type produced by the `iter_fields_and_values` iterator.
//! // tantivy already implements the Value trait for serde_json::Value.
//! type Value<'a> = &'a serde_json::Value;
//! // The iterator which is produced by `iter_fields_and_values`.
//! // Often this is a simple new-type wrapper unless you like super long generics.
@@ -97,11 +94,10 @@
//! implementation for.
//!
//! ## Implementing custom values
//! In order to allow documents to return custom types, they must implement
//! the [Value] trait which provides a way for Tantivy to get a `ReferenceValue` that it can then
//! index and store.
//! Internally, Tantivy only works with `ReferenceValue` which is an enum that tries to borrow
//! as much data as it can
//! as much data as it can, in order to allow documents to return custom types, they must implement
//! the `Value` trait which provides a way for Tantivy to get a `ReferenceValue` that it can then
//! index and store.
//!
//! Values can just as easily be customised as documents by implementing the `Value` trait.
//!
@@ -109,9 +105,9 @@
//! hold references of the data held by the parent [Document] which can then be passed
//! on to the [ReferenceValue].
//!
//! This is why [Value] is implemented for `&'a serde_json::Value` and
//! [&'a tantivy::schema::document::OwnedValue](OwnedValue) but not for their owned counterparts, as
//! we cannot satisfy the lifetime bounds necessary when indexing the documents.
//! This is why `Value` is implemented for `&'a serde_json::Value` and `&'a
//! tantivy::schema::Value` but not for their owned counterparts, as we cannot satisfy the lifetime
//! bounds necessary when indexing the documents.
//!
//! ### A note about returning values
//! The custom value type does not have to be the type stored by the document, instead the
@@ -169,13 +165,10 @@ use std::mem;
pub(crate) use self::de::BinaryDocumentDeserializer;
pub use self::de::{
ArrayAccess, DeserializeError, DocumentDeserialize, DocumentDeserializeSeed,
DocumentDeserializer, ObjectAccess, ValueDeserialize, ValueDeserializer, ValueType,
ValueVisitor,
};
pub use self::default_document::{
CompactDocArrayIter, CompactDocObjectIter, CompactDocValue, DocParsingError, TantivyDocument,
ArrayAccess, DeserializeError, DocumentDeserialize, DocumentDeserializer, ObjectAccess,
ValueDeserialize, ValueDeserializer, ValueType, ValueVisitor,
};
pub use self::default_document::{DocParsingError, TantivyDocument};
pub use self::owned_value::OwnedValue;
pub(crate) use self::se::BinaryDocumentSerializer;
pub use self::value::{ReferenceValue, ReferenceValueLeaf, Value};
@@ -236,7 +229,7 @@ pub trait Document: Send + Sync + 'static {
let field_name = schema.get_field_name(field);
let values: Vec<OwnedValue> = field_values
.into_iter()
.map(|val| OwnedValue::from(val.as_value()))
.map(|val| val.as_value().into())
.collect();
field_map.insert(field_name.to_string(), values);
}

