rust/tantivy
1
0
Fork 0
mirror of https://github.com/quickwit-oss/tantivy.git synced 2026-03-25 22:50:42 +00:00
Code Issues Packages Projects Releases Wiki Activity

Compare commits

base: rust:main
Branches Tags
rust:main rust:optim rust:storage_abstraction rust:composite-agg-rebase rust:composite-agg rust:postings-writer-enum-codec5 rust:congxie/supportAgg rust:postings-writer-enum-codec4 rust:segmentreader_trait rust:stuhood.intersection-termination-upstream rust:postings-writer-enum-codec3 rust:postings-writer-enum-codec2 rust:paradedb/fix-agg-validation rust:postings-writer-enum rust:paul.masurel/skip-first-doc-on-intersection rust:stuhood.lazy-scorers-blocks-upstream rust:paul.masurel/seek_into_the_danger_zone_result rust:paul.masurel/search_into_danger_zone_fastfield rust:stuhood.lazy-scorers-blocks 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:composite-agg
Branches Tags
rust:optim rust:storage_abstraction rust:main rust:composite-agg-rebase rust:composite-agg rust:postings-writer-enum-codec5 rust:congxie/supportAgg rust:postings-writer-enum-codec4 rust:segmentreader_trait rust:stuhood.intersection-termination-upstream rust:postings-writer-enum-codec3 rust:postings-writer-enum-codec2 rust:paradedb/fix-agg-validation rust:postings-writer-enum rust:paul.masurel/skip-first-doc-on-intersection rust:stuhood.lazy-scorers-blocks-upstream rust:paul.masurel/seek_into_the_danger_zone_result rust:paul.masurel/search_into_danger_zone_fastfield rust:stuhood.lazy-scorers-blocks 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
main ... composite-

Author SHA1 Message Date
cong.xie
e9c6383b8c fix: adapt composite aggregation for Quickwit compatibility 
Fixes on top of PR #2714 ("Add composite aggregation") to make it work
with Quickwit's current codebase and Postcard serialization:

- Rewrite SegmentCompositeCollector to match current
  SegmentAggregationCollector trait signatures (collect,
  add_intermediate_aggregation_result, prepare_max_bucket)
- Remove Clone derive from CompositeBucketCollector (incompatible
  with dyn SegmentAggregationCollector)
- Add custom serde for FxHashMap entries in
  IntermediateCompositeBucketResult (Postcard requires known sequence
  length)
- Rewrite AfterKey Serialize/Deserialize to output raw values instead
  of internal "type:value" format, matching Elasticsearch wire format
- Remove unused imports and tracing::warn calls

Made-with: Cursor
 2026-03-13 15:38:25 -04:00
Remi Dettai
d662415b81 Add composite aggregation 2026-03-13 10:32:41 -04:00
45 changed files with 3089 additions and 971 deletions
Expand all files Collapse all files

87
.claude/skills/update-changelog/SKILL.md
View File

@@ -1,87 +0,0 @@
---
name: update-changelog
description: Update CHANGELOG.md with merged PRs since the last changelog update, categorized by type
---
# Update Changelog
This skill updates CHANGELOG.md with merged PRs that aren't already listed.
## Step 1: Determine the changelog scope
Read `CHANGELOG.md` to identify the current unreleased version section at the top (e.g., `Tantivy 0.26 (Unreleased)`).
Collect all PR numbers already mentioned in the unreleased section by extracting `#NNNN` references.
## Step 2: Find merged PRs not yet in the changelog
Use `gh` to list recently merged PRs from the upstream repo:
```bash
gh pr list --repo quickwit-oss/tantivy --state merged --limit 100 --json number,title,author,labels,mergedAt
```
Filter out any PRs whose number already appears in the unreleased section of the changelog.
## Step 3: Consolidate related PRs
Before categorizing, group PRs that belong to the same logical change. This is critical for producing a clean changelog. Use PR descriptions, titles, cross-references, and the files touched to identify relationships.
**Merge follow-up PRs into the original:**
- If a PR is a bugfix, refinement, or follow-up to another PR in the same unreleased cycle, combine them into a single changelog entry with multiple `[#N](url)` links.
- Also consolidate PRs that touch the same feature area even if not explicitly linked — e.g., a PR fixing an edge case in a new API should be folded into the entry for the PR that introduced that API.
**Filter out bugfixes on unreleased features:**
- If a bugfix PR fixes something introduced by another PR in the **same unreleased version**, it must NOT appear as a separate Bugfixes entry. Instead, silently fold it into the original feature/improvement entry. The changelog should describe the final shipped state, not the development history.
- To detect this: check if the bugfix PR references or reverts changes from another PR in the same release cycle, or if it touches code that was newly added (not present in the previous release).
## Step 4: Review the actual code diff
**Do not rely on PR titles or descriptions alone.** For every candidate PR, run `gh pr diff <number> --repo quickwit-oss/tantivy` and read the actual changes. PR titles are often misleading — the diff is the source of truth.
**What to look for in the diff:**
- Does it change observable behavior, public API surface, or performance characteristics?
- Is the change something a user of the library would notice or need to know about?
- Could the change break existing code (API changes, removed features)?
**Skip PRs where the diff reveals the change is not meaningful enough for the changelog** — e.g., cosmetic renames, trivial visibility tweaks, test-only changes, etc.
## Step 5: Categorize each PR group
For each PR (or consolidated group) that survived the diff review, determine its category:
- **Bugfixes** — fixes to behavior that existed in the **previous release**. NOT fixes to features introduced in this release cycle.
- **Features/Improvements** — new features, API additions, new options, improvements that change user-facing behavior or add new capabilities.
- **Performance** — optimizations, speed improvements, memory reductions. **If a PR adds new API whose primary purpose is enabling a performance optimization, categorize it as Performance, not Features.** The deciding question is: does a user benefit from this because of new functionality, or because things got faster/leaner? For example, a new trait method that exists solely to enable cheaper intersection ordering is Performance, not a Feature.
If a PR doesn't clearly fit any category (e.g., CI-only changes, internal refactors with no user-facing impact, dependency bumps with no behavior change), skip it — not everything belongs in the changelog.
When unclear, use your best judgment or ask the user.
## Step 6: Format entries
Each entry must follow this exact format:
```
- Description [#NUMBER](https://github.com/quickwit-oss/tantivy/pull/NUMBER)(@author)
```
Rules:
- The description should be concise and describe the user-facing change (not the implementation). Describe the final shipped state, not the incremental development steps.
- Use sub-categories with bold headers when multiple entries relate to the same area (e.g., `- **Aggregation**` with indented entries beneath). Follow the existing grouping style in the changelog.
- Author is the GitHub username from the PR, prefixed with `@`. For consolidated entries, include all contributing authors.
- For consolidated PRs, list all PR links in a single entry: `[#100](url) [#110](url)` (see existing entries for examples).
## Step 7: Present changes to the user
Show the user the proposed changelog entries grouped by category **before** editing the file. Ask for confirmation or adjustments.
## Step 8: Update CHANGELOG.md
Insert the new entries into the appropriate sections of the unreleased version block. If a section doesn't exist yet, create it following the order: Bugfixes, Features/Improvements, Performance.
Append new entries at the end of each section (before the next section header or version header).
## Step 9: Verify
Read back the updated unreleased section and display it to the user for final review.

48
CHANGELOG.md
View File

@@ -1,51 +1,3 @@
Tantivy 0.26 (Unreleased)
================================
## Bugfixes
- Align float query coercion during search with the columnar coercion rules [#2692](https://github.com/quickwit-oss/tantivy/pull/2692)(@fulmicoton)
- Fix lenient elastic range queries with trailing closing parentheses [#2816](https://github.com/quickwit-oss/tantivy/pull/2816)(@evance-br)
- Fix intersection `seek()` advancing below current doc id [#2812](https://github.com/quickwit-oss/tantivy/pull/2812)(@fulmicoton)
- Fix phrase query prefixed with `*` [#2751](https://github.com/quickwit-oss/tantivy/pull/2751)(@Darkheir)
- Fix `vint` buffer overflow during index creation [#2778](https://github.com/quickwit-oss/tantivy/pull/2778)(@rebasedming)
- Fix integer overflow in `ExpUnrolledLinkedList` for large datasets [#2735](https://github.com/quickwit-oss/tantivy/pull/2735)(@mdashti)
- Fix integer overflow in segment sorting and merge policy truncation [#2846](https://github.com/quickwit-oss/tantivy/pull/2846)(@anaslimem)
- Fix merging of intermediate aggregation results [#2719](https://github.com/quickwit-oss/tantivy/pull/2719)(@PSeitz)
- Fix deduplicate doc counts in term aggregation for multi-valued fields [#2854](https://github.com/quickwit-oss/tantivy/pull/2854)(@nuri-yoo)
## Features/Improvements
- **Aggregation**
- Add filter aggregation [#2711](https://github.com/quickwit-oss/tantivy/pull/2711)(@mdashti)
- Add include/exclude filtering for term aggregations [#2717](https://github.com/quickwit-oss/tantivy/pull/2717)(@PSeitz)
- Add public accessors for intermediate aggregation results [#2829](https://github.com/quickwit-oss/tantivy/pull/2829)(@congx4)
- Replace HyperLogLog++ with Apache DataSketches HLL for cardinality aggregation [#2837](https://github.com/quickwit-oss/tantivy/pull/2837) [#2842](https://github.com/quickwit-oss/tantivy/pull/2842)(@congx4)
- Add composite aggregation [#2856](https://github.com/quickwit-oss/tantivy/pull/2856)(@fulmicoton)
- **Fast Fields**
- Add fast field fallback for `TermQuery` when the field is not indexed [#2693](https://github.com/quickwit-oss/tantivy/pull/2693)(@PSeitz-dd)
- Add fast field support for `Bytes` values [#2830](https://github.com/quickwit-oss/tantivy/pull/2830)(@mdashti)
- **Query Parser**
- Add support for regexes in the query grammar [#2677](https://github.com/quickwit-oss/tantivy/pull/2677) [#2818](https://github.com/quickwit-oss/tantivy/pull/2818)(@Darkheir)
- Deduplicate queries in query parser [#2698](https://github.com/quickwit-oss/tantivy/pull/2698)(@PSeitz-dd)
- Add erased `SortKeyComputer` for sorting on column types unknown until runtime [#2770](https://github.com/quickwit-oss/tantivy/pull/2770) [#2790](https://github.com/quickwit-oss/tantivy/pull/2790)(@stuhood @PSeitz)
- Add natural-order-with-none-highest support in `TopDocs::order_by` [#2780](https://github.com/quickwit-oss/tantivy/pull/2780)(@stuhood)
- Move stemming behing `stemmer` feature flag [#2791](https://github.com/quickwit-oss/tantivy/pull/2791)(@fulmicoton)
- Make `DeleteMeta`, `AddOperation`, `advance_deletes`, `with_max_doc`, `serializer` module, and `delete_queue` public [#2762](https://github.com/quickwit-oss/tantivy/pull/2762) [#2765](https://github.com/quickwit-oss/tantivy/pull/2765) [#2766](https://github.com/quickwit-oss/tantivy/pull/2766) [#2835](https://github.com/quickwit-oss/tantivy/pull/2835)(@philippemnoel @PSeitz)
- Make `Language` hashable [#2763](https://github.com/quickwit-oss/tantivy/pull/2763)(@philippemnoel)
- Improve `space_usage` reporting for JSON fields and columnar data [#2761](https://github.com/quickwit-oss/tantivy/pull/2761)(@PSeitz-dd)
- Split `Term` into `Term` and `IndexingTerm` [#2744](https://github.com/quickwit-oss/tantivy/pull/2744) [#2750](https://github.com/quickwit-oss/tantivy/pull/2750)(@PSeitz-dd @PSeitz)
## Performance
- **Aggregation**
- Large speed up and memory reduction for nested high cardinality aggregations by using one collector per request instead of one per bucket, and adding `PagedTermMap` for faster medium cardinality term aggregations [#2715](https://github.com/quickwit-oss/tantivy/pull/2715) [#2759](https://github.com/quickwit-oss/tantivy/pull/2759)(@PSeitz @PSeitz-dd)
- Optimize low-cardinality term aggregations by using a `Vec` instead of a `HashMap` [#2740](https://github.com/quickwit-oss/tantivy/pull/2740)(@fulmicoton-dd)
- Optimize `ExistsQuery` for a high number of dynamic columns [#2694](https://github.com/quickwit-oss/tantivy/pull/2694)(@PSeitz-dd)
- Add lazy scorers to stop score evaluation early when a doc won't reach the top-K threshold [#2726](https://github.com/quickwit-oss/tantivy/pull/2726) [#2777](https://github.com/quickwit-oss/tantivy/pull/2777)(@fulmicoton @stuhood)
- Add `DocSet::cost()` and use it to order scorers in intersections [#2707](https://github.com/quickwit-oss/tantivy/pull/2707)(@PSeitz)
- Add `collect_block` support for collector wrappers [#2727](https://github.com/quickwit-oss/tantivy/pull/2727)(@stuhood)
- Optimize saturated posting lists by replacing them with `AllScorer` in boolean queries [#2745](https://github.com/quickwit-oss/tantivy/pull/2745) [#2760](https://github.com/quickwit-oss/tantivy/pull/2760) [#2774](https://github.com/quickwit-oss/tantivy/pull/2774)(@fulmicoton @mdashti @trinity-1686a)
- Add `seek_danger` on `DocSet` for more efficient intersections [#2538](https://github.com/quickwit-oss/tantivy/pull/2538) [#2810](https://github.com/quickwit-oss/tantivy/pull/2810)(@PSeitz @stuhood @fulmicoton)
- Skip column traversal in `RangeDocSet` when query range does not overlap with column bounds [#2783](https://github.com/quickwit-oss/tantivy/pull/2783)(@ChangRui-Ryan)
- Speed up exclude queries by supporting multiple excluded `DocSet`s without intermediate union [#2825](https://github.com/quickwit-oss/tantivy/pull/2825)(@PSeitz)
Tantivy 0.25
================================

12
Cargo.toml
View File

@@ -11,7 +11,7 @@ repository = "https://github.com/quickwit-oss/tantivy"
readme = "README.md"
keywords = ["search", "information", "retrieval"]
edition = "2021"
rust-version = "1.86"
rust-version = "1.85"
exclude = ["benches/*.json", "benches/*.txt"]
[dependencies]
@@ -27,7 +27,7 @@ regex = { version = "1.5.5", default-features = false, features = [
aho-corasick = "1.0"
tantivy-fst = "0.5"
memmap2 = { version = "0.9.0", optional = true }
lz4_flex = { version = "0.13", default-features = false, optional = true }
lz4_flex = { version = "0.12", default-features = false, optional = true }
zstd = { version = "0.13", optional = true, default-features = false }
tempfile = { version = "3.12.0", optional = true }
log = "0.4.16"
@@ -47,7 +47,7 @@ rustc-hash = "2.0.0"
thiserror = "2.0.1"
htmlescape = "0.3.1"
fail = { version = "0.5.0", optional = true }
time = { version = "0.3.47", features = ["serde-well-known"] }
time = { version = "0.3.35", features = ["serde-well-known"] }
smallvec = "1.8.0"
rayon = "1.5.2"
lru = "0.16.3"
@@ -64,7 +64,7 @@ query-grammar = { version = "0.25.0", path = "./query-grammar", package = "tanti
tantivy-bitpacker = { version = "0.9", path = "./bitpacker" }
common = { version = "0.10", path = "./common/", package = "tantivy-common" }
tokenizer-api = { version = "0.6", path = "./tokenizer-api", package = "tantivy-tokenizer-api" }
sketches-ddsketch = { version = "0.4", features = ["use_serde"] }
sketches-ddsketch = { path = "./sketches-ddsketch", features = ["use_serde"] }
datasketches = "0.2.0"
futures-util = { version = "0.3.28", optional = true }
futures-channel = { version = "0.3.28", optional = true }
@@ -86,7 +86,7 @@ futures = "0.3.21"
paste = "1.0.11"
more-asserts = "0.3.1"
rand_distr = "0.5"
time = { version = "0.3.47", features = ["serde-well-known", "macros"] }
time = { version = "0.3.10", features = ["serde-well-known", "macros"] }
postcard = { version = "1.0.4", features = [
"use-std",
], default-features = false }
@@ -144,6 +144,7 @@ members = [
"sstable",
"tokenizer-api",
"columnar",
"sketches-ddsketch",
]
# Following the "fail" crate best practises, we isolate
@@ -201,3 +202,4 @@ harness = false
[[bench]]
name = "regex_all_terms"
harness = false

10
benches/agg_bench.rs
View File

@@ -1,5 +1,6 @@
use binggan::plugins::PeakMemAllocPlugin;
use binggan::{black_box, InputGroup, PeakMemAlloc, INSTRUMENTED_SYSTEM};
use common::DateTime;
use rand::distr::weighted::WeightedIndex;
use rand::rngs::StdRng;
use rand::seq::IndexedRandom;
@@ -10,7 +11,7 @@ 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, DateTime, Index, Term};
use tantivy::{doc, Index, Term};
#[global_allocator]
pub static GLOBAL: &PeakMemAlloc<std::alloc::System> = &INSTRUMENTED_SYSTEM;
@@ -319,7 +320,6 @@ fn terms_many_json_mixed_type_with_avg_sub_agg(index: &Index) {
});
execute_agg(index, agg_req);
}
fn composite_term_few(index: &Index) {
let agg_req = json!({
"my_ctf": {
@@ -354,6 +354,7 @@ fn composite_term_many_page_1000_with_avg_sub_agg(index: &Index) {
{ "text_many_terms": { "terms": { "field": "text_many_terms" } } }
],
"size": 1000,
},
"aggs": {
"average_f64": { "avg": { "field": "score_f64" } }
@@ -571,7 +572,6 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
let text_field_all_unique_terms =
schema_builder.add_text_field("text_all_unique_terms", STRING | FAST);
let text_field_many_terms = schema_builder.add_text_field("text_many_terms", STRING | FAST);
let text_field_few_terms = schema_builder.add_text_field("text_few_terms", STRING | FAST);
let text_field_few_terms_status =
schema_builder.add_text_field("text_few_terms_status", STRING | FAST);
let text_field_1000_terms_zipf =
@@ -600,7 +600,6 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
let log_level_distribution =
WeightedIndex::new(status_field_data.iter().map(|item| item.1)).unwrap();
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)
@@ -636,8 +635,6 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
text_field_all_unique_terms => "coolo",
text_field_many_terms => "cool",
text_field_many_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms => "cool",
text_field_few_terms_status => log_level_sample_a,
text_field_few_terms_status => log_level_sample_b,
text_field_1000_terms_zipf => term_1000_a.as_str(),
@@ -668,7 +665,6 @@ fn get_test_index_bench(cardinality: Cardinality) -> tantivy::Result<Index> {
json_field => json,
text_field_all_unique_terms => format!("unique_term_{}", rng.random::<u64>()),
text_field_many_terms => many_terms_data.choose(&mut rng).unwrap().to_string(),
text_field_few_terms => few_terms_data.choose(&mut rng).unwrap().to_string(),
text_field_few_terms_status => status_field_data[log_level_distribution.sample(&mut rng)].0,
text_field_1000_terms_zipf => terms_1000[zipf_1000.sample(&mut rng) as usize - 1].as_str(),
score_field => val as u64,

124
columnar/src/block_accessor.rs
View File

@@ -58,78 +58,6 @@ impl<T: PartialOrd + Copy + std::fmt::Debug + Send + Sync + 'static + Default>
}
}
/// Like `fetch_block_with_missing`, but deduplicates (doc_id, value) pairs
/// so that each unique value per document is returned only once.
///
/// This is necessary for correct document counting in aggregations,
/// where multi-valued fields can produce duplicate entries that inflate counts.
#[inline]
pub fn fetch_block_with_missing_unique_per_doc(
&mut self,
docs: &[u32],
accessor: &Column<T>,
missing: Option<T>,
) where
T: Ord,
{
self.fetch_block_with_missing(docs, accessor, missing);
if accessor.index.get_cardinality().is_multivalue() {
self.dedup_docid_val_pairs();
}
}
/// Removes duplicate (doc_id, value) pairs from the caches.
///
/// After `fetch_block`, entries are sorted by doc_id, but values within
/// the same doc may not be sorted (e.g. `(0,1), (0,2), (0,1)`).
/// We group consecutive entries by doc_id, sort values within each group
/// if it has more than 2 elements, then deduplicate adjacent pairs.
///
/// Skips entirely if no doc_id appears more than once in the block.
fn dedup_docid_val_pairs(&mut self)
where T: Ord {
if self.docid_cache.len() <= 1 {
return;
}
// Quick check: if no consecutive doc_ids are equal, no dedup needed.
let has_multivalue = self.docid_cache.windows(2).any(|w| w[0] == w[1]);
if !has_multivalue {
return;
}
// Sort values within each doc_id group so duplicates become adjacent.
let mut start = 0;
while start < self.docid_cache.len() {
let doc = self.docid_cache[start];
let mut end = start + 1;
while end < self.docid_cache.len() && self.docid_cache[end] == doc {
end += 1;
}
if end - start > 2 {
self.val_cache[start..end].sort();
}
start = end;
}
// Now duplicates are adjacent — deduplicate in place.
let mut write = 0;
for read in 1..self.docid_cache.len() {
if self.docid_cache[read] != self.docid_cache[write]
|| self.val_cache[read] != self.val_cache[write]
{
write += 1;
if write != read {
self.docid_cache[write] = self.docid_cache[read];
self.val_cache[write] = self.val_cache[read];
}
}
}
let new_len = write + 1;
self.docid_cache.truncate(new_len);
self.val_cache.truncate(new_len);
}
#[inline]
pub fn iter_vals(&self) -> impl Iterator<Item = T> + '_ {
self.val_cache.iter().cloned()
@@ -235,56 +163,4 @@ mod tests {
assert_eq!(missing_docs, vec![1, 2, 3, 4, 5]);
}
#[test]
fn test_dedup_docid_val_pairs_consecutive() {
let mut accessor = ColumnBlockAccessor::<u64>::default();
accessor.docid_cache = vec![0, 0, 2, 3];
accessor.val_cache = vec![10, 10, 10, 10];
accessor.dedup_docid_val_pairs();
assert_eq!(accessor.docid_cache, vec![0, 2, 3]);
assert_eq!(accessor.val_cache, vec![10, 10, 10]);
}
#[test]
fn test_dedup_docid_val_pairs_non_consecutive() {
// (0,1), (0,2), (0,1) — duplicate value not adjacent
let mut accessor = ColumnBlockAccessor::<u64>::default();
accessor.docid_cache = vec![0, 0, 0];
accessor.val_cache = vec![1, 2, 1];
accessor.dedup_docid_val_pairs();
assert_eq!(accessor.docid_cache, vec![0, 0]);
assert_eq!(accessor.val_cache, vec![1, 2]);
}
#[test]
fn test_dedup_docid_val_pairs_multi_doc() {
// doc 0: values [3, 1, 3], doc 1: values [5, 5]
let mut accessor = ColumnBlockAccessor::<u64>::default();
accessor.docid_cache = vec![0, 0, 0, 1, 1];
accessor.val_cache = vec![3, 1, 3, 5, 5];
accessor.dedup_docid_val_pairs();
assert_eq!(accessor.docid_cache, vec![0, 0, 1]);
assert_eq!(accessor.val_cache, vec![1, 3, 5]);
}
#[test]
fn test_dedup_docid_val_pairs_no_duplicates() {
let mut accessor = ColumnBlockAccessor::<u64>::default();
accessor.docid_cache = vec![0, 0, 1];
accessor.val_cache = vec![1, 2, 3];
accessor.dedup_docid_val_pairs();
assert_eq!(accessor.docid_cache, vec![0, 0, 1]);
assert_eq!(accessor.val_cache, vec![1, 2, 3]);
}
#[test]
fn test_dedup_docid_val_pairs_single_element() {
let mut accessor = ColumnBlockAccessor::<u64>::default();
accessor.docid_cache = vec![0];
accessor.val_cache = vec![1];
accessor.dedup_docid_val_pairs();
assert_eq!(accessor.docid_cache, vec![0]);
assert_eq!(accessor.val_cache, vec![1]);
}
}

35
columnar/src/column_values/u128_based/compact_space/mod.rs
View File

@@ -448,6 +448,26 @@ impl CompactSpaceDecompressor {
Ok(decompressor)
}
/// Finds the next compact space value for a given u128 value
pub fn u128_to_next_compact(&self, value: u128) -> CompactHit {
// Try to convert to compact space
match self.u128_to_compact(value) {
// Value is in compact space, return its compact representation
Ok(compact) => CompactHit::Exact(compact),
// Value is not in compact space
Err(pos) => {
if pos >= self.params.compact_space.ranges_mapping.len() {
// Value is beyond all ranges, no next value exists
CompactHit::AfterLast
} else {
// Get the next range and return its start compact value
let next_range = &self.params.compact_space.ranges_mapping[pos];
CompactHit::Next(next_range.compact_start)
}
}
}
}
/// Converting to compact space for the decompressor is more complex, since we may get values
/// which are outside the compact space. e.g. if we map
/// 1000 => 5
@@ -459,21 +479,6 @@ impl CompactSpaceDecompressor {
self.params.compact_space.u128_to_compact(value)
}
/// Finds the next compact space value for a given u128 value.
pub fn u128_to_next_compact(&self, value: u128) -> CompactHit {
match self.u128_to_compact(value) {
Ok(compact) => CompactHit::Exact(compact),
Err(pos) => {
if pos >= self.params.compact_space.ranges_mapping.len() {
CompactHit::AfterLast
} else {
let next_range = &self.params.compact_space.ranges_mapping[pos];
CompactHit::Next(next_range.compact_start)
}
}
}
}
fn compact_to_u128(&self, compact: u32) -> u128 {
self.params.compact_space.compact_to_u128(compact)
}

3
common/Cargo.toml
View File

@@ -15,10 +15,11 @@ repository = "https://github.com/quickwit-oss/tantivy"
byteorder = "1.4.3"
ownedbytes = { version= "0.9", path="../ownedbytes" }
async-trait = "0.1"
time = { version = "0.3.47", features = ["serde-well-known"] }
time = { version = "0.3.10", features = ["serde-well-known"] }
serde = { version = "1.0.136", features = ["derive"] }
[dev-dependencies]
binggan = "0.14.0"
proptest = "1.0.0"
rand = "0.9"

4
common/src/writer.rs
View File

@@ -62,9 +62,7 @@ impl<W: TerminatingWrite> TerminatingWrite for CountingWriter<W> {
pub struct AntiCallToken(());
/// Trait used to indicate when no more write need to be done on a writer
///
/// Thread-safety is enforced at the call sites that require it.
pub trait TerminatingWrite: Write {
pub trait TerminatingWrite: Write + Send + Sync {
/// Indicate that the writer will no longer be used. Internally call terminate_ref.
fn terminate(mut self) -> io::Result<()>
where Self: Sized {

27
sketches-ddsketch/Cargo.toml Normal file
View File

@@ -0,0 +1,27 @@
[package]
name = "sketches-ddsketch"
version = "0.3.0"
authors = ["Mike Heffner <mikeh@fesnel.com>"]
edition = "2018"
license = "Apache-2.0"
readme = "README.md"
repository = "https://github.com/mheffner/rust-sketches-ddsketch"
homepage = "https://github.com/mheffner/rust-sketches-ddsketch"
description = """
A direct port of the Golang DDSketch implementation.
"""
exclude = [".gitignore"]
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
serde = { package = "serde", version = "1.0", optional = true, features = ["derive", "serde_derive"] }
[dev-dependencies]
approx = "0.5.1"
rand = "0.8.5"
rand_distr = "0.4.3"
[features]
use_serde = ["serde", "serde/derive"]

201
sketches-ddsketch/LICENSE Normal file
View File

@@ -0,0 +1,201 @@
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [2019] [Mike Heffner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

11
sketches-ddsketch/Makefile Normal file
View File

@@ -0,0 +1,11 @@
clean:
cargo clean
test:
cargo test
test_logs:
cargo test -- --nocapture
test_performance:
cargo test --release --jobs 1 test_performance -- --ignored --nocapture

37
sketches-ddsketch/README.md Normal file
View File

@@ -0,0 +1,37 @@
# sketches-ddsketch
This is a direct port of the [Golang](https://github.com/DataDog/sketches-go)
[DDSketch](https://arxiv.org/pdf/1908.10693.pdf) quantile sketch implementation
to Rust. DDSketch is a fully-mergeable quantile sketch with relative-error
guarantees and is extremely fast.
# DDSketch
* Sketch size automatically grows as needed, starting with 128 bins.
* Extremely fast sample insertion and sketch merges.
## Usage
```rust
use sketches_ddsketch::{Config, DDSketch};
let config = Config::defaults();
let mut sketch = DDSketch::new(c);
sketch.add(1.0);
sketch.add(1.0);
sketch.add(1.0);
// Get p=50%
let quantile = sketch.quantile(0.5).unwrap();
assert_eq!(quantile, Some(1.0));
```
## Performance
No performance tuning has been done with this implementation of the port, so we
would expect similar profiles to the original implementation.
Out of the box we see can achieve over 70M sample inserts/sec and 350K sketch
merges/sec. All tests run on a single core Intel i7 processor with 4.2Ghz max
clock.

98
sketches-ddsketch/src/config.rs Normal file
View File

@@ -0,0 +1,98 @@
#[cfg(feature = "use_serde")]
use serde::{Deserialize, Serialize};
const DEFAULT_MAX_BINS: u32 = 2048;
const DEFAULT_ALPHA: f64 = 0.01;
const DEFAULT_MIN_VALUE: f64 = 1.0e-9;
/// The configuration struct for constructing a `DDSketch`
#[derive(Copy, Clone, Debug, PartialEq)]
#[cfg_attr(feature = "use_serde", derive(Serialize, Deserialize))]
pub struct Config {
pub max_num_bins: u32,
pub gamma: f64,
pub(crate) gamma_ln: f64,
pub(crate) min_value: f64,
pub offset: i32,
}
fn log_gamma(value: f64, gamma_ln: f64) -> f64 {
value.ln() / gamma_ln
}
impl Config {
/// Construct a new `Config` struct with specific parameters. If you are unsure of how to
/// configure this, the `defaults` method constructs a `Config` with built-in defaults.
///
/// `max_num_bins` is the max number of bins the DDSketch will grow to, in steps of 128 bins.
pub fn new(alpha: f64, max_num_bins: u32, min_value: f64) -> Self {
// Aligned with Java's LogarithmicMapping / LogLikeIndexMapping:
// gamma = (1 + alpha) / (1 - alpha) (correctingFactor=1 for LogarithmicMapping)
// gamma_ln = gamma.ln() (not ln_1p, to match Java's Math.log(gamma))
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/mapping/LogLikeIndexMapping.java (gamma() static method)
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/mapping/LogarithmicMapping.java (constructor, correctingFactor()=1)
let gamma = (1.0 + alpha) / (1.0 - alpha);
let gamma_ln = gamma.ln();
Config {
max_num_bins,
gamma,
gamma_ln,
min_value,
offset: 1 - (log_gamma(min_value, gamma_ln) as i32),
}
}
/// Return a `Config` using built-in default settings
pub fn defaults() -> Self {
Self::new(DEFAULT_ALPHA, DEFAULT_MAX_BINS, DEFAULT_MIN_VALUE)
}
pub fn key(&self, v: f64) -> i32 {
// Aligned with Java's LogLikeIndexMapping.index(): floor-based indexing.
// Java uses `(int) index` / `(int) index - 1` which is equivalent to floor().
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/mapping/LogLikeIndexMapping.java (index() method)
self.log_gamma(v).floor() as i32
}
pub fn value(&self, key: i32) -> f64 {
// Aligned with Java's LogLikeIndexMapping.value():
// lowerBound(index) * (1 + relativeAccuracy)
// = logInverse((index - indexOffset) / multiplier) * (1 + relativeAccuracy)
// = gamma^key * 2*gamma/(gamma+1)
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/mapping/LogLikeIndexMapping.java (value() and lowerBound() methods)
self.pow_gamma(key) * (2.0 * self.gamma / (1.0 + self.gamma))
}
pub fn log_gamma(&self, value: f64) -> f64 {
log_gamma(value, self.gamma_ln)
}
pub fn pow_gamma(&self, key: i32) -> f64 {
((key as f64) * self.gamma_ln).exp()
}
pub fn min_possible(&self) -> f64 {
self.min_value
}
/// Reconstruct a Config from a gamma value (as decoded from the binary format).
/// Uses default max_num_bins and min_value.
/// See Java: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/mapping/LogarithmicMapping.java (LogarithmicMapping(double gamma, double indexOffset) constructor)
pub(crate) fn from_gamma(gamma: f64) -> Self {
let gamma_ln = gamma.ln();
Config {
max_num_bins: DEFAULT_MAX_BINS,
gamma,
gamma_ln,
min_value: DEFAULT_MIN_VALUE,
offset: 1 - (log_gamma(DEFAULT_MIN_VALUE, gamma_ln) as i32),
}
}
}
impl Default for Config {
fn default() -> Self {
Self::new(DEFAULT_ALPHA, DEFAULT_MAX_BINS, DEFAULT_MIN_VALUE)
}
}