41
src/schema/document/owned_value.rs
View File

@@ -8,7 +8,6 @@ use serde::de::{MapAccess, SeqAccess};
use time::format_description::well_known::Rfc3339;
use time::OffsetDateTime;
use super::existing_type_impls::can_be_rfc3339_date_time;
use super::ReferenceValueLeaf;
use crate::schema::document::{
ArrayAccess, DeserializeError, ObjectAccess, ReferenceValue, Value, ValueDeserialize,
@@ -66,13 +65,13 @@ impl<'a> Value<'a> for &'a OwnedValue {
match self {
OwnedValue::Null => ReferenceValueLeaf::Null.into(),
OwnedValue::Str(val) => ReferenceValueLeaf::Str(val).into(),
OwnedValue::PreTokStr(val) => ReferenceValueLeaf::PreTokStr(val.clone().into()).into(),
OwnedValue::PreTokStr(val) => ReferenceValueLeaf::PreTokStr(val).into(),
OwnedValue::U64(val) => ReferenceValueLeaf::U64(*val).into(),
OwnedValue::I64(val) => ReferenceValueLeaf::I64(*val).into(),
OwnedValue::F64(val) => ReferenceValueLeaf::F64(*val).into(),
OwnedValue::Bool(val) => ReferenceValueLeaf::Bool(*val).into(),
OwnedValue::Date(val) => ReferenceValueLeaf::Date(*val).into(),
OwnedValue::Facet(val) => ReferenceValueLeaf::Facet(val.encoded_str()).into(),
OwnedValue::Facet(val) => ReferenceValueLeaf::Facet(val).into(),
OwnedValue::Bytes(val) => ReferenceValueLeaf::Bytes(val).into(),
OwnedValue::IpAddr(val) => ReferenceValueLeaf::IpAddr(*val).into(),
OwnedValue::Array(array) => ReferenceValue::Array(array.iter()),
@@ -149,7 +148,7 @@ impl ValueDeserialize for OwnedValue {
fn visit_object<'de, A>(&self, mut access: A) -> Result<Self::Value, DeserializeError>
where A: ObjectAccess<'de> {
let mut elements = Vec::with_capacity(access.size_hint());
let mut elements = Vec::new();
while let Some((key, value)) = access.next_entry()? {
elements.push((key, value));
@@ -168,7 +167,6 @@ impl Eq for OwnedValue {}
impl serde::Serialize for OwnedValue {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where S: serde::Serializer {
use serde::ser::SerializeMap;
match *self {
OwnedValue::Null => serializer.serialize_unit(),
OwnedValue::Str(ref v) => serializer.serialize_str(v),
@@ -182,13 +180,7 @@ impl serde::Serialize for OwnedValue {
}
OwnedValue::Facet(ref facet) => facet.serialize(serializer),
OwnedValue::Bytes(ref bytes) => serializer.serialize_str(&BASE64.encode(bytes)),
OwnedValue::Object(ref obj) => {
let mut map = serializer.serialize_map(Some(obj.len()))?;
for (k, v) in obj {
map.serialize_entry(k, v)?;
}
map.end()
}
OwnedValue::Object(ref obj) => obj.serialize(serializer),
OwnedValue::IpAddr(ref ip_v6) => {
// Ensure IpV4 addresses get serialized as IpV4, but excluding IpV6 loopback.
if let Some(ip_v4) = ip_v6.to_ipv4_mapped() {
@@ -256,7 +248,10 @@ impl<'de> serde::Deserialize<'de> for OwnedValue {
fn visit_map<A>(self, mut map: A) -> Result<Self::Value, A::Error>
where A: MapAccess<'de> {
let mut object = map.size_hint().map(Vec::with_capacity).unwrap_or_default();
let mut object =
map.size_hint()
.map(Vec::with_capacity)
.unwrap_or_default();
while let Some((key, value)) = map.next_entry()? {
object.push((key, value));
}
@@ -278,13 +273,11 @@ impl<'a, V: Value<'a>> From<ReferenceValue<'a, V>> for OwnedValue {
ReferenceValueLeaf::I64(val) => OwnedValue::I64(val),
ReferenceValueLeaf::F64(val) => OwnedValue::F64(val),
ReferenceValueLeaf::Date(val) => OwnedValue::Date(val),
ReferenceValueLeaf::Facet(val) => {
OwnedValue::Facet(Facet::from_encoded_string(val.to_string()))
}
ReferenceValueLeaf::Facet(val) => OwnedValue::Facet(val.clone()),
ReferenceValueLeaf::Bytes(val) => OwnedValue::Bytes(val.to_vec()),
ReferenceValueLeaf::IpAddr(val) => OwnedValue::IpAddr(val),
ReferenceValueLeaf::Bool(val) => OwnedValue::Bool(val),
ReferenceValueLeaf::PreTokStr(val) => OwnedValue::PreTokStr(*val.clone()),
ReferenceValueLeaf::PreTokStr(val) => OwnedValue::PreTokStr(val.clone()),
},
ReferenceValue::Array(val) => {
OwnedValue::Array(val.map(|v| v.as_value().into()).collect())
@@ -376,6 +369,16 @@ impl From<BTreeMap<String, OwnedValue>> for OwnedValue {
}
}
fn can_be_rfc3339_date_time(text: &str) -> bool {
if let Some(&first_byte) = text.as_bytes().first() {
if first_byte.is_ascii_digit() {
return true;
}
}
false
}
impl From<serde_json::Value> for OwnedValue {
fn from(value: serde_json::Value) -> Self {
match value {
@@ -416,8 +419,7 @@ impl From<serde_json::Value> for OwnedValue {
impl From<serde_json::Map<String, serde_json::Value>> for OwnedValue {
fn from(map: serde_json::Map<String, serde_json::Value>) -> Self {
let object: Vec<(String, OwnedValue)> = map
.into_iter()
let object: Vec<(String, OwnedValue)> = map.into_iter()
.map(|(key, value)| (key, OwnedValue::from(value)))
.collect();
OwnedValue::Object(object)
@@ -463,7 +465,6 @@ mod tests {
let mut doc = TantivyDocument::default();
doc.add_bytes(bytes_field, "".as_bytes());
let json_string = doc.to_json(&schema);
assert_eq!(json_string, r#"{"my_bytes":[""]}"#);
}