385
sketches-ddsketch/src/ddsketch.rs Normal file
View File

@@ -0,0 +1,385 @@
use std::{error, fmt};
#[cfg(feature = "use_serde")]
use serde::{Deserialize, Serialize};
use crate::config::Config;
use crate::store::Store;
type Result<T> = std::result::Result<T, DDSketchError>;
/// General error type for DDSketch, represents either an invalid quantile or an
/// incompatible merge operation.
#[derive(Debug, Clone)]
pub enum DDSketchError {
Quantile,
Merge,
}
impl fmt::Display for DDSketchError {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
match self {
DDSketchError::Quantile => {
write!(f, "Invalid quantile, must be between 0 and 1 (inclusive)")
}
DDSketchError::Merge => write!(f, "Can not merge sketches with different configs"),
}
}
}
impl error::Error for DDSketchError {
fn source(&self) -> Option<&(dyn error::Error + 'static)> {
// Generic
None
}
}
/// This struct represents a [DDSketch](https://arxiv.org/pdf/1908.10693.pdf)
#[derive(Clone)]
#[cfg_attr(feature = "use_serde", derive(Serialize, Deserialize))]
pub struct DDSketch {
pub(crate) config: Config,
pub(crate) store: Store,
pub(crate) negative_store: Store,
pub(crate) min: f64,
pub(crate) max: f64,
pub(crate) sum: f64,
pub(crate) zero_count: u64,
}
impl Default for DDSketch {
fn default() -> Self {
Self::new(Default::default())
}
}
// XXX: functions should return Option<> in the case of empty
impl DDSketch {
/// Construct a `DDSketch`. Requires a `Config` specifying the parameters of the sketch
pub fn new(config: Config) -> Self {
DDSketch {
config,
store: Store::new(config.max_num_bins as usize),
negative_store: Store::new(config.max_num_bins as usize),
min: f64::INFINITY,
max: f64::NEG_INFINITY,
sum: 0.0,
zero_count: 0,
}
}
/// Add the sample to the sketch
pub fn add(&mut self, v: f64) {
if v > self.config.min_possible() {
let key = self.config.key(v);
self.store.add(key);
} else if v < -self.config.min_possible() {
let key = self.config.key(-v);
self.negative_store.add(key);
} else {
self.zero_count += 1;
}
if v < self.min {
self.min = v;
}
if self.max < v {
self.max = v;
}
self.sum += v;
}
/// Return the quantile value for quantiles between 0.0 and 1.0. Result is an error, represented
/// as DDSketchError::Quantile if the requested quantile is outside of that range.
///
/// If the sketch is empty the result is None, else Some(v) for the quantile value.
pub fn quantile(&self, q: f64) -> Result<Option<f64>> {
if !(0.0..=1.0).contains(&q) {
return Err(DDSketchError::Quantile);
}
if self.empty() {
return Ok(None);
}
if q == 0.0 {
return Ok(Some(self.min));
} else if q == 1.0 {
return Ok(Some(self.max));
}
let rank = (q * (self.count() as f64 - 1.0)) as u64;
let quantile;
if rank < self.negative_store.count() {
let reversed_rank = self.negative_store.count() - rank - 1;
let key = self.negative_store.key_at_rank(reversed_rank);
quantile = -self.config.value(key);
} else if rank < self.zero_count + self.negative_store.count() {
quantile = 0.0;
} else {
let key = self
.store
.key_at_rank(rank - self.zero_count - self.negative_store.count());
quantile = self.config.value(key);
}
Ok(Some(quantile))
}
/// Returns the minimum value seen, or None if sketch is empty
pub fn min(&self) -> Option<f64> {
if self.empty() {
None
} else {
Some(self.min)
}
}
/// Returns the maximum value seen, or None if sketch is empty
pub fn max(&self) -> Option<f64> {
if self.empty() {
None
} else {
Some(self.max)
}
}
/// Returns the sum of values seen, or None if sketch is empty
pub fn sum(&self) -> Option<f64> {
if self.empty() {
None
} else {
Some(self.sum)
}
}
/// Returns the number of values added to the sketch
pub fn count(&self) -> usize {
(self.store.count() + self.zero_count + self.negative_store.count()) as usize
}
/// Returns the length of the underlying `Store`. This is mainly only useful for understanding
/// how much the sketch has grown given the inserted values.
pub fn length(&self) -> usize {
self.store.length() as usize + self.negative_store.length() as usize
}
/// Merge the contents of another sketch into this one. The sketch that is merged into this one
/// is unchanged after the merge.
pub fn merge(&mut self, o: &DDSketch) -> Result<()> {
if self.config != o.config {
return Err(DDSketchError::Merge);
}
let was_empty = self.store.count() == 0;
// Merge the stores
self.store.merge(&o.store);
self.negative_store.merge(&o.negative_store);
self.zero_count += o.zero_count;
// Need to ensure we don't override min/max with initializers
// if either store were empty
if was_empty {
self.min = o.min;
self.max = o.max;
} else if o.store.count() > 0 {
if o.min < self.min {
self.min = o.min
}
if o.max > self.max {
self.max = o.max;
}
}
self.sum += o.sum;
Ok(())
}
fn empty(&self) -> bool {
self.count() == 0
}
/// Encode this sketch into the Java-compatible binary format used by
/// `com.datadoghq.sketch.ddsketch.DDSketchWithExactSummaryStatistics`.
pub fn to_java_bytes(&self) -> Vec<u8> {
crate::encoding::encode_to_java_bytes(self)
}
/// Decode a sketch from the Java-compatible binary format.
/// Accepts bytes produced by Java's `DDSketchWithExactSummaryStatistics.encode()`
/// with or without the `0x02` version prefix.
pub fn from_java_bytes(
bytes: &[u8],
) -> std::result::Result<Self, crate::encoding::DecodeError> {
crate::encoding::decode_from_java_bytes(bytes)
}
}
#[cfg(test)]
mod tests {
use approx::assert_relative_eq;
use crate::{Config, DDSketch};
#[test]
fn test_add_zero() {
let alpha = 0.01;
let c = Config::new(alpha, 2048, 10e-9);
let mut dd = DDSketch::new(c);
dd.add(0.0);
}
#[test]
fn test_quartiles() {
let alpha = 0.01;
let c = Config::new(alpha, 2048, 10e-9);
let mut dd = DDSketch::new(c);
// Initialize sketch with {1.0, 2.0, 3.0, 4.0}
for i in 1..5 {
dd.add(i as f64);
}
// We expect the following mappings from quantile to value:
// [0,0.33]: 1.0, (0.34,0.66]: 2.0, (0.67,0.99]: 3.0, (0.99, 1.0]: 4.0
let test_cases = vec![
(0.0, 1.0),
(0.25, 1.0),
(0.33, 1.0),
(0.34, 2.0),
(0.5, 2.0),
(0.66, 2.0),
(0.67, 3.0),
(0.75, 3.0),
(0.99, 3.0),
(1.0, 4.0),
];
for (q, val) in test_cases {
assert_relative_eq!(dd.quantile(q).unwrap().unwrap(), val, max_relative = alpha);
}
}
#[test]
fn test_neg_quartiles() {
let alpha = 0.01;
let c = Config::new(alpha, 2048, 10e-9);
let mut dd = DDSketch::new(c);
// Initialize sketch with {1.0, 2.0, 3.0, 4.0}
for i in 1..5 {
dd.add(-i as f64);
}
let test_cases = vec![
(0.0, -4.0),
(0.25, -4.0),
(0.5, -3.0),
(0.75, -2.0),
(1.0, -1.0),
];
for (q, val) in test_cases {
assert_relative_eq!(dd.quantile(q).unwrap().unwrap(), val, max_relative = alpha);
}
}
#[test]
fn test_simple_quantile() {
let c = Config::defaults();
let mut dd = DDSketch::new(c);
for i in 1..101 {
dd.add(i as f64);
}
assert_eq!(dd.quantile(0.95).unwrap().unwrap().ceil(), 95.0);
assert!(dd.quantile(-1.01).is_err());
assert!(dd.quantile(1.01).is_err());
}
#[test]
fn test_empty_sketch() {
let c = Config::defaults();
let dd = DDSketch::new(c);
assert_eq!(dd.quantile(0.98).unwrap(), None);
assert_eq!(dd.max(), None);
assert_eq!(dd.min(), None);
assert_eq!(dd.sum(), None);
assert_eq!(dd.count(), 0);
assert!(dd.quantile(1.01).is_err());
}
#[test]
fn test_basic_histogram_data() {
let values = &[
0.754225035,
0.752900282,
0.752812246,
0.752602367,
0.754310155,
0.753525981,
0.752981082,
0.752715536,
0.751667941,
0.755079054,
0.753528150,
0.755188464,
0.752508723,
0.750064549,
0.753960428,
0.751139298,
0.752523560,
0.753253428,
0.753498342,
0.751858358,
0.752104636,
0.753841300,
0.754467374,
0.753814334,
0.750881719,
0.753182556,
0.752576884,
0.753945708,
0.753571911,
0.752314573,
0.752586651,
];
let c = Config::defaults();
let mut dd = DDSketch::new(c);
for value in values {
dd.add(*value);
}
assert_eq!(dd.max(), Some(0.755188464));
assert_eq!(dd.min(), Some(0.750064549));
assert_eq!(dd.count(), 31);
assert_eq!(dd.sum(), Some(23.343630625000003));
assert!(dd.quantile(0.25).unwrap().is_some());
assert!(dd.quantile(0.5).unwrap().is_some());
assert!(dd.quantile(0.75).unwrap().is_some());
}
#[test]
fn test_length() {
let mut dd = DDSketch::default();
assert_eq!(dd.length(), 0);
dd.add(1.0);
assert_eq!(dd.length(), 128);
dd.add(2.0);
dd.add(3.0);
assert_eq!(dd.length(), 128);
dd.add(-1.0);
assert_eq!(dd.length(), 256);
dd.add(-2.0);
dd.add(-3.0);
assert_eq!(dd.length(), 256);
}
}

813
sketches-ddsketch/src/encoding.rs Normal file
View File

@@ -0,0 +1,813 @@
//! Java-compatible binary encoding/decoding for DDSketch.
//!
//! This module implements the binary format used by the Java
//! `com.datadoghq.sketch.ddsketch.DDSketchWithExactSummaryStatistics` class
//! from the DataDog/sketches-java library. It enables cross-language
//! serialization so that sketches produced in Rust can be deserialized
//! and merged by Java consumers.
use std::fmt;
use crate::config::Config;
use crate::ddsketch::DDSketch;
use crate::store::Store;
// ---------------------------------------------------------------------------
// Flag byte layout
//
// Each flag byte packs a 2-bit type ordinal in the low bits and a 6-bit
// subflag in the upper bits: (subflag << 2) | type_ordinal
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/encoding/Flag.java
// ---------------------------------------------------------------------------
/// The 2-bit type field occupying the low bits of every flag byte.
#[repr(u8)]
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
enum FlagType {
SketchFeatures = 0,
PositiveStore = 1,
IndexMapping = 2,
NegativeStore = 3,
}
impl FlagType {
fn from_byte(b: u8) -> Option<Self> {
match b & 0x03 {
0 => Some(Self::SketchFeatures),
1 => Some(Self::PositiveStore),
2 => Some(Self::IndexMapping),
3 => Some(Self::NegativeStore),
_ => None,
}
}
}
/// Construct a flag byte from a subflag and a type.
const fn flag(subflag: u8, flag_type: FlagType) -> u8 {
(subflag << 2) | (flag_type as u8)
}
// Pre-computed flag bytes for the sketch features we encode/decode.
const FLAG_INDEX_MAPPING_LOG: u8 = flag(0, FlagType::IndexMapping); // 0x02
const FLAG_ZERO_COUNT: u8 = flag(1, FlagType::SketchFeatures); // 0x04
const FLAG_COUNT: u8 = flag(0x28, FlagType::SketchFeatures); // 0xA0
const FLAG_SUM: u8 = flag(0x21, FlagType::SketchFeatures); // 0x84
const FLAG_MIN: u8 = flag(0x22, FlagType::SketchFeatures); // 0x88
const FLAG_MAX: u8 = flag(0x23, FlagType::SketchFeatures); // 0x8C
/// BinEncodingMode subflags for store flag bytes.
/// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/encoding/BinEncodingMode.java
#[repr(u8)]
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
enum BinEncodingMode {
IndexDeltasAndCounts = 1,
IndexDeltas = 2,
ContiguousCounts = 3,
}
impl BinEncodingMode {
fn from_subflag(subflag: u8) -> Option<Self> {
match subflag {
1 => Some(Self::IndexDeltasAndCounts),
2 => Some(Self::IndexDeltas),
3 => Some(Self::ContiguousCounts),
_ => None,
}
}
}
const VAR_DOUBLE_ROTATE_DISTANCE: u32 = 6;
const MAX_VAR_LEN_64: usize = 9;
const DEFAULT_MAX_BINS: u32 = 2048;
// ---------------------------------------------------------------------------
// Error type
// ---------------------------------------------------------------------------
#[derive(Debug, Clone)]
pub enum DecodeError {
UnexpectedEof,
InvalidFlag(u8),
InvalidData(String),
}
impl fmt::Display for DecodeError {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match self {
Self::UnexpectedEof => write!(f, "unexpected end of input"),
Self::InvalidFlag(b) => write!(f, "invalid flag byte: 0x{b:02X}"),
Self::InvalidData(msg) => write!(f, "invalid data: {msg}"),
}
}
}
impl std::error::Error for DecodeError {}
// ---------------------------------------------------------------------------
// VarEncoding — bit-exact port of Java VarEncodingHelper
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/encoding/VarEncodingHelper.java
// ---------------------------------------------------------------------------
fn encode_unsigned_var_long(out: &mut Vec<u8>, mut value: u64) {
let length = ((63 - value.leading_zeros() as i32) / 7).clamp(0, 8);
for _ in 0..length {
out.push((value as u8) | 0x80);
value >>= 7;
}
out.push(value as u8);
}
fn decode_unsigned_var_long(input: &mut &[u8]) -> Result<u64, DecodeError> {
let mut value: u64 = 0;
let mut shift: u32 = 0;
loop {
let next = read_byte(input)?;
if next < 0x80 || shift == 56 {
return Ok(value | (u64::from(next) << shift));
}
value |= (u64::from(next) & 0x7F) << shift;
shift += 7;
}
}
/// ZigZag encode then var-long encode.
fn encode_signed_var_long(out: &mut Vec<u8>, value: i64) {
let encoded = ((value >> 63) ^ (value << 1)) as u64;
encode_unsigned_var_long(out, encoded);
}
fn decode_signed_var_long(input: &mut &[u8]) -> Result<i64, DecodeError> {
let encoded = decode_unsigned_var_long(input)?;
Ok(((encoded >> 1) as i64) ^ -((encoded & 1) as i64))
}
fn double_to_var_bits(value: f64) -> u64 {
let bits = f64::to_bits(value + 1.0).wrapping_sub(f64::to_bits(1.0));
bits.rotate_left(VAR_DOUBLE_ROTATE_DISTANCE)
}
fn var_bits_to_double(bits: u64) -> f64 {
f64::from_bits(
bits.rotate_right(VAR_DOUBLE_ROTATE_DISTANCE)
.wrapping_add(f64::to_bits(1.0)),
) - 1.0
}
fn encode_var_double(out: &mut Vec<u8>, value: f64) {
let mut bits = double_to_var_bits(value);
for _ in 0..MAX_VAR_LEN_64 - 1 {
let next = (bits >> 57) as u8;
bits <<= 7;
if bits == 0 {
out.push(next);
return;
}
out.push(next | 0x80);
}
out.push((bits >> 56) as u8);
}
fn decode_var_double(input: &mut &[u8]) -> Result<f64, DecodeError> {
let mut bits: u64 = 0;
let mut shift: i32 = 57; // 8*8 - 7
loop {
let next = read_byte(input)?;
if shift == 1 {
bits |= u64::from(next);
break;
}
if next < 0x80 {
bits |= u64::from(next) << shift;
break;
}
bits |= (u64::from(next) & 0x7F) << shift;
shift -= 7;
}
Ok(var_bits_to_double(bits))
}
// ---------------------------------------------------------------------------
// Byte-level helpers
// ---------------------------------------------------------------------------
fn read_byte(input: &mut &[u8]) -> Result<u8, DecodeError> {
match input.split_first() {
Some((&byte, rest)) => {
*input = rest;
Ok(byte)
}
None => Err(DecodeError::UnexpectedEof),
}
}
fn write_f64_le(out: &mut Vec<u8>, value: f64) {
out.extend_from_slice(&value.to_le_bytes());
}
fn read_f64_le(input: &mut &[u8]) -> Result<f64, DecodeError> {
if input.len() < 8 {
return Err(DecodeError::UnexpectedEof);
}
let (bytes, rest) = input.split_at(8);
*input = rest;
// bytes is guaranteed to be length 8 by the split_at above.
let arr = [
bytes[0], bytes[1], bytes[2], bytes[3], bytes[4], bytes[5], bytes[6], bytes[7],
];
Ok(f64::from_le_bytes(arr))
}
// ---------------------------------------------------------------------------
// Store encoding/decoding
// See: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/store/DenseStore.java (encode/decode methods)
// ---------------------------------------------------------------------------
/// Collect non-zero bins in the store as (absolute_index, count) pairs.
///
/// Allocation is acceptable here: this runs once per encode and the Vec
/// has at most `max_num_bins` entries.
fn collect_non_zero_bins(store: &Store) -> Vec<(i32, u64)> {
if store.count == 0 {
return Vec::new();
}
let start = (store.min_key - store.offset) as usize;
let end = ((store.max_key - store.offset + 1) as usize).min(store.bins.len());
store.bins[start..end]
.iter()
.enumerate()
.filter(|&(_, &count)| count > 0)
.map(|(i, &count)| (start as i32 + i as i32 + store.offset, count))
.collect()
}
fn encode_store(out: &mut Vec<u8>, store: &Store, flag_type: FlagType) {
let bins = collect_non_zero_bins(store);
if bins.is_empty() {
return;
}
out.push(flag(BinEncodingMode::IndexDeltasAndCounts as u8, flag_type));
encode_unsigned_var_long(out, bins.len() as u64);
let mut prev_index: i64 = 0;
for &(index, count) in &bins {
encode_signed_var_long(out, i64::from(index) - prev_index);
encode_var_double(out, count as f64);
prev_index = i64::from(index);
}
}
fn decode_store(input: &mut &[u8], subflag: u8, bin_limit: usize) -> Result<Store, DecodeError> {
let mode = BinEncodingMode::from_subflag(subflag).ok_or_else(|| {
DecodeError::InvalidData(format!("unknown bin encoding mode subflag: {subflag}"))
})?;
let num_bins = decode_unsigned_var_long(input)? as usize;
let mut store = Store::new(bin_limit);
match mode {
BinEncodingMode::IndexDeltasAndCounts => {
let mut index: i64 = 0;
for _ in 0..num_bins {
index += decode_signed_var_long(input)?;
let count = decode_var_double(input)?;
store.add_count(index as i32, count as u64);
}
}
BinEncodingMode::IndexDeltas => {
let mut index: i64 = 0;
for _ in 0..num_bins {
index += decode_signed_var_long(input)?;
store.add_count(index as i32, 1);
}
}
BinEncodingMode::ContiguousCounts => {
let start_index = decode_signed_var_long(input)?;
let index_delta = decode_signed_var_long(input)?;
let mut index = start_index;
for _ in 0..num_bins {
let count = decode_var_double(input)?;
store.add_count(index as i32, count as u64);
index += index_delta;
}
}
}
Ok(store)
}
// ---------------------------------------------------------------------------
// Top-level encode / decode
// ---------------------------------------------------------------------------
/// Encode a DDSketch into the Java-compatible binary format.
///
/// The output follows the encoding order of
/// `DDSketchWithExactSummaryStatistics.encode()` then `DDSketch.encode()`:
///
/// 1. Summary statistics: COUNT, MIN, MAX (if count > 0)
/// 2. SUM (if sum != 0)
/// 3. Index mapping (LOG layout): gamma, indexOffset
/// 4. Zero count (if > 0)
/// 5. Positive store bins
/// 6. Negative store bins
pub fn encode_to_java_bytes(sketch: &DDSketch) -> Vec<u8> {
let mut out = Vec::new();
let count = sketch.count() as f64;
// Summary statistics (DDSketchWithExactSummaryStatistics.encode)
if count != 0.0 {
out.push(FLAG_COUNT);
encode_var_double(&mut out, count);
out.push(FLAG_MIN);
write_f64_le(&mut out, sketch.min);
out.push(FLAG_MAX);
write_f64_le(&mut out, sketch.max);
}
if sketch.sum != 0.0 {
out.push(FLAG_SUM);
write_f64_le(&mut out, sketch.sum);
}
// DDSketch.encode: index mapping + zero count + stores
out.push(FLAG_INDEX_MAPPING_LOG);
write_f64_le(&mut out, sketch.config.gamma);
write_f64_le(&mut out, 0.0_f64);
if sketch.zero_count != 0 {
out.push(FLAG_ZERO_COUNT);
encode_var_double(&mut out, sketch.zero_count as f64);
}
encode_store(&mut out, &sketch.store, FlagType::PositiveStore);
encode_store(&mut out, &sketch.negative_store, FlagType::NegativeStore);
out
}
/// Decode a DDSketch from the Java-compatible binary format.
///
/// Accepts bytes with or without a `0x02` version prefix.
pub fn decode_from_java_bytes(bytes: &[u8]) -> Result<DDSketch, DecodeError> {
if bytes.is_empty() {
return Err(DecodeError::UnexpectedEof);
}
let mut input = bytes;
// Skip optional version prefix (0x02 followed by a valid flag byte).
if input.len() >= 2 && input[0] == 0x02 && is_valid_flag_byte(input[1]) {
input = &input[1..];
}
let mut gamma: Option<f64> = None;
let mut zero_count: f64 = 0.0;
let mut sum: f64 = 0.0;
let mut min: f64 = f64::INFINITY;
let mut max: f64 = f64::NEG_INFINITY;
let mut positive_store: Option<Store> = None;
let mut negative_store: Option<Store> = None;
while !input.is_empty() {
let flag_byte = read_byte(&mut input)?;
let flag_type =
FlagType::from_byte(flag_byte).ok_or(DecodeError::InvalidFlag(flag_byte))?;
let subflag = flag_byte >> 2;
match flag_type {
FlagType::IndexMapping => {
gamma = Some(read_f64_le(&mut input)?);
let _index_offset = read_f64_le(&mut input)?;
}
FlagType::SketchFeatures => match flag_byte {
FLAG_ZERO_COUNT => zero_count += decode_var_double(&mut input)?,
FLAG_COUNT => {
let _count = decode_var_double(&mut input)?;
}
FLAG_SUM => sum = read_f64_le(&mut input)?,
FLAG_MIN => min = read_f64_le(&mut input)?,
FLAG_MAX => max = read_f64_le(&mut input)?,
_ => return Err(DecodeError::InvalidFlag(flag_byte)),
},
FlagType::PositiveStore => {
positive_store = Some(decode_store(
&mut input,
subflag,
DEFAULT_MAX_BINS as usize,
)?);
}
FlagType::NegativeStore => {
negative_store = Some(decode_store(
&mut input,
subflag,
DEFAULT_MAX_BINS as usize,
)?);
}
}
}
let g = gamma.unwrap_or_else(|| Config::defaults().gamma);
let config = Config::from_gamma(g);
let store = positive_store.unwrap_or_else(|| Store::new(config.max_num_bins as usize));
let neg = negative_store.unwrap_or_else(|| Store::new(config.max_num_bins as usize));
Ok(DDSketch {
config,
store,
negative_store: neg,
min,
max,
sum,
zero_count: zero_count as u64,
})
}
/// Check whether a byte is a valid flag byte for the DDSketch binary format.
fn is_valid_flag_byte(b: u8) -> bool {
// Known sketch-feature flags
if matches!(
b,
FLAG_ZERO_COUNT | FLAG_COUNT | FLAG_SUM | FLAG_MIN | FLAG_MAX | FLAG_INDEX_MAPPING_LOG
) {
return true;
}
let Some(flag_type) = FlagType::from_byte(b) else {
return false;
};
let subflag = b >> 2;
match flag_type {
FlagType::PositiveStore | FlagType::NegativeStore => (1..=3).contains(&subflag),
FlagType::IndexMapping => subflag <= 4, // LOG=0, LOG_LINEAR=1 .. LOG_QUARTIC=4
_ => false,
}
}
// ---------------------------------------------------------------------------
// Tests
// ---------------------------------------------------------------------------
#[cfg(test)]
mod tests {
use super::*;
use crate::{Config, DDSketch};
// --- VarEncoding unit tests ---
#[test]
fn test_unsigned_var_long_zero() {
let mut buf = Vec::new();
encode_unsigned_var_long(&mut buf, 0);
assert_eq!(buf, [0x00]);
let mut input = buf.as_slice();
assert_eq!(decode_unsigned_var_long(&mut input).unwrap(), 0);
assert!(input.is_empty());
}
#[test]
fn test_unsigned_var_long_small() {
let mut buf = Vec::new();
encode_unsigned_var_long(&mut buf, 1);
assert_eq!(buf, [0x01]);
let mut input = buf.as_slice();
assert_eq!(decode_unsigned_var_long(&mut input).unwrap(), 1);
}
#[test]
fn test_unsigned_var_long_128() {
let mut buf = Vec::new();
encode_unsigned_var_long(&mut buf, 128);
assert_eq!(buf, [0x80, 0x01]);
let mut input = buf.as_slice();
assert_eq!(decode_unsigned_var_long(&mut input).unwrap(), 128);
}
#[test]
fn test_unsigned_var_long_roundtrip() {
for v in [0u64, 1, 127, 128, 255, 256, 16383, 16384, u64::MAX] {
let mut buf = Vec::new();
encode_unsigned_var_long(&mut buf, v);
let mut input = buf.as_slice();
let decoded = decode_unsigned_var_long(&mut input).unwrap();
assert_eq!(decoded, v, "roundtrip failed for {}", v);
assert!(input.is_empty());
}
}
#[test]
fn test_signed_var_long_roundtrip() {
for v in [0i64, 1, -1, 63, -64, 64, -65, i64::MAX, i64::MIN] {
let mut buf = Vec::new();
encode_signed_var_long(&mut buf, v);
let mut input = buf.as_slice();
let decoded = decode_signed_var_long(&mut input).unwrap();
assert_eq!(decoded, v, "roundtrip failed for {}", v);
assert!(input.is_empty());
}
}
#[test]
fn test_var_double_roundtrip() {
for v in [0.0, 1.0, 2.0, 5.0, 15.0, 42.0, 100.0, 1e-9, 1e15, 0.5, 7.77] {
let mut buf = Vec::new();
encode_var_double(&mut buf, v);
let mut input = buf.as_slice();
let decoded = decode_var_double(&mut input).unwrap();
assert!(
(decoded - v).abs() < 1e-15 || decoded == v,
"roundtrip failed for {}: got {}",
v,
decoded,
);
assert!(input.is_empty());
}
}
#[test]
fn test_var_double_small_integers() {
let mut buf = Vec::new();
encode_var_double(&mut buf, 1.0);
assert_eq!(buf.len(), 1, "VarDouble(1.0) should be 1 byte");
buf.clear();
encode_var_double(&mut buf, 5.0);
assert_eq!(buf.len(), 1, "VarDouble(5.0) should be 1 byte");
}
// --- DDSketch encode/decode roundtrip tests ---
#[test]
fn test_encode_empty_sketch() {
let sketch = DDSketch::new(Config::defaults());
let bytes = sketch.to_java_bytes();
assert!(!bytes.is_empty());
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 0);
assert_eq!(decoded.min(), None);
assert_eq!(decoded.max(), None);
assert_eq!(decoded.sum(), None);
}
#[test]
fn test_encode_simple_sketch() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [1.0, 2.0, 3.0, 4.0, 5.0] {
sketch.add(v);
}
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 5);
assert_eq!(decoded.min(), Some(1.0));
assert_eq!(decoded.max(), Some(5.0));
assert_eq!(decoded.sum(), Some(15.0));
assert_quantiles_match(&sketch, &decoded, &[0.5, 0.9, 0.95, 0.99]);
}
#[test]
fn test_encode_single_value() {
let mut sketch = DDSketch::new(Config::defaults());
sketch.add(42.0);
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 1);
assert_eq!(decoded.min(), Some(42.0));
assert_eq!(decoded.max(), Some(42.0));
assert_eq!(decoded.sum(), Some(42.0));
}
#[test]
fn test_encode_negative_values() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [-3.0, -1.0, 2.0, 5.0] {
sketch.add(v);
}
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 4);
assert_eq!(decoded.min(), Some(-3.0));
assert_eq!(decoded.max(), Some(5.0));
assert_eq!(decoded.sum(), Some(3.0));
assert_quantiles_match(&sketch, &decoded, &[0.0, 0.25, 0.5, 0.75, 1.0]);
}
#[test]
fn test_encode_with_zero_value() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [0.0, 1.0, 2.0] {
sketch.add(v);
}
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 3);
assert_eq!(decoded.min(), Some(0.0));
assert_eq!(decoded.max(), Some(2.0));
assert_eq!(decoded.sum(), Some(3.0));
assert_eq!(decoded.zero_count, 1);
}
#[test]
fn test_encode_large_range() {
let mut sketch = DDSketch::new(Config::defaults());
sketch.add(0.001);
sketch.add(1_000_000.0);
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 2);
assert_eq!(decoded.min(), Some(0.001));
assert_eq!(decoded.max(), Some(1_000_000.0));
}
#[test]
fn test_encode_with_version_prefix() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [1.0, 2.0, 3.0] {
sketch.add(v);
}
let bytes = sketch.to_java_bytes();
// Simulate Java's toByteArrayV2: prepend 0x02
let mut v2_bytes = vec![0x02];
v2_bytes.extend_from_slice(&bytes);
let decoded = DDSketch::from_java_bytes(&v2_bytes).unwrap();
assert_eq!(decoded.count(), 3);
assert_eq!(decoded.min(), Some(1.0));
assert_eq!(decoded.max(), Some(3.0));
}
#[test]
fn test_byte_level_encoding() {
let mut sketch = DDSketch::new(Config::defaults());
sketch.add(1.0);
let bytes = sketch.to_java_bytes();
assert_eq!(bytes[0], FLAG_COUNT, "first byte should be COUNT flag");
assert!(
bytes.contains(&FLAG_INDEX_MAPPING_LOG),
"should contain index mapping flag"
);
}
// --- Cross-language golden byte tests ---
//
// Golden bytes generated by Java's DDSketchWithExactSummaryStatistics.encode()
// using LogarithmicMapping(0.01) + CollapsingLowestDenseStore(2048).
const GOLDEN_SIMPLE: &str = "a00588000000000000f03f8c0000000000001440840000000000002e4002fd4a815abf52f03f000000000000000005050002440228021e021602";
const GOLDEN_SINGLE: &str = "a0028800000000000045408c000000000000454084000000000000454002fd4a815abf52f03f00000000000000000501f40202";
const GOLDEN_NEGATIVE: &str = "a084408800000000000008c08c000000000000144084000000000000084002fd4a815abf52f03f0000000000000000050244025c02070200026c02";
const GOLDEN_ZERO: &str = "a0048800000000000000008c000000000000004084000000000000084002fd4a815abf52f03f00000000000000000402050200024402";
const GOLDEN_EMPTY: &str = "02fd4a815abf52f03f0000000000000000";
const GOLDEN_MANY: &str = "a08d1488000000000000f03f8c0000000000005940840000000000bab34002fd4a815abf52f03f000000000000000005550002440228021e021602120210020c020c020c0208020a020802060208020602060206020602040206020402040204020402040204020402040204020202040202020402020204020202020204020202020202020402020202020202020202020202020202020202020202020202020202020203020202020202020302020202020302020202020302020203020202030202020302030202020302030203020202030203020302030202";
fn hex_to_bytes(hex: &str) -> Vec<u8> {
(0..hex.len())
.step_by(2)
.map(|i| u8::from_str_radix(&hex[i..i + 2], 16).unwrap())
.collect()
}
fn bytes_to_hex(bytes: &[u8]) -> String {
bytes.iter().map(|b| format!("{b:02x}")).collect()
}
fn assert_golden(label: &str, sketch: &DDSketch, golden_hex: &str) {
let bytes = sketch.to_java_bytes();
let expected = hex_to_bytes(golden_hex);
assert_eq!(
bytes,
expected,
"Rust encoding doesn't match Java golden bytes for {}.\nRust: {}\nJava: {}",
label,
bytes_to_hex(&bytes),
golden_hex,
);
}
fn assert_quantiles_match(a: &DDSketch, b: &DDSketch, quantiles: &[f64]) {
for &q in quantiles {
let va = a.quantile(q).unwrap().unwrap();
let vb = b.quantile(q).unwrap().unwrap();
assert!(
(va - vb).abs() / va.abs().max(1e-15) < 1e-12,
"quantile({}) mismatch: {} vs {}",
q,
va,
vb,
);
}
}
#[test]
fn test_cross_language_simple() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [1.0, 2.0, 3.0, 4.0, 5.0] {
sketch.add(v);
}
assert_golden("SIMPLE", &sketch, GOLDEN_SIMPLE);
}
#[test]
fn test_cross_language_single() {
let mut sketch = DDSketch::new(Config::defaults());
sketch.add(42.0);
assert_golden("SINGLE", &sketch, GOLDEN_SINGLE);
}
#[test]
fn test_cross_language_negative() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [-3.0, -1.0, 2.0, 5.0] {
sketch.add(v);
}
assert_golden("NEGATIVE", &sketch, GOLDEN_NEGATIVE);
}
#[test]
fn test_cross_language_zero() {
let mut sketch = DDSketch::new(Config::defaults());
for v in [0.0, 1.0, 2.0] {
sketch.add(v);
}
assert_golden("ZERO", &sketch, GOLDEN_ZERO);
}
#[test]
fn test_cross_language_empty() {
let sketch = DDSketch::new(Config::defaults());
assert_golden("EMPTY", &sketch, GOLDEN_EMPTY);
}
#[test]
fn test_cross_language_many() {
let mut sketch = DDSketch::new(Config::defaults());
for i in 1..=100 {
sketch.add(i as f64);
}
assert_golden("MANY", &sketch, GOLDEN_MANY);
}
#[test]
fn test_decode_java_golden_bytes() {
for (name, hex) in [
("SIMPLE", GOLDEN_SIMPLE),
("SINGLE", GOLDEN_SINGLE),
("NEGATIVE", GOLDEN_NEGATIVE),
("ZERO", GOLDEN_ZERO),
("EMPTY", GOLDEN_EMPTY),
("MANY", GOLDEN_MANY),
] {
let bytes = hex_to_bytes(hex);
let result = DDSketch::from_java_bytes(&bytes);
assert!(
result.is_ok(),
"failed to decode {}: {:?}",
name,
result.err()
);
}
}
#[test]
fn test_encode_decode_many_values() {
let mut sketch = DDSketch::new(Config::defaults());
for i in 1..=100 {
sketch.add(i as f64);
}
let bytes = sketch.to_java_bytes();
let decoded = DDSketch::from_java_bytes(&bytes).unwrap();
assert_eq!(decoded.count(), 100);
assert_eq!(decoded.min(), Some(1.0));
assert_eq!(decoded.max(), Some(100.0));
assert_eq!(decoded.sum(), Some(5050.0));
let alpha = 0.01;
let orig_p95 = sketch.quantile(0.95).unwrap().unwrap();
let dec_p95 = decoded.quantile(0.95).unwrap().unwrap();
assert!(
(orig_p95 - dec_p95).abs() / orig_p95 < alpha,
"p95 mismatch: {} vs {}",
orig_p95,
dec_p95,
);
}
}