9
src/schema/document/se.rs
View File

@@ -25,7 +25,6 @@ where W: Write
/// Attempts to serialize a given document and write the output
/// to the writer.
#[inline]
pub(crate) fn serialize_doc<D>(&mut self, doc: &D) -> io::Result<()>
where D: Document {
let stored_field_values = || {
@@ -122,7 +121,7 @@ where W: Write
ReferenceValueLeaf::Facet(val) => {
self.write_type_code(type_codes::HIERARCHICAL_FACET_CODE)?;
Cow::Borrowed(val).serialize(self.writer)
val.serialize(self.writer)
}
ReferenceValueLeaf::Bytes(val) => {
self.write_type_code(type_codes::BYTES_CODE)?;
@@ -429,7 +428,7 @@ mod tests {
);
let facet = Facet::from_text("/hello/world").unwrap();
let result = serialize_value(ReferenceValueLeaf::Facet(facet.encoded_str()).into());
let result = serialize_value(ReferenceValueLeaf::Facet(&facet).into());
let expected = binary_repr!(
type_codes::HIERARCHICAL_FACET_CODE => Facet::from_text("/hello/world").unwrap(),
);
@@ -442,8 +441,7 @@ mod tests {
text: "hello, world".to_string(),
tokens: vec![Token::default(), Token::default()],
};
let result =
serialize_value(ReferenceValueLeaf::PreTokStr(pre_tok_str.clone().into()).into());
let result = serialize_value(ReferenceValueLeaf::PreTokStr(&pre_tok_str).into());
let expected = binary_repr!(
type_codes::EXT_CODE, type_codes::TOK_STR_EXT_CODE => pre_tok_str,
);
@@ -680,7 +678,6 @@ mod tests {
);
}
#[inline]
fn serialize_doc<D: Document>(doc: &D, schema: &Schema) -> Vec<u8> {
let mut writer = Vec::new();

111
src/schema/document/value.rs
View File

@@ -3,6 +3,7 @@ use std::net::Ipv6Addr;
use common::DateTime;
use crate::schema::Facet;
use crate::tokenizer::PreTokenizedString;
/// A single field value.
@@ -27,7 +28,7 @@ pub trait Value<'a>: Send + Sync + Debug {
}
#[inline]
/// If the Value is a leaf, returns the associated leaf. Returns None otherwise.
/// If the Value is a String, returns the associated str. Returns None otherwise.
fn as_leaf(&self) -> Option<ReferenceValueLeaf<'a>> {
if let ReferenceValue::Leaf(val) = self.as_value() {
Some(val)
@@ -81,9 +82,8 @@ pub trait Value<'a>: Send + Sync + Debug {
#[inline]
/// If the Value is a pre-tokenized string, returns the associated string. Returns None
/// otherwise.
fn as_pre_tokenized_text(&self) -> Option<Box<PreTokenizedString>> {
self.as_leaf()
.and_then(|leaf| leaf.into_pre_tokenized_text())
fn as_pre_tokenized_text(&self) -> Option<&'a PreTokenizedString> {
self.as_leaf().and_then(|leaf| leaf.as_pre_tokenized_text())
}
#[inline]
@@ -94,7 +94,7 @@ pub trait Value<'a>: Send + Sync + Debug {
#[inline]
/// If the Value is a facet, returns the associated facet. Returns None otherwise.
fn as_facet(&self) -> Option<&'a str> {
fn as_facet(&self) -> Option<&'a Facet> {
self.as_leaf().and_then(|leaf| leaf.as_facet())
}
@@ -132,7 +132,7 @@ pub trait Value<'a>: Send + Sync + Debug {
}
/// A enum representing a leaf value for tantivy to index.
#[derive(Clone, Debug, PartialEq)]
#[derive(Clone, Copy, Debug, PartialEq)]
pub enum ReferenceValueLeaf<'a> {
/// A null value.
Null,
@@ -146,9 +146,8 @@ pub enum ReferenceValueLeaf<'a> {
F64(f64),
/// Date/time with nanoseconds precision
Date(DateTime),
/// Facet string needs to match the format of
/// [Facet::encoded_str](crate::schema::Facet::encoded_str).
Facet(&'a str),
/// Facet
Facet(&'a Facet),
/// Arbitrarily sized byte array
Bytes(&'a [u8]),
/// IpV6 Address. Internally there is no IpV4, it needs to be converted to `Ipv6Addr`.
@@ -156,70 +155,7 @@ pub enum ReferenceValueLeaf<'a> {
/// Bool value
Bool(bool),
/// Pre-tokenized str type,
PreTokStr(Box<PreTokenizedString>),
}
impl From<u64> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: u64) -> Self {
ReferenceValueLeaf::U64(value)
}
}
impl From<i64> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: i64) -> Self {
ReferenceValueLeaf::I64(value)
}
}
impl From<f64> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: f64) -> Self {
ReferenceValueLeaf::F64(value)
}
}
impl From<bool> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: bool) -> Self {
ReferenceValueLeaf::Bool(value)
}
}
impl<'a> From<&'a str> for ReferenceValueLeaf<'a> {
#[inline]
fn from(value: &'a str) -> Self {
ReferenceValueLeaf::Str(value)
}
}
impl<'a> From<&'a [u8]> for ReferenceValueLeaf<'a> {
#[inline]
fn from(value: &'a [u8]) -> Self {
ReferenceValueLeaf::Bytes(value)
}
}
impl From<DateTime> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: DateTime) -> Self {
ReferenceValueLeaf::Date(value)
}
}
impl From<Ipv6Addr> for ReferenceValueLeaf<'_> {
#[inline]
fn from(value: Ipv6Addr) -> Self {
ReferenceValueLeaf::IpAddr(value)
}
}
impl From<PreTokenizedString> for ReferenceValueLeaf<'_> {
#[inline]
fn from(val: PreTokenizedString) -> Self {
ReferenceValueLeaf::PreTokStr(Box::new(val))
}
PreTokStr(&'a PreTokenizedString),
}
impl<'a, T: Value<'a> + ?Sized> From<ReferenceValueLeaf<'a>> for ReferenceValue<'a, T> {
@@ -323,9 +259,9 @@ impl<'a> ReferenceValueLeaf<'a> {
}
#[inline]
/// If the Value is a pre-tokenized string, consumes it and returns the string.
/// Returns None otherwise.
pub fn into_pre_tokenized_text(self) -> Option<Box<PreTokenizedString>> {
/// If the Value is a pre-tokenized string, returns the associated string. Returns None
/// otherwise.
pub fn as_pre_tokenized_text(&self) -> Option<&'a PreTokenizedString> {
if let Self::PreTokStr(val) = self {
Some(val)
} else {
@@ -345,7 +281,7 @@ impl<'a> ReferenceValueLeaf<'a> {
#[inline]
/// If the Value is a facet, returns the associated facet. Returns None otherwise.
pub fn as_facet(&self) -> Option<&'a str> {
pub fn as_facet(&self) -> Option<&'a Facet> {
if let Self::Facet(val) = self {
Some(val)
} else {
@@ -386,16 +322,6 @@ where V: Value<'a>
}
}
#[inline]
/// If the Value is a leaf, consume it and return the leaf. Returns None otherwise.
pub fn into_leaf(self) -> Option<ReferenceValueLeaf<'a>> {
if let Self::Leaf(val) = self {
Some(val)
} else {
None
}
}
#[inline]
/// If the Value is a String, returns the associated str. Returns None otherwise.
pub fn as_str(&self) -> Option<&'a str> {
@@ -439,11 +365,10 @@ where V: Value<'a>
}
#[inline]
/// If the Value is a pre-tokenized string, consumes it and returns the string.
/// Returns None otherwise.
pub fn into_pre_tokenized_text(self) -> Option<Box<PreTokenizedString>> {
self.into_leaf()
.and_then(|leaf| leaf.into_pre_tokenized_text())
/// If the Value is a pre-tokenized string, returns the associated string. Returns None
/// otherwise.
pub fn as_pre_tokenized_text(&self) -> Option<&'a PreTokenizedString> {
self.as_leaf().and_then(|leaf| leaf.as_pre_tokenized_text())
}
#[inline]
@@ -454,7 +379,7 @@ where V: Value<'a>
#[inline]
/// If the Value is a facet, returns the associated facet. Returns None otherwise.
pub fn as_facet(&self) -> Option<&'a str> {
pub fn as_facet(&self) -> Option<&'a Facet> {
self.as_leaf().and_then(|leaf| leaf.as_facet())
}