52
sketches-ddsketch/src/lib.rs Normal file
View File

@@ -0,0 +1,52 @@
//! This crate provides a direct port of the [Golang](https://github.com/DataDog/sketches-go)
//! [DDSketch](https://arxiv.org/pdf/1908.10693.pdf) implementation to Rust. All efforts
//! have been made to keep this as close to the original implementation as possible, with a few
//! tweaks to get closer to idiomatic Rust.
//!
//! # Usage
//!
//! Add multiple samples to a DDSketch and invoke the `quantile` method to pull any quantile from
//! 0.0* to *1.0*.
//!
//! ```rust
//! use sketches_ddsketch::{Config, DDSketch};
//!
//! let c = Config::defaults();
//! let mut d = DDSketch::new(c);
//!
//! d.add(1.0);
//! d.add(1.0);
//! d.add(1.0);
//!
//! let q = d.quantile(0.50).unwrap();
//!
//! assert!(q < Some(1.02));
//! assert!(q > Some(0.98));
//! ```
//!
//! Sketches can also be merged.
//!
//! ```rust
//! use sketches_ddsketch::{Config, DDSketch};
//!
//! let c = Config::defaults();
//! let mut d1 = DDSketch::new(c);
//! let mut d2 = DDSketch::new(c);
//!
//! d1.add(1.0);
//! d2.add(2.0);
//! d2.add(2.0);
//!
//! d1.merge(&d2);
//!
//! assert_eq!(d1.count(), 3);
//! ```
pub use self::config::Config;
pub use self::ddsketch::{DDSketch, DDSketchError};
pub use self::encoding::DecodeError;
mod config;
mod ddsketch;
pub mod encoding;
mod store;

252
sketches-ddsketch/src/store.rs Normal file
View File

@@ -0,0 +1,252 @@
#[cfg(feature = "use_serde")]
use serde::{Deserialize, Serialize};
const CHUNK_SIZE: i32 = 128;
// Divide the `dividend` by the `divisor`, rounding towards positive infinity.
//
// Similar to the nightly only `std::i32::div_ceil`.
fn div_ceil(dividend: i32, divisor: i32) -> i32 {
(dividend + divisor - 1) / divisor
}
/// CollapsingLowestDenseStore
#[derive(Clone, Debug)]
#[cfg_attr(feature = "use_serde", derive(Serialize, Deserialize))]
pub struct Store {
pub(crate) bins: Vec<u64>,
pub(crate) count: u64,
pub(crate) min_key: i32,
pub(crate) max_key: i32,
pub(crate) offset: i32,
pub(crate) bin_limit: usize,
is_collapsed: bool,
}
impl Store {
pub fn new(bin_limit: usize) -> Self {
Store {
bins: Vec::new(),
count: 0,
min_key: i32::MAX,
max_key: i32::MIN,
offset: 0,
bin_limit,
is_collapsed: false,
}
}
/// Return the number of bins.
pub fn length(&self) -> i32 {
self.bins.len() as i32
}
pub fn is_empty(&self) -> bool {
self.bins.is_empty()
}
pub fn add(&mut self, key: i32) {
let idx = self.get_index(key);
self.bins[idx] += 1;
self.count += 1;
}
/// See Java: https://github.com/DataDog/sketches-java/blob/master/src/main/java/com/datadoghq/sketch/ddsketch/store/DenseStore.java (add(int index, double count) method)
pub(crate) fn add_count(&mut self, key: i32, count: u64) {
let idx = self.get_index(key);
self.bins[idx] += count;
self.count += count;
}
fn get_index(&mut self, key: i32) -> usize {
if key < self.min_key {
if self.is_collapsed {
return 0;
}
self.extend_range(key, None);
if self.is_collapsed {
return 0;
}
} else if key > self.max_key {
self.extend_range(key, None);
}
(key - self.offset) as usize
}
fn extend_range(&mut self, key: i32, second_key: Option<i32>) {
let second_key = second_key.unwrap_or(key);
let new_min_key = i32::min(key, i32::min(second_key, self.min_key));
let new_max_key = i32::max(key, i32::max(second_key, self.max_key));
if self.is_empty() {
let new_len = self.get_new_length(new_min_key, new_max_key);
self.bins.resize(new_len, 0);
self.offset = new_min_key;
self.adjust(new_min_key, new_max_key);
} else if new_min_key >= self.min_key && new_max_key < self.offset + self.length() {
self.min_key = new_min_key;
self.max_key = new_max_key;
} else {
// Grow bins
let new_length = self.get_new_length(new_min_key, new_max_key);
if new_length > self.length() as usize {
self.bins.resize(new_length, 0);
}
self.adjust(new_min_key, new_max_key);
}
}
fn get_new_length(&self, new_min_key: i32, new_max_key: i32) -> usize {
let desired_length = new_max_key - new_min_key + 1;
usize::min(
(CHUNK_SIZE * div_ceil(desired_length, CHUNK_SIZE)) as usize,
self.bin_limit,
)
}
fn adjust(&mut self, new_min_key: i32, new_max_key: i32) {
if new_max_key - new_min_key + 1 > self.length() {
let new_min_key = new_max_key - self.length() + 1;
if new_min_key >= self.max_key {
// Put everything in the first bin.
self.offset = new_min_key;
self.min_key = new_min_key;
self.bins.fill(0);
self.bins[0] = self.count;
} else {
let shift = self.offset - new_min_key;
if shift < 0 {
let collapse_start_index = (self.min_key - self.offset) as usize;
let collapse_end_index = (new_min_key - self.offset) as usize;
let collapsed_count: u64 = self.bins[collapse_start_index..collapse_end_index]
.iter()
.sum();
let zero_len = (new_min_key - self.min_key) as usize;
self.bins.splice(
collapse_start_index..collapse_end_index,
std::iter::repeat_n(0, zero_len),
);
self.bins[collapse_end_index] += collapsed_count;
}
self.min_key = new_min_key;
self.shift_bins(shift);
}
self.max_key = new_max_key;
self.is_collapsed = true;
} else {
self.center_bins(new_min_key, new_max_key);
self.min_key = new_min_key;
self.max_key = new_max_key;
}
}
fn shift_bins(&mut self, shift: i32) {
if shift > 0 {
let shift = shift as usize;
self.bins.rotate_right(shift);
for idx in 0..shift {
self.bins[idx] = 0;
}
} else {
let shift = shift.unsigned_abs() as usize;
for idx in 0..shift {
self.bins[idx] = 0;
}
self.bins.rotate_left(shift);
}
self.offset -= shift;
}
fn center_bins(&mut self, new_min_key: i32, new_max_key: i32) {
let middle_key = new_min_key + (new_max_key - new_min_key + 1) / 2;
let shift = self.offset + self.length() / 2 - middle_key;
self.shift_bins(shift)
}
pub fn key_at_rank(&self, rank: u64) -> i32 {
let mut n = 0;
for (i, bin) in self.bins.iter().enumerate() {
n += *bin;
if n > rank {
return i as i32 + self.offset;
}
}
self.max_key
}
pub fn count(&self) -> u64 {
self.count
}
pub fn merge(&mut self, other: &Store) {
if other.count == 0 {
return;
}
if self.count == 0 {
self.copy(other);
return;
}
if other.min_key < self.min_key || other.max_key > self.max_key {
self.extend_range(other.min_key, Some(other.max_key));
}
let collapse_start_index = other.min_key - other.offset;
let mut collapse_end_index = i32::min(self.min_key, other.max_key + 1) - other.offset;
if collapse_end_index > collapse_start_index {
let collapsed_count: u64 = self.bins
[collapse_start_index as usize..collapse_end_index as usize]
.iter()
.sum();
self.bins[0] += collapsed_count;
} else {
collapse_end_index = collapse_start_index;
}
for key in (collapse_end_index + other.offset)..(other.max_key + 1) {
self.bins[(key - self.offset) as usize] += other.bins[(key - other.offset) as usize]
}
self.count += other.count;
}
fn copy(&mut self, o: &Store) {
self.bins = o.bins.clone();
self.count = o.count;
self.min_key = o.min_key;
self.max_key = o.max_key;
self.offset = o.offset;
self.bin_limit = o.bin_limit;
self.is_collapsed = o.is_collapsed;
}
}
#[cfg(test)]
mod tests {
use crate::store::Store;
#[test]
fn test_simple_store() {
let mut s = Store::new(2048);
for i in 0..2048 {
s.add(i);
}
}
#[test]
fn test_simple_store_rev() {
let mut s = Store::new(2048);
for i in (0..2048).rev() {
s.add(i);
}
}
}

88
sketches-ddsketch/tests/common/dataset.rs Normal file
View File

@@ -0,0 +1,88 @@
use std::cmp::Ordering;
use std::f64::NAN;
pub struct Dataset {
values: Vec<f64>,
sum: f64,
sorted: bool,
}
fn cmp_f64(a: &f64, b: &f64) -> Ordering {
assert!(!a.is_nan() && !b.is_nan());
if a < b {
return Ordering::Less;
} else if a > b {
return Ordering::Greater;
} else {
return Ordering::Equal;
}
}
impl Dataset {
pub fn new() -> Self {
Dataset {
values: Vec::new(),
sum: 0.0,
sorted: false,
}
}
pub fn add(&mut self, value: f64) {
self.values.push(value);
self.sum += value;
self.sorted = false;
}
// pub fn quantile(&mut self, q: f64) -> f64 {
// self.lower_quantile(q)
// }
pub fn lower_quantile(&mut self, q: f64) -> f64 {
if q < 0.0 || q > 1.0 || self.values.len() == 0 {
return NAN;
}
self.sort();
let rank = q * (self.values.len() - 1) as f64;
self.values[rank.floor() as usize]
}
pub fn upper_quantile(&mut self, q: f64) -> f64 {
if q < 0.0 || q > 1.0 || self.values.len() == 0 {
return NAN;
}
self.sort();
let rank = q * (self.values.len() - 1) as f64;
self.values[rank.ceil() as usize]
}
pub fn min(&mut self) -> f64 {
self.sort();
self.values[0]
}
pub fn max(&mut self) -> f64 {
self.sort();
self.values[self.values.len() - 1]
}
pub fn sum(&self) -> f64 {
self.sum
}
pub fn count(&self) -> usize {
self.values.len()
}
fn sort(&mut self) {
if self.sorted {
return;
}
self.values.sort_by(cmp_f64);
self.sorted = true;
}
}

100
sketches-ddsketch/tests/common/generator.rs Normal file
View File

@@ -0,0 +1,100 @@
extern crate rand;
extern crate rand_distr;
use rand::prelude::*;
pub trait Generator {
fn generate(&mut self) -> f64;
}
// Constant generator
//
pub struct Constant {
value: f64,
}
impl Constant {
pub fn new(value: f64) -> Self {
Constant { value }
}
}
impl Generator for Constant {
fn generate(&mut self) -> f64 {
self.value
}
}
// Linear generator
//
pub struct Linear {
current_value: f64,
step: f64,
}
impl Linear {
pub fn new(start_value: f64, step: f64) -> Self {
Linear {
current_value: start_value,
step,
}
}
}
impl Generator for Linear {
fn generate(&mut self) -> f64 {
let value = self.current_value;
self.current_value += self.step;
value
}
}
// Normal distribution generator
//
pub struct Normal {
distr: rand_distr::Normal<f64>,
}
impl Normal {
pub fn new(mean: f64, stddev: f64) -> Self {
Normal {
distr: rand_distr::Normal::new(mean, stddev).unwrap(),
}
}
}
impl Generator for Normal {
fn generate(&mut self) -> f64 {
self.distr.sample(&mut rand::thread_rng())
}
}
// Lognormal distribution generator
//
pub struct Lognormal {
distr: rand_distr::LogNormal<f64>,
}
impl Lognormal {
pub fn new(mean: f64, stddev: f64) -> Self {
Lognormal {
distr: rand_distr::LogNormal::new(mean, stddev).unwrap(),
}
}
}
impl Generator for Lognormal {
fn generate(&mut self) -> f64 {
self.distr.sample(&mut rand::thread_rng())
}
}
// Exponential distribution generator
//
pub struct Exponential {
distr: rand_distr::Exp<f64>,
}
impl Exponential {
pub fn new(lambda: f64) -> Self {
Exponential {
distr: rand_distr::Exp::new(lambda).unwrap(),
}
}
}
impl Generator for Exponential {
fn generate(&mut self) -> f64 {
self.distr.sample(&mut rand::thread_rng())
}
}

2
sketches-ddsketch/tests/common/mod.rs Normal file
View File

@@ -0,0 +1,2 @@
pub mod dataset;
pub mod generator;

316
sketches-ddsketch/tests/test_ddsketch.rs Normal file
View File

@@ -0,0 +1,316 @@
mod common;
use std::time::Instant;
use common::dataset::Dataset;
use common::generator;
use common::generator::Generator;
use sketches_ddsketch::{Config, DDSketch};
const TEST_ALPHA: f64 = 0.01;
const TEST_MAX_BINS: u32 = 1024;
const TEST_MIN_VALUE: f64 = 1.0e-9;
// Used for float equality
const TEST_ERROR_THRESH: f64 = 1.0e-9;
const TEST_SIZES: [usize; 5] = [3, 5, 10, 100, 1000];
const TEST_QUANTILES: [f64; 10] = [0.0, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95, 0.99, 0.999, 1.0];
#[test]
fn test_constant() {
evaluate_sketches(|| Box::new(generator::Constant::new(42.0)));
}
#[test]
fn test_linear() {
evaluate_sketches(|| Box::new(generator::Linear::new(0.0, 1.0)));
}
#[test]
fn test_normal() {
evaluate_sketches(|| Box::new(generator::Normal::new(35.0, 1.0)));
}
#[test]
fn test_lognormal() {
evaluate_sketches(|| Box::new(generator::Lognormal::new(0.0, 2.0)));
}
#[test]
fn test_exponential() {
evaluate_sketches(|| Box::new(generator::Exponential::new(2.0)));
}
fn evaluate_test_sizes(f: impl Fn(usize)) {
for sz in &TEST_SIZES {
f(*sz);
}
}
fn evaluate_sketches(gen_factory: impl Fn() -> Box<dyn generator::Generator>) {
evaluate_test_sizes(|sz: usize| {
let mut generator = gen_factory();
evaluate_sketch(sz, &mut generator);
});
}
fn new_config() -> Config {
Config::new(TEST_ALPHA, TEST_MAX_BINS, TEST_MIN_VALUE)
}
fn assert_float_eq(a: f64, b: f64) {
assert!((a - b).abs() < TEST_ERROR_THRESH, "{} != {}", a, b);
}
fn evaluate_sketch(count: usize, generator: &mut Box<dyn generator::Generator>) {
let c = new_config();
let mut g = DDSketch::new(c);
let mut d = Dataset::new();
for _i in 0..count {
let value = generator.generate();
g.add(value);
d.add(value);
}
compare_sketches(&mut d, &g);
}
fn compare_sketches(d: &mut Dataset, g: &DDSketch) {
for q in &TEST_QUANTILES {
let lower = d.lower_quantile(*q);
let upper = d.upper_quantile(*q);
let min_expected;
if lower < 0.0 {
min_expected = lower * (1.0 + TEST_ALPHA);
} else {
min_expected = lower * (1.0 - TEST_ALPHA);
}
let max_expected;
if upper > 0.0 {
max_expected = upper * (1.0 + TEST_ALPHA);
} else {
max_expected = upper * (1.0 - TEST_ALPHA);
}
let quantile = g.quantile(*q).unwrap().unwrap();
assert!(
min_expected <= quantile,
"Lower than min, quantile: {}, wanted {} <= {}",
*q,
min_expected,
quantile
);
assert!(
quantile <= max_expected,
"Higher than max, quantile: {}, wanted {} <= {}",
*q,
quantile,
max_expected
);
// verify that calls do not modify result (not mut so not possible?)
let quantile2 = g.quantile(*q).unwrap().unwrap();
assert_eq!(quantile, quantile2);
}
assert_eq!(g.min().unwrap(), d.min());
assert_eq!(g.max().unwrap(), d.max());
assert_float_eq(g.sum().unwrap(), d.sum());
assert_eq!(g.count(), d.count());
}
#[test]
fn test_merge_normal() {
evaluate_test_sizes(|sz: usize| {
let c = new_config();
let mut d = Dataset::new();
let mut g1 = DDSketch::new(c);
let mut generator1 = generator::Normal::new(35.0, 1.0);
for _ in (0..sz).step_by(3) {
let value = generator1.generate();
g1.add(value);
d.add(value);
}
let mut g2 = DDSketch::new(c);
let mut generator2 = generator::Normal::new(50.0, 2.0);
for _ in (1..sz).step_by(3) {
let value = generator2.generate();
g2.add(value);
d.add(value);
}
g1.merge(&g2).unwrap();
let mut g3 = DDSketch::new(c);
let mut generator3 = generator::Normal::new(40.0, 0.5);
for _ in (2..sz).step_by(3) {
let value = generator3.generate();
g3.add(value);
d.add(value);
}
g1.merge(&g3).unwrap();
compare_sketches(&mut d, &g1);
});
}
#[test]
fn test_merge_empty() {
evaluate_test_sizes(|sz: usize| {
let c = new_config();
let mut d = Dataset::new();
let mut g1 = DDSketch::new(c);
let mut g2 = DDSketch::new(c);
let mut generator = generator::Exponential::new(5.0);
for _ in 0..sz {
let value = generator.generate();
g2.add(value);
d.add(value);
}
g1.merge(&g2).unwrap();
compare_sketches(&mut d, &g1);
let g3 = DDSketch::new(c);
g2.merge(&g3).unwrap();
compare_sketches(&mut d, &g2);
});
}
#[test]
fn test_merge_mixed() {
evaluate_test_sizes(|sz: usize| {
let c = new_config();
let mut d = Dataset::new();
let mut g1 = DDSketch::new(c);
let mut generator1 = generator::Normal::new(100.0, 1.0);
for _ in (0..sz).step_by(3) {
let value = generator1.generate();
g1.add(value);
d.add(value);
}
let mut g2 = DDSketch::new(c);
let mut generator2 = generator::Exponential::new(5.0);
for _ in (1..sz).step_by(3) {
let value = generator2.generate();
g2.add(value);
d.add(value);
}
g1.merge(&g2).unwrap();
let mut g3 = DDSketch::new(c);
let mut generator3 = generator::Exponential::new(0.1);
for _ in (2..sz).step_by(3) {
let value = generator3.generate();
g3.add(value);
d.add(value);
}
g1.merge(&g3).unwrap();
compare_sketches(&mut d, &g1);
})
}
#[test]
fn test_merge_incompatible() {
let c1 = Config::new(TEST_ALPHA, TEST_MAX_BINS, TEST_MIN_VALUE);
let c2 = Config::new(TEST_ALPHA * 2.0, TEST_MAX_BINS, TEST_MIN_VALUE);
let mut d1 = DDSketch::new(c1);
let d2 = DDSketch::new(c2);
assert!(d1.merge(&d2).is_err());
let c3 = Config::new(TEST_ALPHA, TEST_MAX_BINS, TEST_MIN_VALUE * 10.0);
let d3 = DDSketch::new(c3);
assert!(d1.merge(&d3).is_err());
let c4 = Config::new(TEST_ALPHA, TEST_MAX_BINS * 2, TEST_MIN_VALUE);
let d4 = DDSketch::new(c4);
assert!(d1.merge(&d4).is_err());
// the same should work
let c5 = Config::new(TEST_ALPHA, TEST_MAX_BINS, TEST_MIN_VALUE);
let dsame = DDSketch::new(c5);
assert!(d1.merge(&dsame).is_ok());
}
#[test]
#[ignore]
fn test_performance_insert() {
let c = Config::defaults();
let mut g = DDSketch::new(c);
let mut gen = generator::Normal::new(1000.0, 500.0);
let count = 300_000_000;
let mut values = Vec::new();
for _ in 0..count {
values.push(gen.generate());
}
let start_time = Instant::now();
for value in values {
g.add(value);
}
// This simply ensures the operations don't get optimzed out as ignored
let quantile = g.quantile(0.50).unwrap().unwrap();
let elapsed = start_time.elapsed().as_micros() as f64;
let elapsed = elapsed / 1_000_000.0;
println!(
"RESULT: p50={:.2} => Added {}M samples in {:2} secs ({:.2}M samples/sec)",
quantile,
count / 1_000_000,
elapsed,
(count as f64) / 1_000_000.0 / elapsed
);
}
#[test]
#[ignore]
fn test_performance_merge() {
let c = Config::defaults();
let mut gen = generator::Normal::new(1000.0, 500.0);
let merge_count = 500_000;
let sample_count = 1_000;
let mut sketches = Vec::new();
for _ in 0..merge_count {
let mut d = DDSketch::new(c);
for _ in 0..sample_count {
d.add(gen.generate());
}
sketches.push(d);
}
let mut base = DDSketch::new(c);
let start_time = Instant::now();
for sketch in &sketches {
base.merge(sketch).unwrap();
}
let elapsed = start_time.elapsed().as_micros() as f64;
let elapsed = elapsed / 1_000_000.0;
println!(
"RESULT: Merged {} sketches in {:2} secs ({:.2} merges/sec)",
merge_count,
elapsed,
(merge_count as f64) / elapsed
);
}