39
src/schema/field_type.rs
View File

@@ -568,21 +568,21 @@ mod tests {
let schema = schema_builder.build();
let doc = TantivyDocument::parse_json(&schema, r#"{"id": 100}"#).unwrap();
assert_eq!(
OwnedValue::Str("100".to_string()),
doc.get_first(text_field).unwrap().into()
&OwnedValue::Str("100".to_string()),
doc.get_first(text_field).unwrap()
);
let doc = TantivyDocument::parse_json(&schema, r#"{"id": true}"#).unwrap();
assert_eq!(
OwnedValue::Str("true".to_string()),
doc.get_first(text_field).unwrap().into()
&OwnedValue::Str("true".to_string()),
doc.get_first(text_field).unwrap()
);
// Not sure if this null coercion is the best approach
let doc = TantivyDocument::parse_json(&schema, r#"{"id": null}"#).unwrap();
assert_eq!(
OwnedValue::Str("null".to_string()),
doc.get_first(text_field).unwrap().into()
&OwnedValue::Str("null".to_string()),
doc.get_first(text_field).unwrap()
);
}
@@ -595,18 +595,9 @@ mod tests {
let schema = schema_builder.build();
let doc_json = r#"{"i64": "100", "u64": "100", "f64": "100"}"#;
let doc = TantivyDocument::parse_json(&schema, doc_json).unwrap();
assert_eq!(
OwnedValue::I64(100),
doc.get_first(i64_field).unwrap().into()
);
assert_eq!(
OwnedValue::U64(100),
doc.get_first(u64_field).unwrap().into()
);
assert_eq!(
OwnedValue::F64(100.0),
doc.get_first(f64_field).unwrap().into()
);
assert_eq!(&OwnedValue::I64(100), doc.get_first(i64_field).unwrap());
assert_eq!(&OwnedValue::U64(100), doc.get_first(u64_field).unwrap());
assert_eq!(&OwnedValue::F64(100.0), doc.get_first(f64_field).unwrap());
}
#[test]
@@ -616,17 +607,11 @@ mod tests {
let schema = schema_builder.build();
let doc_json = r#"{"bool": "true"}"#;
let doc = TantivyDocument::parse_json(&schema, doc_json).unwrap();
assert_eq!(
OwnedValue::Bool(true),
doc.get_first(bool_field).unwrap().into()
);
assert_eq!(&OwnedValue::Bool(true), doc.get_first(bool_field).unwrap());
let doc_json = r#"{"bool": "false"}"#;
let doc = TantivyDocument::parse_json(&schema, doc_json).unwrap();
assert_eq!(
OwnedValue::Bool(false),
doc.get_first(bool_field).unwrap().into()
);
assert_eq!(&OwnedValue::Bool(false), doc.get_first(bool_field).unwrap());
}
#[test]
@@ -661,7 +646,7 @@ mod tests {
let doc = TantivyDocument::parse_json(&schema, doc_json).unwrap();
let date = doc.get_first(date_field).unwrap();
// Time zone is converted to UTC
assert_eq!("Leaf(Date(2019-10-12T05:20:50.52Z))", format!("{date:?}"));
assert_eq!("Date(2019-10-12T05:20:50.52Z)", format!("{date:?}"));
}
#[test]