16
src/aggregation/accessor_helpers.rs
View File

@@ -95,11 +95,21 @@ pub(crate) fn get_all_ff_reader_or_empty(
allowed_column_types: Option<&[ColumnType]>,
fallback_type: ColumnType,
) -> crate::Result<Vec<(columnar::Column<u64>, ColumnType)>> {
let ff_fields = reader.fast_fields();
let mut ff_field_with_type =
ff_fields.u64_lenient_for_type_all(allowed_column_types, field_name)?;
let mut ff_field_with_type = get_all_ff_readers(reader, field_name, allowed_column_types)?;
if ff_field_with_type.is_empty() {
ff_field_with_type.push((Column::build_empty_column(reader.num_docs()), fallback_type));
}
Ok(ff_field_with_type)
}
/// Get all fast field reader.
pub(crate) fn get_all_ff_readers(
reader: &SegmentReader,
field_name: &str,
allowed_column_types: Option<&[ColumnType]>,
) -> crate::Result<Vec<(columnar::Column<u64>, ColumnType)>> {
let ff_fields = reader.fast_fields();
let ff_field_with_type =
ff_fields.u64_lenient_for_type_all(allowed_column_types, field_name)?;
Ok(ff_field_with_type)
}

108
src/aggregation/agg_data.rs
View File

@@ -9,12 +9,12 @@ use crate::aggregation::accessor_helpers::{
get_numeric_or_date_column_types,
};
use crate::aggregation::agg_req::{Aggregation, AggregationVariants, Aggregations};
pub use crate::aggregation::bucket::{CompositeAggReqData, CompositeSourceAccessors};
use crate::aggregation::bucket::{
build_segment_filter_collector, build_segment_range_collector, CompositeAggReqData,
CompositeAggregation, CompositeSourceAccessors, FilterAggReqData, HistogramAggReqData,
HistogramBounds, IncludeExcludeParam, MissingTermAggReqData, RangeAggReqData,
SegmentHistogramCollector, TermMissingAgg, TermsAggReqData, TermsAggregation,
TermsAggregationInternal,
build_segment_filter_collector, build_segment_range_collector, CompositeAggregation,
FilterAggReqData, HistogramAggReqData, HistogramBounds, IncludeExcludeParam,
MissingTermAggReqData, RangeAggReqData, SegmentCompositeCollector, SegmentHistogramCollector,
TermMissingAgg, TermsAggReqData, TermsAggregation, TermsAggregationInternal,
};
use crate::aggregation::metric::{
build_segment_stats_collector, AverageAggregation, CardinalityAggReqData,
@@ -143,8 +143,14 @@ impl AggregationsSegmentCtx {
.as_deref_mut()
.expect("histogram_req_data slot is empty (taken)")
}
#[inline]
pub(crate) fn get_composite_req_data_mut(&mut self, idx: usize) -> &mut CompositeAggReqData {
self.per_request.composite_req_data[idx]
.as_deref_mut()
.expect("composite_req_data slot is empty (taken)")
}
// ---------- take / put (terms, histogram, range) ----------
// ---------- take / put (terms, histogram, range, composite) ----------
/// Move out the boxed Histogram request at `idx`, leaving `None`.
#[inline]
@@ -232,6 +238,8 @@ pub struct PerRequestAggSegCtx {
pub range_req_data: Vec<Option<Box<RangeAggReqData>>>,
/// FilterAggReqData contains the request data for a filter aggregation.
pub filter_req_data: Vec<Option<Box<FilterAggReqData>>>,
/// CompositeAggReqData contains the request data for a composite aggregation.
pub composite_req_data: Vec<Option<Box<CompositeAggReqData>>>,
/// Shared by avg, min, max, sum, stats, extended_stats, count
pub stats_metric_req_data: Vec<MetricAggReqData>,
/// CardinalityAggReqData contains the request data for a cardinality aggregation.
@@ -240,8 +248,6 @@ pub struct PerRequestAggSegCtx {
pub top_hits_req_data: Vec<TopHitsAggReqData>,
/// MissingTermAggReqData contains the request data for a missing term aggregation.
pub missing_term_req_data: Vec<MissingTermAggReqData>,
/// CompositeAggReqData contains the request data for a composite aggregation.
pub composite_req_data: Vec<Option<Box<CompositeAggReqData>>>,
/// Request tree used to build collectors.
pub agg_tree: Vec<AggRefNode>,
@@ -292,7 +298,7 @@ impl PerRequestAggSegCtx {
+ self
.composite_req_data
.iter()
.map(|b| b.as_ref().map(|d| d.get_memory_consumption()).unwrap_or(0))
.map(|t| t.as_ref().unwrap().get_memory_consumption())
.sum::<usize>()
+ self.agg_tree.len() * std::mem::size_of::<AggRefNode>()
}
@@ -330,7 +336,7 @@ impl PerRequestAggSegCtx {
.expect("filter_req_data slot is empty (taken)")
.name
.as_str(),
AggKind::Composite => self.composite_req_data[idx]
AggKind::Composite => &self.composite_req_data[idx]
.as_deref()
.expect("composite_req_data slot is empty (taken)")
.name
@@ -461,11 +467,9 @@ pub(crate) fn build_segment_agg_collector(
)?)),
AggKind::Range => Ok(build_segment_range_collector(req, node)?),
AggKind::Filter => build_segment_filter_collector(req, node),
AggKind::Composite => Ok(Box::new(
crate::aggregation::bucket::SegmentCompositeCollector::from_req_and_validate(
req, node,
)?,
)),
AggKind::Composite => Ok(Box::new(SegmentCompositeCollector::from_req_and_validate(
req, node,
)?)),
}
}
@@ -760,14 +764,6 @@ fn build_nodes(
children,
}])
}
AggregationVariants::Composite(composite_req) => Ok(vec![build_composite_node(
agg_name,
reader,
segment_ordinal,
data,
&req.sub_aggregation,
composite_req,
)?]),
AggregationVariants::Filter(filter_req) => {
// Build the query and evaluator upfront
let schema = reader.schema();
@@ -799,38 +795,17 @@ fn build_nodes(
children,
}])
}
AggregationVariants::Composite(composite_req) => Ok(vec![build_composite_node(
agg_name,
reader,
segment_ordinal,
data,
&req.sub_aggregation,
composite_req,
)?]),
}
}
fn build_composite_node(
agg_name: &str,
reader: &SegmentReader,
_segment_ordinal: SegmentOrdinal,
data: &mut AggregationsSegmentCtx,
sub_aggs: &Aggregations,
req: &CompositeAggregation,
) -> crate::Result<AggRefNode> {
let mut composite_accessors = Vec::with_capacity(req.sources.len());
for source in &req.sources {
let source_after_key_opt = req.after.get(source.name()).map(|k| &k.0);
let source_accessor =
CompositeSourceAccessors::build_for_source(reader, source, source_after_key_opt)?;
composite_accessors.push(source_accessor);
}
let agg = CompositeAggReqData {
name: agg_name.to_string(),
req: req.clone(),
composite_accessors,
};
let idx = data.push_composite_req_data(agg);
let children = build_children(sub_aggs, reader, _segment_ordinal, data)?;
Ok(AggRefNode {
kind: AggKind::Composite,
idx_in_req_data: idx,
children,
})
}
fn build_children(
aggs: &Aggregations,
reader: &SegmentReader,
@@ -1023,6 +998,35 @@ fn build_terms_or_cardinality_nodes(
Ok(nodes)
}
fn build_composite_node(
agg_name: &str,
reader: &SegmentReader,
segment_ordinal: SegmentOrdinal,
data: &mut AggregationsSegmentCtx,
sub_aggs: &Aggregations,
req: &CompositeAggregation,
) -> crate::Result<AggRefNode> {
let mut composite_accessors = Vec::with_capacity(req.sources.len());
for source in &req.sources {
let source_after_key_opt = req.after.get(source.name()).map(|k| &k.0);
let source_accessor =
CompositeSourceAccessors::build_for_source(reader, source, source_after_key_opt)?;
composite_accessors.push(source_accessor);
}
let agg = CompositeAggReqData {
name: agg_name.to_string(),
req: req.clone(),
composite_accessors,
};
let idx = data.push_composite_req_data(agg);
let children = build_children(sub_aggs, reader, segment_ordinal, data)?;
Ok(AggRefNode {
kind: AggKind::Composite,
idx_in_req_data: idx,
children,
})
}
/// Builds a single BitSet of allowed term ordinals for a string dictionary column according to
/// include/exclude parameters.
fn build_allowed_term_ids_for_str(

9
src/aggregation/agg_req.rs
View File

@@ -32,14 +32,15 @@ use rustc_hash::FxHashMap;
use serde::{Deserialize, Serialize};
use super::bucket::{
CompositeAggregation, DateHistogramAggregationReq, FilterAggregation, HistogramAggregation,
RangeAggregation, TermsAggregation,
DateHistogramAggregationReq, FilterAggregation, HistogramAggregation, RangeAggregation,
TermsAggregation,
};
use super::metric::{
AverageAggregation, CardinalityAggregationReq, CountAggregation, ExtendedStatsAggregation,
MaxAggregation, MinAggregation, PercentilesAggregationReq, StatsAggregation, SumAggregation,
TopHitsAggregationReq,
};
use crate::aggregation::bucket::CompositeAggregation;
/// The top-level aggregation request structure, which contains [`Aggregation`] and their user
/// defined names. It is also used in buckets aggregations to define sub-aggregations.
@@ -134,7 +135,7 @@ pub enum AggregationVariants {
/// Filter documents into a single bucket.
#[serde(rename = "filter")]
Filter(FilterAggregation),
/// Multi-dimensional, paginable bucket aggregation.
/// Put data into multi level paginated buckets.
#[serde(rename = "composite")]
Composite(CompositeAggregation),
@@ -186,7 +187,7 @@ impl AggregationVariants {
AggregationVariants::Composite(composite) => composite
.sources
.iter()
.map(|source| source.field())
.map(|source_map| source_map.field())
.collect(),
AggregationVariants::Average(avg) => vec![avg.field_name()],
AggregationVariants::Count(count) => vec![count.field_name()],

55
src/aggregation/agg_result.rs
View File

@@ -9,12 +9,12 @@ use rustc_hash::FxHashMap;
use serde::{Deserialize, Serialize};
use super::bucket::GetDocCount;
use super::intermediate_agg_result::CompositeIntermediateKey;
use super::metric::{
ExtendedStats, PercentilesMetricResult, SingleMetricResult, Stats, TopHitsMetricResult,
};
use super::{AggregationError, Key};
use crate::aggregation::bucket::AfterKey;
use crate::aggregation::intermediate_agg_result::CompositeIntermediateKey;
use crate::TantivyError;
#[derive(Clone, Default, Debug, PartialEq, Serialize, Deserialize)]
@@ -160,9 +160,11 @@ pub enum BucketResult {
},
/// This is the filter result - a single bucket with sub-aggregations
Filter(FilterBucketResult),
/// This is the composite result
/// This is the composite aggregation result
Composite {
/// The buckets
///
/// See [`CompositeAggregation`](super::bucket::CompositeAggregation)
buckets: Vec<CompositeBucketEntry>,
/// The key to start after when paginating
#[serde(skip_serializing_if = "FxHashMap::is_empty")]
@@ -351,6 +353,10 @@ pub struct FilterBucketResult {
pub sub_aggregations: AggregationResults,
}
/// The JSON mappable key to identify a composite bucket.
///
/// This is similar to `Key`, but composite keys can also be boolean and null.
///
/// Note the type information loss compared to `CompositeIntermediateKey`.
/// Pagination is performed using `AfterKey`, which encodes type information.
#[derive(Clone, Debug, Serialize, Deserialize)]
@@ -392,7 +398,15 @@ impl PartialEq for CompositeKey {
(Self::I64(l), Self::I64(r)) => l == r,
(Self::U64(l), Self::U64(r)) => l == r,
(Self::Null, Self::Null) => true,
_ => false,
(
Self::Bool(_)
| Self::Str(_)
| Self::F64(_)
| Self::I64(_)
| Self::U64(_)
| Self::Null,
_,
) => false,
}
}
}
@@ -401,6 +415,7 @@ impl From<CompositeIntermediateKey> for CompositeKey {
match value {
CompositeIntermediateKey::Str(s) => Self::Str(s),
CompositeIntermediateKey::IpAddr(s) => {
// Prefer to use the IPv4 representation if possible
if let Some(ip) = s.to_ipv4_mapped() {
Self::Str(ip.to_string())
} else {
@@ -411,13 +426,43 @@ impl From<CompositeIntermediateKey> for CompositeKey {
CompositeIntermediateKey::Bool(f) => Self::Bool(f),
CompositeIntermediateKey::U64(f) => Self::U64(f),
CompositeIntermediateKey::I64(f) => Self::I64(f),
CompositeIntermediateKey::DateTime(f) => Self::I64(f / 1_000_000), // ns to ms
CompositeIntermediateKey::DateTime(f) => Self::I64(f / 1_000_000), // Convert ns to ms
CompositeIntermediateKey::Null => Self::Null,
}
}
}
/// Composite bucket entry with a multi-dimensional key.
/// This is the default entry for a bucket, which contains a composite key, count, and optionally
/// sub-aggregations.
/// ...
/// "my_composite": {
/// "buckets": [
/// {
/// "key": {
/// "date": 1494201600000,
/// "product": "rocky"
/// },
/// "doc_count": 5
/// },
/// {
/// "key": {
/// "date": 1494201600000,
/// "product": "balboa"
/// },
/// "doc_count": 2
/// },
/// {
/// "key": {
/// "date": 1494201700000,
/// "product": "john"
/// },
/// "doc_count": 3
/// }
/// ]
/// }
/// ...
/// }
/// ```
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct CompositeBucketEntry {
/// The identifier of the bucket.