46
src/schema/field_value.rs Normal file
View File

@@ -0,0 +1,46 @@
use crate::schema::{Field, OwnedValue};
/// `FieldValue` holds together a `Field` and its `Value`.
#[allow(missing_docs)]
#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub struct FieldValue {
pub field: Field,
pub value: OwnedValue,
}
impl FieldValue {
/// Constructor
pub fn new(field: Field, value: OwnedValue) -> FieldValue {
FieldValue { field, value }
}
/// Field accessor
pub fn field(&self) -> Field {
self.field
}
/// Value accessor
pub fn value(&self) -> &OwnedValue {
&self.value
}
}
impl From<FieldValue> for OwnedValue {
fn from(field_value: FieldValue) -> Self {
field_value.value
}
}
/// A helper wrapper for creating standard iterators
/// out of the fields iterator trait.
pub struct FieldValueIter<'a>(pub(crate) std::slice::Iter<'a, FieldValue>);
impl<'a> Iterator for FieldValueIter<'a> {
type Item = (Field, &'a OwnedValue);
fn next(&mut self) -> Option<Self::Item> {
self.0
.next()
.map(|field_value| (field_value.field, &field_value.value))
}
}

3
src/schema/flags.rs
View File

@@ -1,6 +1,7 @@
use std::ops::BitOr;
use crate::schema::{DateOptions, NumericOptions, TextOptions};
use crate::schema::{NumericOptions, TextOptions};
use crate::DateOptions;
#[derive(Clone)]
pub struct StoredFlag;

8
src/schema/mod.rs
View File

@@ -114,6 +114,7 @@ pub(crate) mod term;
mod field_entry;
mod field_type;
mod field_value;
mod bytes_options;
mod date_time_options;
@@ -129,6 +130,8 @@ mod text_options;
use columnar::ColumnType;
pub use self::bytes_options::BytesOptions;
#[allow(deprecated)]
pub use self::date_time_options::DatePrecision;
pub use self::date_time_options::{DateOptions, DateTimePrecision, DATE_TIME_PRECISION_INDEXED};
pub use self::document::{DocParsingError, Document, OwnedValue, TantivyDocument, Value};
pub(crate) use self::facet::FACET_SEP_BYTE;
@@ -137,14 +140,17 @@ pub use self::facet_options::FacetOptions;
pub use self::field::Field;
pub use self::field_entry::FieldEntry;
pub use self::field_type::{FieldType, Type};
pub use self::field_value::FieldValue;
pub use self::flags::{COERCE, FAST, INDEXED, STORED};
pub use self::index_record_option::IndexRecordOption;
pub use self::ip_options::{IntoIpv6Addr, IpAddrOptions};
pub use self::json_object_options::JsonObjectOptions;
pub use self::named_field_document::NamedFieldDocument;
#[allow(deprecated)]
pub use self::numeric_options::IntOptions;
pub use self::numeric_options::NumericOptions;
pub use self::schema::{Schema, SchemaBuilder};
pub use self::term::{Term, ValueBytes};
pub use self::term::{Term, ValueBytes, JSON_END_OF_PATH};
pub use self::text_options::{TextFieldIndexing, TextOptions, STRING, TEXT};
/// Validator for a potential `field_name`.

4
src/schema/numeric_options.rs
View File

@@ -5,6 +5,10 @@ use serde::{Deserialize, Serialize};
use super::flags::CoerceFlag;
use crate::schema::flags::{FastFlag, IndexedFlag, SchemaFlagList, StoredFlag};
#[deprecated(since = "0.17.0", note = "Use NumericOptions instead.")]
/// Deprecated use [`NumericOptions`] instead.
pub type IntOptions = NumericOptions;
/// Define how an `u64`, `i64`, or `f64` field should be handled by tantivy.
#[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Default)]
#[serde(from = "NumericOptionsDeser")]

12
src/schema/schema.rs
View File

@@ -645,15 +645,15 @@ mod tests {
let doc =
TantivyDocument::convert_named_doc(&schema, NamedFieldDocument(named_doc_map)).unwrap();
assert_eq!(
doc.get_all(title).map(OwnedValue::from).collect::<Vec<_>>(),
doc.get_all(title).collect::<Vec<_>>(),
vec![
OwnedValue::from("title1".to_string()),
OwnedValue::from("title2".to_string())
&OwnedValue::from("title1".to_string()),
&OwnedValue::from("title2".to_string())
]
);
assert_eq!(
doc.get_all(val).map(OwnedValue::from).collect::<Vec<_>>(),
vec![OwnedValue::from(14u64), OwnedValue::from(-1i64)]
doc.get_all(val).collect::<Vec<_>>(),
vec![&OwnedValue::from(14u64), &OwnedValue::from(-1i64)]
);
}
@@ -682,7 +682,7 @@ mod tests {
let schema = schema_builder.build();
{
let doc = TantivyDocument::parse_json(&schema, "{}").unwrap();
assert!(doc.field_values().next().is_none());
assert!(doc.field_values().is_empty());
}
{
let doc = TantivyDocument::parse_json(