51
src/aggregation/bucket/composite/accessors.rs
View File

@@ -1,9 +1,10 @@
use std::fmt::Debug;
use std::net::Ipv6Addr;
use columnar::column_values::{CompactHit, CompactSpaceU64Accessor};
use columnar::{Column, ColumnType, MonotonicallyMappableToU64, StrColumn, TermOrdHit};
use crate::aggregation::accessor_helpers::get_numeric_or_date_column_types;
use crate::aggregation::accessor_helpers::{get_all_ff_readers, get_numeric_or_date_column_types};
use crate::aggregation::bucket::composite::numeric_types::num_proj;
use crate::aggregation::bucket::composite::numeric_types::num_proj::ProjectedNumber;
use crate::aggregation::bucket::composite::ToTypePaginationOrder;
@@ -115,14 +116,11 @@ impl CompositeSourceAccessors {
ColumnType::IpAddr,
// ColumnType::Bytes Unsupported
];
let mut columns_and_types = reader
.fast_fields()
.u64_lenient_for_type_all(Some(&allowed_column_types), &source.field)?;
let mut columns_and_types =
get_all_ff_readers(reader, &source.field, Some(&allowed_column_types))?;
// Sort columns by their pagination order and determine which to skip
columns_and_types.sort_by_key(|(_, col_type): &(Column, ColumnType)| {
col_type.column_pagination_order()
});
columns_and_types.sort_by_key(|(_, col_type)| col_type.column_pagination_order());
if source.order == Order::Desc {
columns_and_types.reverse();
}
@@ -150,7 +148,7 @@ impl CompositeSourceAccessors {
{
match source_after_key_opt {
Some(after_key) => PrecomputedAfterKey::precompute(
first_col,
&first_col,
after_key,
&source.field,
source.missing_order,
@@ -174,11 +172,11 @@ impl CompositeSourceAccessors {
})
}
CompositeAggregationSource::Histogram(source) => {
let column_and_types: Vec<(Column, ColumnType)> =
reader.fast_fields().u64_lenient_for_type_all(
Some(get_numeric_or_date_column_types()),
&source.field,
)?;
let column_and_types: Vec<(Column, ColumnType)> = get_all_ff_readers(
reader,
&source.field,
Some(get_numeric_or_date_column_types()),
)?;
let source_collectors: Vec<CompositeAccessor> = column_and_types
.into_iter()
.map(|(column, column_type)| {
@@ -214,9 +212,8 @@ impl CompositeSourceAccessors {
})
}
CompositeAggregationSource::DateHistogram(source) => {
let column_and_types = reader
.fast_fields()
.u64_lenient_for_type_all(Some(&[ColumnType::DateTime]), &source.field)?;
let column_and_types =
get_all_ff_readers(reader, &source.field, Some(&[ColumnType::DateTime]))?;
let date_histogram_interval =
PrecomputedDateInterval::from_date_histogram_source_intervals(
&source.fixed_interval,
@@ -342,7 +339,7 @@ impl PrecomputedDateInterval {
.to_string(),
)),
(Some(fixed_interval), None) => {
let fixed_interval_ms = parse_into_milliseconds(fixed_interval)?;
let fixed_interval_ms = parse_into_milliseconds(&fixed_interval)?;
Ok(PrecomputedDateInterval::FixedNanoseconds(
fixed_interval_ms * 1_000_000,
))
@@ -370,16 +367,6 @@ pub enum PrecomputedAfterKey {
AfterLast,
}
impl From<CompactHit> for PrecomputedAfterKey {
fn from(hit: CompactHit) -> Self {
match hit {
CompactHit::Exact(ord) => PrecomputedAfterKey::Exact(ord as u64),
CompactHit::Next(ord) => PrecomputedAfterKey::Next(ord as u64),
CompactHit::AfterLast => PrecomputedAfterKey::AfterLast,
}
}
}
impl From<TermOrdHit> for PrecomputedAfterKey {
fn from(hit: TermOrdHit) -> Self {
match hit {
@@ -390,6 +377,16 @@ impl From<TermOrdHit> for PrecomputedAfterKey {
}
}
impl From<CompactHit> for PrecomputedAfterKey {
fn from(hit: CompactHit) -> Self {
match hit {
CompactHit::Exact(ord) => PrecomputedAfterKey::Exact(ord as u64),
CompactHit::Next(ord) => PrecomputedAfterKey::Next(ord as u64),
CompactHit::AfterLast => PrecomputedAfterKey::AfterLast,
}
}
}
impl<T: MonotonicallyMappableToU64> From<ProjectedNumber<T>> for PrecomputedAfterKey {
fn from(num: ProjectedNumber<T>) -> Self {
match num {

10
src/aggregation/bucket/composite/calendar_interval.rs
View File

@@ -8,8 +8,9 @@ const NS_IN_DAY: i64 = Nanosecond::per_t::<i128>(Day) as i64;
pub(super) fn try_year_bucket(timestamp_ns: i64) -> crate::Result<i64> {
year_bucket_using_time_crate(timestamp_ns).map_err(|e| {
crate::TantivyError::InvalidArgument(format!(
"Failed to compute year bucket for timestamp {}: {e}",
timestamp_ns
"Failed to compute year bucket for timestamp {}: {}",
timestamp_ns,
e.to_string()
))
})
}
@@ -19,8 +20,9 @@ pub(super) fn try_year_bucket(timestamp_ns: i64) -> crate::Result<i64> {
pub(super) fn try_month_bucket(timestamp_ns: i64) -> crate::Result<i64> {
month_bucket_using_time_crate(timestamp_ns).map_err(|e| {
crate::TantivyError::InvalidArgument(format!(
"Failed to compute month bucket for timestamp {}: {e}",
timestamp_ns
"Failed to compute month bucket for timestamp {}: {}",
timestamp_ns,
e.to_string()
))
})
}