75
src/schema/term.rs
View File

@@ -3,16 +3,22 @@ use std::net::Ipv6Addr;
use std::{fmt, str};
use columnar::{MonotonicallyMappableToU128, MonotonicallyMappableToU64};
use common::json_path_writer::{JSON_END_OF_PATH, JSON_PATH_SEGMENT_SEP_STR};
use common::JsonPathWriter;
use super::date_time_options::DATE_TIME_PRECISION_INDEXED;
use super::Field;
use crate::fastfield::FastValue;
use crate::json_utils::split_json_path;
use crate::schema::{Facet, Type};
use crate::DateTime;
/// Separates the different segments of a json path.
pub const JSON_PATH_SEGMENT_SEP: u8 = 1u8;
pub const JSON_PATH_SEGMENT_SEP_STR: &str =
unsafe { std::str::from_utf8_unchecked(&[JSON_PATH_SEGMENT_SEP]) };
/// Separates the json path and the value in
/// a JSON term binary representation.
pub const JSON_END_OF_PATH: u8 = 0u8;
/// Term represents the value that the token can take.
/// It's a serialized representation over different types.
///
@@ -35,28 +41,6 @@ impl Term {
Term(data)
}
/// Creates a term from a json path.
///
/// The json path can address a nested value in a JSON object.
/// e.g. `{"k8s": {"node": {"id": 5}}}` can be addressed via `k8s.node.id`.
///
/// In case there are dots in the field name, and the `expand_dots_enabled` parameter is not
/// set they need to be escaped with a backslash.
/// e.g. `{"k8s.node": {"id": 5}}` can be addressed via `k8s\.node.id`.
pub fn from_field_json_path(field: Field, json_path: &str, expand_dots_enabled: bool) -> Term {
let paths = split_json_path(json_path);
let mut json_path = JsonPathWriter::with_expand_dots(expand_dots_enabled);
for path in paths {
json_path.push(&path);
}
json_path.set_end();
let mut term = Term::with_type_and_field(Type::Json, field);
term.append_bytes(json_path.as_str().as_bytes());
term
}
pub(crate) fn with_type_and_field(typ: Type, field: Field) -> Term {
let mut term = Self::with_capacity(8);
term.set_field_and_type(field, typ);
@@ -185,11 +169,7 @@ impl Term {
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) {
pub(crate) fn append_type_and_fast_value<T: FastValue>(&mut self, val: T) {
self.0.push(T::to_type().to_code());
let value = if T::to_type() == Type::Date {
DateTime::from_u64(val.to_u64())
@@ -201,15 +181,6 @@ impl Term {
self.0.extend(value.to_be_bytes().as_ref());
}
/// Append a string type marker + string to a term.
/// This is used in JSON type to append a str after the path.
///
/// It will not clear existing bytes.
pub fn append_type_and_str(&mut self, val: &str) {
self.0.push(Type::Str.to_code());
self.0.extend(val.as_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());
@@ -221,6 +192,11 @@ impl Term {
self.0.extend(bytes);
}
/// Set the texts only, keeping the field untouched.
pub fn set_text(&mut self, text: &str) {
self.set_bytes(text.as_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);
@@ -257,6 +233,27 @@ impl Term {
}
&mut self.0[len_before..]
}
/// Appends a JSON_PATH_SEGMENT_SEP to the term.
/// Only used for JSON type.
#[inline]
pub fn add_json_path_separator(&mut self) {
self.0.push(JSON_PATH_SEGMENT_SEP);
}
/// Sets the current end to JSON_END_OF_PATH.
/// Only used for JSON type.
#[inline]
pub fn set_json_path_end(&mut self) {
let buffer_len = self.0.len();
self.0[buffer_len - 1] = JSON_END_OF_PATH;
}
/// Sets the current end to JSON_PATH_SEGMENT_SEP.
/// Only used for JSON type.
#[inline]
pub fn set_json_path_separator(&mut self) {
let buffer_len = self.0.len();
self.0[buffer_len - 1] = JSON_PATH_SEGMENT_SEP;
}
}
impl<B> Term<B>

4
src/space_usage/mod.rs
View File

@@ -12,8 +12,8 @@ use std::collections::HashMap;
use common::ByteCount;
use serde::{Deserialize, Serialize};
use crate::index::SegmentComponent;
use crate::schema::Field;
use crate::SegmentComponent;
/// Enum containing any of the possible space usage results for segment components.
pub enum ComponentSpaceUsage {
@@ -115,7 +115,7 @@ impl SegmentSpaceUsage {
/// Use the components directly if this is somehow in performance critical code.
pub fn component(&self, component: SegmentComponent) -> ComponentSpaceUsage {
use self::ComponentSpaceUsage::*;
use crate::index::SegmentComponent::*;
use crate::SegmentComponent::*;
match component {
Postings => PerField(self.postings().clone()),
Positions => PerField(self.positions().clone()),

14
src/store/mod.rs
View File

@@ -59,6 +59,7 @@ pub mod tests {
use super::*;
use crate::directory::{Directory, RamDirectory, WritePtr};
use crate::fastfield::AliveBitSet;
use crate::schema::document::Value;
use crate::schema::{
self, Schema, TantivyDocument, TextFieldIndexing, TextOptions, STORED, TEXT,
};
@@ -91,8 +92,8 @@ pub mod tests {
StoreWriter::new(writer, compressor, blocksize, separate_thread).unwrap();
for i in 0..num_docs {
let mut doc = TantivyDocument::default();
doc.add_text(field_body, LOREM);
doc.add_text(field_title, format!("Doc {i}"));
doc.add_field_value(field_body, LOREM.to_string());
doc.add_field_value(field_title, format!("Doc {i}"));
store_writer.store(&doc, &schema).unwrap();
}
store_writer.close().unwrap();
@@ -118,7 +119,7 @@ pub mod tests {
let store = StoreReader::open(store_file, 10)?;
for i in 0..NUM_DOCS as u32 {
assert_eq!(
store
*store
.get::<TantivyDocument>(i)?
.get_first(field_title)
.unwrap()
@@ -130,12 +131,7 @@ pub mod tests {
for doc in store.iter::<TantivyDocument>(Some(&alive_bitset)) {
let doc = doc?;
let title_content = doc
.get_first(field_title)
.unwrap()
.as_str()
.unwrap()
.to_string();
let title_content = doc.get_first(field_title).unwrap().as_str().unwrap();
if !title_content.starts_with("Doc ") {
panic!("unexpected title_content {title_content}");
}

78
src/store/reader.rs
View File

@@ -1,6 +1,5 @@
use std::io;
use std::iter::Sum;
use std::marker::PhantomData;
use std::num::NonZeroUsize;
use std::ops::{AddAssign, Range};
use std::sync::atomic::{AtomicUsize, Ordering};
@@ -15,14 +14,10 @@ use super::Decompressor;
use crate::directory::FileSlice;
use crate::error::DataCorruption;
use crate::fastfield::AliveBitSet;
use crate::schema::document::{
BinaryDocumentDeserializer, DocumentDeserialize, DocumentDeserializeSeed,
};
use crate::schema::document::{BinaryDocumentDeserializer, Document, DocumentDeserialize};
use crate::space_usage::StoreSpaceUsage;
use crate::store::index::Checkpoint;
use crate::DocId;
#[cfg(feature = "quickwit")]
use crate::Executor;
pub(crate) const DOCSTORE_CACHE_CAPACITY: usize = 100;
@@ -204,21 +199,11 @@ impl StoreReader {
/// It should not be called to score documents
/// for instance.
pub fn get<D: DocumentDeserialize>(&self, doc_id: DocId) -> crate::Result<D> {
self.get_seed(doc_id, PhantomData)
}
/// A stateful version of [`get`][Self::get].
pub fn get_seed<T: DocumentDeserializeSeed>(
&self,
doc_id: DocId,
seed: T,
) -> crate::Result<T::Value> {
let mut doc_bytes = self.get_document_bytes(doc_id)?;
let deserializer = BinaryDocumentDeserializer::from_reader(&mut doc_bytes)
.map_err(crate::TantivyError::from)?;
seed.deserialize(deserializer)
.map_err(crate::TantivyError::from)
D::deserialize(deserializer).map_err(crate::TantivyError::from)
}
/// Returns raw bytes of a given document.
@@ -250,27 +235,16 @@ impl StoreReader {
/// Iterator over all Documents in their order as they are stored in the doc store.
/// Use this, if you want to extract all Documents from the doc store.
/// The `alive_bitset` has to be forwarded from the `SegmentReader` or the results may be wrong.
pub fn iter<'a: 'b, 'b, D: DocumentDeserialize + 'b>(
pub fn iter<'a: 'b, 'b, D: Document + DocumentDeserialize>(
&'b self,
alive_bitset: Option<&'a AliveBitSet>,
) -> impl Iterator<Item = crate::Result<D>> + 'b {
self.iter_seed(alive_bitset, &PhantomData)
}
/// A stateful variant of [`iter`][Self::iter].
pub fn iter_seed<'a: 'b, 'b, T: DocumentDeserializeSeed + Clone + 'b>(
&'b self,
alive_bitset: Option<&'a AliveBitSet>,
seed: &'b T,
) -> impl Iterator<Item = crate::Result<T::Value>> + 'b {
self.iter_raw(alive_bitset).map(|doc_bytes_res| {
let mut doc_bytes = doc_bytes_res?;
let deserializer = BinaryDocumentDeserializer::from_reader(&mut doc_bytes)
.map_err(crate::TantivyError::from)?;
seed.clone()
.deserialize(deserializer)
.map_err(crate::TantivyError::from)
D::deserialize(deserializer).map_err(crate::TantivyError::from)
})
}
@@ -367,11 +341,7 @@ impl StoreReader {
/// In most cases use [`get_async`](Self::get_async)
///
/// Loads and decompresses a block asynchronously.
async fn read_block_async(
&self,
checkpoint: &Checkpoint,
executor: &Executor,
) -> io::Result<Block> {
async fn read_block_async(&self, checkpoint: &Checkpoint) -> io::Result<Block> {
let cache_key = checkpoint.byte_range.start;
if let Some(block) = self.cache.get_from_cache(checkpoint.byte_range.start) {
return Ok(block);
@@ -383,12 +353,8 @@ impl StoreReader {
.read_bytes_async()
.await?;
let decompressor = self.decompressor;
let maybe_decompressed_block = executor
.spawn_blocking(move || decompressor.decompress(compressed_block.as_ref()))
.await
.expect("decompression panicked");
let decompressed_block = OwnedBytes::new(maybe_decompressed_block?);
let decompressed_block =
OwnedBytes::new(self.decompressor.decompress(compressed_block.as_ref())?);
self.cache
.put_into_cache(cache_key, decompressed_block.clone());
@@ -397,38 +363,19 @@ impl StoreReader {
}
/// Reads raw bytes of a given document asynchronously.
pub async fn get_document_bytes_async(
&self,
doc_id: DocId,
executor: &Executor,
) -> crate::Result<OwnedBytes> {
pub async fn get_document_bytes_async(&self, doc_id: DocId) -> crate::Result<OwnedBytes> {
let checkpoint = self.block_checkpoint(doc_id)?;
let block = self.read_block_async(&checkpoint, executor).await?;
let block = self.read_block_async(&checkpoint).await?;
Self::get_document_bytes_from_block(block, doc_id, &checkpoint)
}
/// Fetches a document asynchronously. Async version of [`get`](Self::get).
pub async fn get_async<D: DocumentDeserialize>(
&self,
doc_id: DocId,
executor: &Executor,
) -> crate::Result<D> {
self.get_async_seed(doc_id, executor, PhantomData).await
}
/// A stateful variant of [`get_async`][Self::get_async].
pub async fn get_async_seed<T: DocumentDeserializeSeed>(
&self,
doc_id: DocId,
executor: &Executor,
seed: T,
) -> crate::Result<T::Value> {
let mut doc_bytes = self.get_document_bytes_async(doc_id, executor).await?;
pub async fn get_async<D: DocumentDeserialize>(&self, doc_id: DocId) -> crate::Result<D> {
let mut doc_bytes = self.get_document_bytes_async(doc_id).await?;
let deserializer = BinaryDocumentDeserializer::from_reader(&mut doc_bytes)
.map_err(crate::TantivyError::from)?;
seed.deserialize(deserializer)
.map_err(crate::TantivyError::from)
D::deserialize(deserializer).map_err(crate::TantivyError::from)
}
}
@@ -438,6 +385,7 @@ mod tests {
use super::*;
use crate::directory::RamDirectory;
use crate::schema::document::Value;
use crate::schema::{Field, TantivyDocument};
use crate::store::tests::write_lorem_ipsum_store;
use crate::store::Compressor;

6
sstable/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-sstable"
version = "0.3.0"
version = "0.2.0"
edition = "2021"
license = "MIT"
homepage = "https://github.com/quickwit-oss/tantivy"
@@ -10,8 +10,8 @@ categories = ["database-implementations", "data-structures", "compression"]
description = "sstables for tantivy"
[dependencies]
common = {version= "0.7", path="../common", package="tantivy-common"}
tantivy-bitpacker = { version= "0.6", path="../bitpacker" }
common = {version= "0.6", path="../common", package="tantivy-common"}
tantivy-bitpacker = { version= "0.5", path="../bitpacker" }
tantivy-fst = "0.5"
# experimental gives us access to Decompressor::upper_bound
zstd = { version = "0.13", features = ["experimental"] }

1
sstable/src/lib.rs
View File

@@ -1,5 +1,6 @@
use std::io::{self, Write};
use std::ops::Range;
use std::usize;
use merge::ValueMerger;

4
stacker/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-stacker"
version = "0.3.0"
version = "0.2.0"
edition = "2021"
license = "MIT"
homepage = "https://github.com/quickwit-oss/tantivy"
@@ -9,7 +9,7 @@ description = "term hashmap used for indexing"
[dependencies]
murmurhash32 = "0.3"
common = { version = "0.7", path = "../common/", package = "tantivy-common" }
common = { version = "0.6", path = "../common/", package = "tantivy-common" }
ahash = { version = "0.8.11", default-features = false, optional = true }
rand_distr = "0.4.3"

2
tokenizer-api/Cargo.toml
View File

@@ -1,6 +1,6 @@
[package]
name = "tantivy-tokenizer-api"
version = "0.3.0"
version = "0.2.0"
license = "MIT"
edition = "2021"
description = "Tokenizer API of tantivy"