631
src/aggregation/bucket/composite/collector.rs
View File

@@ -1,5 +1,4 @@
use std::fmt::Debug;
use std::mem;
use std::net::Ipv6Addr;
use columnar::column_values::CompactSpaceU64Accessor;
@@ -21,94 +20,75 @@ use crate::aggregation::bucket::composite::map::{DynArrayHeapMap, MAX_DYN_ARRAY_
use crate::aggregation::bucket::{
CalendarInterval, CompositeAggregationSource, MissingOrder, Order,
};
use crate::aggregation::cached_sub_aggs::{CachedSubAggs, HighCardSubAggCache};
use crate::aggregation::intermediate_agg_result::{
CompositeIntermediateKey, IntermediateAggregationResult, IntermediateAggregationResults,
IntermediateBucketResult, IntermediateCompositeBucketEntry, IntermediateCompositeBucketResult,
};
use crate::aggregation::segment_agg_result::{BucketIdProvider, SegmentAggregationCollector};
use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
use crate::aggregation::BucketId;
use crate::TantivyError;
#[derive(Clone, Debug)]
#[derive(Debug)]
struct CompositeBucketCollector {
count: u32,
bucket_id: BucketId,
}
/// Compact sortable representation of a single source value within a composite key.
///
/// The struct encodes both the column identity and the fast field value in a way
/// that preserves the desired sort order via the derived `Ord` implementation
/// (fields are compared top-to-bottom: `sort_key` first, then `encoded_value`).
///
/// ## `sort_key` encoding
/// - `0` — missing value, sorted first
/// - `1..=254` — present value; the original accessor index is `sort_key - 1`
/// - `u8::MAX` (255) — missing value, sorted last
///
/// ## `encoded_value` encoding
/// - `0` when the field is missing
/// - The raw u64 fast-field representation when order is ascending
/// - Bitwise NOT of the raw u64 when order is descending
#[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord, Default, Hash)]
struct InternalValueRepr {
/// Column index biased by +1 (so 0 and u8::MAX are reserved for missing sentinels).
sort_key: u8,
/// Fast field value, possibly bit-flipped for descending order.
encoded_value: u64,
impl CompositeBucketCollector {
fn new() -> Self {
CompositeBucketCollector { count: 0 }
}
#[inline]
fn collect(&mut self) {
self.count += 1;
}
}
/// The value is represented as a tuple of:
/// - the column index or missing value sentinel
/// - if the value is present, store the accessor index + 1
/// - if the value is missing, store 0 (for missing first) or u8::MAX (for missing last)
/// - the fast field value u64 representation
/// - 0 if the field is missing
/// - regular u64 repr if the ordering is ascending
/// - bitwise NOT of the u64 repr if the ordering is descending
#[derive(Clone, Copy, Debug, PartialEq, Eq, PartialOrd, Ord, Default, Hash)]
struct InternalValueRepr(u8, u64);
impl InternalValueRepr {
#[inline]
fn new_term(raw: u64, accessor_idx: u8, order: Order) -> Self {
let encoded_value = match order {
Order::Asc => raw,
Order::Desc => !raw,
};
InternalValueRepr {
sort_key: accessor_idx + 1,
encoded_value,
match order {
Order::Asc => InternalValueRepr(accessor_idx + 1, raw),
Order::Desc => InternalValueRepr(accessor_idx + 1, !raw),
}
}
/// For histogram sources the column index is irrelevant (always 1).
/// For histogram, the source column does not matter
#[inline]
fn new_histogram(raw: u64, order: Order) -> Self {
let encoded_value = match order {
Order::Asc => raw,
Order::Desc => !raw,
};
InternalValueRepr {
sort_key: 1,
encoded_value,
match order {
Order::Asc => InternalValueRepr(1, raw),
Order::Desc => InternalValueRepr(1, !raw),
}
}
#[inline]
fn new_missing(order: Order, missing_order: MissingOrder) -> Self {
let sort_key = match (missing_order, order) {
(MissingOrder::First, _) | (MissingOrder::Default, Order::Asc) => 0,
(MissingOrder::Last, _) | (MissingOrder::Default, Order::Desc) => u8::MAX,
let column_idx = match (missing_order, order) {
(MissingOrder::First, _) => 0,
(MissingOrder::Last, _) => u8::MAX,
(MissingOrder::Default, Order::Asc) => 0,
(MissingOrder::Default, Order::Desc) => u8::MAX,
};
InternalValueRepr {
sort_key,
encoded_value: 0,
}
InternalValueRepr(column_idx, 0)
}
/// Decode back to `(accessor_idx, raw_value)`.
/// Returns `None` when the value represents a missing field.
#[inline]
fn decode(self, order: Order) -> Option<(u8, u64)> {
if self.sort_key == 0 || self.sort_key == u8::MAX {
if self.0 == u8::MAX || self.0 == 0 {
return None;
}
let raw = match order {
Order::Asc => self.encoded_value,
Order::Desc => !self.encoded_value,
};
Some((self.sort_key - 1, raw))
match order {
Order::Asc => Some((self.0 - 1, self.1)),
Order::Desc => Some((self.0 - 1, !self.1)),
}
}
}
@@ -116,13 +96,8 @@ impl InternalValueRepr {
/// does a conversion to the correct datatype.
#[derive(Debug)]
pub struct SegmentCompositeCollector {
/// One DynArrayHeapMap per parent bucket.
parent_buckets: Vec<DynArrayHeapMap<InternalValueRepr, CompositeBucketCollector>>,
buckets: DynArrayHeapMap<InternalValueRepr, CompositeBucketCollector>,
accessor_idx: usize,
sub_agg: Option<CachedSubAggs<HighCardSubAggCache>>,
bucket_id_provider: BucketIdProvider,
/// Number of sources, needed when creating new DynArrayHeapMaps.
num_sources: usize,
}
impl SegmentAggregationCollector for SegmentCompositeCollector {
@@ -130,14 +105,14 @@ impl SegmentAggregationCollector for SegmentCompositeCollector {
&mut self,
agg_data: &AggregationsSegmentCtx,
results: &mut IntermediateAggregationResults,
parent_bucket_id: BucketId,
_parent_bucket_id: BucketId,
) -> crate::Result<()> {
let name = agg_data
.get_composite_req_data(self.accessor_idx)
.name
.clone();
let buckets = self.add_intermediate_bucket_result(agg_data, parent_bucket_id)?;
let buckets = self.into_intermediate_bucket_result(agg_data)?;
results.push(
name,
IntermediateAggregationResult::Bucket(IntermediateBucketResult::Composite { buckets }),
@@ -146,9 +121,10 @@ impl SegmentAggregationCollector for SegmentCompositeCollector {
Ok(())
}
#[inline]
fn collect(
&mut self,
parent_bucket_id: BucketId,
_parent_bucket_id: BucketId,
docs: &[crate::DocId],
agg_data: &mut AggregationsSegmentCtx,
) -> crate::Result<()> {
@@ -156,22 +132,19 @@ impl SegmentAggregationCollector for SegmentCompositeCollector {
let composite_agg_data = agg_data.take_composite_req_data(self.accessor_idx);
for doc in docs {
let mut visitor = CompositeKeyVisitor {
doc_id: *doc,
composite_agg_data: &composite_agg_data,
buckets: &mut self.parent_buckets[parent_bucket_id as usize],
sub_agg: &mut self.sub_agg,
bucket_id_provider: &mut self.bucket_id_provider,
sub_level_values: SmallVec::new(),
};
visitor.visit(0, true)?;
let mut sub_level_values = SmallVec::new();
recursive_key_visitor(
*doc,
agg_data,
&composite_agg_data,
0,
&mut sub_level_values,
&mut self.buckets,
true,
)?;
}
agg_data.put_back_composite_req_data(self.accessor_idx, composite_agg_data);
if let Some(sub_agg) = &mut self.sub_agg {
sub_agg.check_flush_local(agg_data)?;
}
let mem_delta = self.get_memory_consumption() - mem_pre;
if mem_delta > 0 {
agg_data.context.limits.add_memory_consumed(mem_delta)?;
@@ -180,33 +153,22 @@ impl SegmentAggregationCollector for SegmentCompositeCollector {
Ok(())
}
fn flush(&mut self, agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
if let Some(sub_agg) = &mut self.sub_agg {
sub_agg.flush(agg_data)?;
}
fn prepare_max_bucket(
&mut self,
_max_bucket: BucketId,
_agg_data: &AggregationsSegmentCtx,
) -> crate::Result<()> {
Ok(())
}
fn prepare_max_bucket(
&mut self,
max_bucket: BucketId,
_agg_data: &AggregationsSegmentCtx,
) -> crate::Result<()> {
let required_len = max_bucket as usize + 1;
while self.parent_buckets.len() < required_len {
let map = DynArrayHeapMap::try_new(self.num_sources)?;
self.parent_buckets.push(map);
}
fn flush(&mut self, _agg_data: &mut AggregationsSegmentCtx) -> crate::Result<()> {
Ok(())
}
}
impl SegmentCompositeCollector {
fn get_memory_consumption(&self) -> u64 {
self.parent_buckets
.iter()
.map(|m| m.memory_consumption())
.sum()
self.buckets.memory_consumption()
}
pub(crate) fn from_req_and_validate(
@@ -215,54 +177,34 @@ impl SegmentCompositeCollector {
) -> crate::Result<Self> {
validate_req(req_data, node.idx_in_req_data)?;
let has_sub_aggregations = !node.children.is_empty();
let sub_agg = if has_sub_aggregations {
let sub_agg_collector = build_segment_agg_collectors(req_data, &node.children)?;
Some(CachedSubAggs::new(sub_agg_collector))
} else {
None
};
if !node.children.is_empty() {
let _sub_aggregation = build_segment_agg_collectors(req_data, &node.children)?;
}
let composite_req_data = req_data.get_composite_req_data(node.idx_in_req_data);
let num_sources = composite_req_data.req.sources.len();
Ok(SegmentCompositeCollector {
parent_buckets: vec![DynArrayHeapMap::try_new(num_sources)?],
buckets: DynArrayHeapMap::try_new(composite_req_data.req.sources.len())?,
accessor_idx: node.idx_in_req_data,
sub_agg,
bucket_id_provider: BucketIdProvider::default(),
num_sources,
})
}
#[inline]
fn add_intermediate_bucket_result(
pub(crate) fn into_intermediate_bucket_result(
&mut self,
agg_data: &AggregationsSegmentCtx,
parent_bucket_id: BucketId,
) -> crate::Result<IntermediateCompositeBucketResult> {
let empty_map = DynArrayHeapMap::try_new(self.num_sources)?;
let heap_map = mem::replace(
&mut self.parent_buckets[parent_bucket_id as usize],
empty_map,
);
let mut dict: FxHashMap<Vec<CompositeIntermediateKey>, IntermediateCompositeBucketEntry> =
Default::default();
dict.reserve(heap_map.size());
dict.reserve(self.buckets.size());
let composite_data = agg_data.get_composite_req_data(self.accessor_idx);
for (key_internal_repr, agg) in heap_map.into_iter() {
let buckets = std::mem::replace(
&mut self.buckets,
DynArrayHeapMap::try_new(composite_data.req.sources.len())
.expect("already validated source count"),
);
for (key_internal_repr, agg) in buckets.into_iter() {
let key = resolve_key(&key_internal_repr, composite_data)?;
let mut sub_aggregation_res = IntermediateAggregationResults::default();
if let Some(sub_agg) = &mut self.sub_agg {
sub_agg
.get_sub_agg_collector()
.add_intermediate_aggregation_result(
agg_data,
&mut sub_aggregation_res,
agg.bucket_id,
)?;
}
let sub_aggregation_res = IntermediateAggregationResults::default();
dict.insert(
key,
@@ -303,13 +245,6 @@ fn validate_req(req_data: &mut AggregationsSegmentCtx, accessor_idx: usize) -> c
"composite aggregation 'size' must be > 0".to_string(),
));
}
if composite_data.composite_accessors.len() > MAX_DYN_ARRAY_SIZE {
return Err(TantivyError::InvalidArgument(format!(
"composite aggregation source supports maximum {MAX_DYN_ARRAY_SIZE} sources",
)));
}
let column_types_for_sources = composite_data.composite_accessors.iter().map(|item| {
item.accessors
.iter()
@@ -318,6 +253,11 @@ fn validate_req(req_data: &mut AggregationsSegmentCtx, accessor_idx: usize) -> c
});
for column_types in column_types_for_sources {
if column_types.len() > MAX_DYN_ARRAY_SIZE {
return Err(TantivyError::InvalidArgument(format!(
"composite aggregation source supports maximum {MAX_DYN_ARRAY_SIZE} sources",
)));
}
if column_types.contains(&ColumnType::Bytes) {
return Err(TantivyError::InvalidArgument(
"composite aggregation does not support 'bytes' field type".to_string(),
@@ -328,47 +268,34 @@ fn validate_req(req_data: &mut AggregationsSegmentCtx, accessor_idx: usize) -> c
}
fn collect_bucket_with_limit(
doc_id: crate::DocId,
limit_num_buckets: usize,
agg_data: &mut AggregationsSegmentCtx,
composite_agg_data: &CompositeAggReqData,
buckets: &mut DynArrayHeapMap<InternalValueRepr, CompositeBucketCollector>,
key: &[InternalValueRepr],
sub_agg: &mut Option<CachedSubAggs<HighCardSubAggCache>>,
bucket_id_provider: &mut BucketIdProvider,
) {
let mut record_in_bucket = |bucket: &mut CompositeBucketCollector| {
bucket.count += 1;
if let Some(sub_agg) = sub_agg {
sub_agg.push(bucket.bucket_id, doc_id);
}
};
// We still have room for buckets, just insert
if buckets.size() < limit_num_buckets {
let bucket = buckets.get_or_insert_with(key, || CompositeBucketCollector {
count: 0,
bucket_id: bucket_id_provider.next_bucket_id(),
});
record_in_bucket(bucket);
return;
) -> crate::Result<()> {
if (buckets.size() as u32) < composite_agg_data.req.size {
buckets
.get_or_insert_with(key, CompositeBucketCollector::new)
.collect();
return Ok(());
}
// Map is full, but we can still update the bucket if it already exists
if let Some(bucket) = buckets.get_mut(key) {
record_in_bucket(bucket);
return;
if let Some(entry) = buckets.get_mut(key) {
entry.collect();
return Ok(());
}
// Check if the item qualifies to enter the top-k, and evict the highest if it does
if let Some(highest_key) = buckets.peek_highest() {
if key < highest_key {
buckets.evict_highest();
let bucket = buckets.get_or_insert_with(key, || CompositeBucketCollector {
count: 0,
bucket_id: bucket_id_provider.next_bucket_id(),
});
record_in_bucket(bucket);
buckets
.get_or_insert_with(key, CompositeBucketCollector::new)
.collect();
}
}
let _ = agg_data;
Ok(())
}
/// Converts the composite key from its internal column space representation
@@ -378,7 +305,7 @@ fn resolve_key(
agg_data: &CompositeAggReqData,
) -> crate::Result<Vec<CompositeIntermediateKey>> {
internal_key
.iter()
.into_iter()
.enumerate()
.map(|(idx, val)| {
resolve_internal_value_repr(
@@ -463,190 +390,206 @@ fn resolve_term(
let val: u128 = compact_space_accessor.compact_to_u128(val as u32);
let val = Ipv6Addr::from_u128(val);
CompositeIntermediateKey::IpAddr(val)
} else if *column_type == ColumnType::U64 {
CompositeIntermediateKey::U64(val)
} else if *column_type == ColumnType::I64 {
CompositeIntermediateKey::I64(i64::from_u64(val))
} else {
let val = f64::from_u64(val);
let val: NumericalValue = val.into();
if *column_type == ColumnType::U64 {
CompositeIntermediateKey::U64(val)
} else if *column_type == ColumnType::I64 {
CompositeIntermediateKey::I64(i64::from_u64(val))
} else {
let val = f64::from_u64(val);
let val: NumericalValue = val.into();
match val.normalize() {
NumericalValue::U64(val) => CompositeIntermediateKey::U64(val),
NumericalValue::I64(val) => CompositeIntermediateKey::I64(val),
NumericalValue::F64(val) => CompositeIntermediateKey::F64(val),
match val.normalize() {
NumericalValue::U64(val) => CompositeIntermediateKey::U64(val),
NumericalValue::I64(val) => CompositeIntermediateKey::I64(val),
NumericalValue::F64(val) => CompositeIntermediateKey::F64(val),
}
}
};
Ok(key)
}
/// Browse through the cardinal product obtained by the different values of the doc composite key
/// sources.
///
/// For each of those tuple-key, that are after the limit key, we call collect_bucket_with_limit.
struct CompositeKeyVisitor<'a> {
/// Depth-first walk of the accessors to build the composite key combinations
/// and update the buckets.
fn recursive_key_visitor(
doc_id: crate::DocId,
composite_agg_data: &'a CompositeAggReqData,
buckets: &'a mut DynArrayHeapMap<InternalValueRepr, CompositeBucketCollector>,
sub_agg: &'a mut Option<CachedSubAggs<HighCardSubAggCache>>,
bucket_id_provider: &'a mut BucketIdProvider,
sub_level_values: SmallVec<[InternalValueRepr; MAX_DYN_ARRAY_SIZE]>,
}
agg_data: &mut AggregationsSegmentCtx,
composite_agg_data: &CompositeAggReqData,
source_idx_for_recursion: usize,
sub_level_values: &mut SmallVec<[InternalValueRepr; MAX_DYN_ARRAY_SIZE]>,
buckets: &mut DynArrayHeapMap<InternalValueRepr, CompositeBucketCollector>,
is_on_after_key: bool,
) -> crate::Result<()> {
if source_idx_for_recursion == composite_agg_data.req.sources.len() {
if !is_on_after_key {
collect_bucket_with_limit(
agg_data,
composite_agg_data,
buckets,
sub_level_values,
)?;
}
return Ok(());
}
impl CompositeKeyVisitor<'_> {
/// Depth-first walk of the accessors to build the composite key combinations
/// and update the buckets.
///
/// `source_idx` is the current source index in the recursion.
/// `is_on_after_key` tracks whether we still need to consider the after_key
/// for pruning at this level and below.
fn visit(&mut self, source_idx: usize, is_on_after_key: bool) -> crate::Result<()> {
if source_idx == self.composite_agg_data.req.sources.len() {
if !is_on_after_key {
collect_bucket_with_limit(
self.doc_id,
self.composite_agg_data.req.size as usize,
self.buckets,
&self.sub_level_values,
self.sub_agg,
self.bucket_id_provider,
);
}
let current_level_accessors = &composite_agg_data.composite_accessors[source_idx_for_recursion];
let current_level_source = &composite_agg_data.req.sources[source_idx_for_recursion];
let mut missing = true;
for (accessor_idx, accessor) in current_level_accessors.accessors.iter().enumerate() {
let values = accessor.column.values_for_doc(doc_id);
for value in values {
missing = false;
match current_level_source {
CompositeAggregationSource::Terms(_) => {
let preceeds_after_key_type =
accessor_idx < current_level_accessors.after_key_accessor_idx;
if is_on_after_key && preceeds_after_key_type {
break;
}
let matches_after_key_type =
accessor_idx == current_level_accessors.after_key_accessor_idx;
if matches_after_key_type && is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(value),
Order::Desc => current_level_accessors.after_key.lt(value),
};
if should_skip {
continue;
}
}
sub_level_values.push(InternalValueRepr::new_term(
value,
accessor_idx as u8,
current_level_source.order(),
));
let still_on_after_key =
matches_after_key_type && current_level_accessors.after_key.equals(value);
recursive_key_visitor(
doc_id,
agg_data,
composite_agg_data,
source_idx_for_recursion + 1,
sub_level_values,
buckets,
is_on_after_key && still_on_after_key,
)?;
sub_level_values.pop();
}
CompositeAggregationSource::Histogram(source) => {
let float_value = match accessor.column_type {
ColumnType::U64 => value as f64,
ColumnType::I64 => i64::from_u64(value) as f64,
ColumnType::DateTime => i64::from_u64(value) as f64 / 1_000_000.,
ColumnType::F64 => f64::from_u64(value),
_ => {
panic!(
"unexpected type {:?}. This should not happen",
accessor.column_type
)
}
};
let bucket_index = (float_value / source.interval).floor() as i64;
let bucket_value = i64::to_u64(bucket_index);
if is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(bucket_value),
Order::Desc => current_level_accessors.after_key.lt(bucket_value),
};
if should_skip {
continue;
}
}
sub_level_values.push(InternalValueRepr::new_histogram(
bucket_value,
current_level_source.order(),
));
let still_on_after_key = current_level_accessors.after_key.equals(bucket_value);
recursive_key_visitor(
doc_id,
agg_data,
composite_agg_data,
source_idx_for_recursion + 1,
sub_level_values,
buckets,
is_on_after_key && still_on_after_key,
)?;
sub_level_values.pop();
}
CompositeAggregationSource::DateHistogram(_) => {
let value_ns = match accessor.column_type {
ColumnType::DateTime => i64::from_u64(value),
_ => {
panic!(
"unexpected type {:?}. This should not happen",
accessor.column_type
)
}
};
let bucket_index = match accessor.date_histogram_interval {
PrecomputedDateInterval::FixedNanoseconds(fixed_interval_ns) => {
(value_ns / fixed_interval_ns) * fixed_interval_ns
}
PrecomputedDateInterval::Calendar(CalendarInterval::Year) => {
calendar_interval::try_year_bucket(value_ns)?
}
PrecomputedDateInterval::Calendar(CalendarInterval::Month) => {
calendar_interval::try_month_bucket(value_ns)?
}
PrecomputedDateInterval::Calendar(CalendarInterval::Week) => {
calendar_interval::week_bucket(value_ns)
}
PrecomputedDateInterval::NotApplicable => {
panic!("interval not precomputed for date histogram source")
}
};
let bucket_value = i64::to_u64(bucket_index);
if is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(bucket_value),
Order::Desc => current_level_accessors.after_key.lt(bucket_value),
};
if should_skip {
continue;
}
}
sub_level_values.push(InternalValueRepr::new_histogram(
bucket_value,
current_level_source.order(),
));
let still_on_after_key = current_level_accessors.after_key.equals(bucket_value);
recursive_key_visitor(
doc_id,
agg_data,
composite_agg_data,
source_idx_for_recursion + 1,
sub_level_values,
buckets,
is_on_after_key && still_on_after_key,
)?;
sub_level_values.pop();
}
};
}
}
if missing && current_level_source.missing_bucket() {
if is_on_after_key && current_level_accessors.skip_missing {
return Ok(());
}
let current_level_accessors = &self.composite_agg_data.composite_accessors[source_idx];
let current_level_source = &self.composite_agg_data.req.sources[source_idx];
let mut missing = true;
for (accessor_idx, accessor) in current_level_accessors.accessors.iter().enumerate() {
let values = accessor.column.values_for_doc(self.doc_id);
for value in values {
missing = false;
match current_level_source {
CompositeAggregationSource::Terms(_) => {
let preceeds_after_key_type =
accessor_idx < current_level_accessors.after_key_accessor_idx;
if is_on_after_key && preceeds_after_key_type {
break;
}
let matches_after_key_type =
accessor_idx == current_level_accessors.after_key_accessor_idx;
if matches_after_key_type && is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(value),
Order::Desc => current_level_accessors.after_key.lt(value),
};
if should_skip {
continue;
}
}
self.sub_level_values.push(InternalValueRepr::new_term(
value,
accessor_idx as u8,
current_level_source.order(),
));
let still_on_after_key = matches_after_key_type
&& current_level_accessors.after_key.equals(value);
self.visit(source_idx + 1, is_on_after_key && still_on_after_key)?;
self.sub_level_values.pop();
}
CompositeAggregationSource::Histogram(source) => {
let float_value = match accessor.column_type {
ColumnType::U64 => value as f64,
ColumnType::I64 => i64::from_u64(value) as f64,
ColumnType::DateTime => i64::from_u64(value) as f64 / 1_000_000.,
ColumnType::F64 => f64::from_u64(value),
_ => {
panic!(
"unexpected type {:?}. This should not happen",
accessor.column_type
)
}
};
let bucket_index = (float_value / source.interval).floor() as i64;
let bucket_value = i64::to_u64(bucket_index);
if is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(bucket_value),
Order::Desc => current_level_accessors.after_key.lt(bucket_value),
};
if should_skip {
continue;
}
}
self.sub_level_values.push(InternalValueRepr::new_histogram(
bucket_value,
current_level_source.order(),
));
let still_on_after_key =
current_level_accessors.after_key.equals(bucket_value);
self.visit(source_idx + 1, is_on_after_key && still_on_after_key)?;
self.sub_level_values.pop();
}
CompositeAggregationSource::DateHistogram(_) => {
let value_ns = match accessor.column_type {
ColumnType::DateTime => i64::from_u64(value),
_ => {
panic!(
"unexpected type {:?}. This should not happen",
accessor.column_type
)
}
};
let bucket_index = match accessor.date_histogram_interval {
PrecomputedDateInterval::FixedNanoseconds(fixed_interval_ns) => {
(value_ns / fixed_interval_ns) * fixed_interval_ns
}
PrecomputedDateInterval::Calendar(CalendarInterval::Year) => {
calendar_interval::try_year_bucket(value_ns)?
}
PrecomputedDateInterval::Calendar(CalendarInterval::Month) => {
calendar_interval::try_month_bucket(value_ns)?
}
PrecomputedDateInterval::Calendar(CalendarInterval::Week) => {
calendar_interval::week_bucket(value_ns)
}
PrecomputedDateInterval::NotApplicable => {
panic!("interval not precomputed for date histogram source")
}
};
let bucket_value = i64::to_u64(bucket_index);
if is_on_after_key {
let should_skip = match current_level_source.order() {
Order::Asc => current_level_accessors.after_key.gt(bucket_value),
Order::Desc => current_level_accessors.after_key.lt(bucket_value),
};
if should_skip {
continue;
}
}
self.sub_level_values.push(InternalValueRepr::new_histogram(
bucket_value,
current_level_source.order(),
));
let still_on_after_key =
current_level_accessors.after_key.equals(bucket_value);
self.visit(source_idx + 1, is_on_after_key && still_on_after_key)?;
self.sub_level_values.pop();
}
};
}
}
if missing && current_level_source.missing_bucket() {
if is_on_after_key && current_level_accessors.skip_missing {
return Ok(());
}
self.sub_level_values.push(InternalValueRepr::new_missing(
current_level_source.order(),
current_level_source.missing_order(),
));
self.visit(
source_idx + 1,
is_on_after_key && current_level_accessors.is_after_key_explicit_missing,
)?;
self.sub_level_values.pop();
}
Ok(())
sub_level_values.push(InternalValueRepr::new_missing(
current_level_source.order(),
current_level_source.missing_order(),
));
recursive_key_visitor(
doc_id,
agg_data,
composite_agg_data,
source_idx_for_recursion + 1,
sub_level_values,
buckets,
is_on_after_key && current_level_accessors.is_after_key_explicit_missing,
)?;
sub_level_values.pop();
}
Ok(())
}

37
src/aggregation/bucket/composite/map.rs
View File

@@ -66,6 +66,10 @@ impl<K: Copy + Ord + Clone + 'static, V: 'static, const S: usize> ArrayHeapMap<K
.map(|(k, v)| (SmallVec::from_slice(&k), v)),
)
}
fn values_mut<'a>(&'a mut self) -> Box<dyn Iterator<Item = &'a mut V> + 'a> {
Box::new(self.buckets.values_mut())
}
}
pub(super) const MAX_DYN_ARRAY_SIZE: usize = 16;
@@ -297,6 +301,28 @@ impl<K: Ord + Clone + Copy + 'static, V: 'static> DynArrayHeapMap<K, V> {
DynArrayHeapMapInner::Dim16(map) => map.into_iter(),
}
}
/// Returns an iterator over mutable references to the values in the map.
pub(super) fn values_mut(&mut self) -> impl Iterator<Item = &mut V> {
match &mut self.0 {
DynArrayHeapMapInner::Dim1(map) => map.values_mut(),
DynArrayHeapMapInner::Dim2(map) => map.values_mut(),
DynArrayHeapMapInner::Dim3(map) => map.values_mut(),
DynArrayHeapMapInner::Dim4(map) => map.values_mut(),
DynArrayHeapMapInner::Dim5(map) => map.values_mut(),
DynArrayHeapMapInner::Dim6(map) => map.values_mut(),
DynArrayHeapMapInner::Dim7(map) => map.values_mut(),
DynArrayHeapMapInner::Dim8(map) => map.values_mut(),
DynArrayHeapMapInner::Dim9(map) => map.values_mut(),
DynArrayHeapMapInner::Dim10(map) => map.values_mut(),
DynArrayHeapMapInner::Dim11(map) => map.values_mut(),
DynArrayHeapMapInner::Dim12(map) => map.values_mut(),
DynArrayHeapMapInner::Dim13(map) => map.values_mut(),
DynArrayHeapMapInner::Dim14(map) => map.values_mut(),
DynArrayHeapMapInner::Dim15(map) => map.values_mut(),
DynArrayHeapMapInner::Dim16(map) => map.values_mut(),
}
}
}
#[cfg(test)]
@@ -319,11 +345,20 @@ mod tests {
assert_eq!(map.size(), 1);
assert_eq!(map.peek_highest(), Some(&key1[..]));
// mutable iterator
{
let mut mut_iter = map.values_mut();
let v = mut_iter.next().unwrap();
assert_eq!(*v, "a");
*v = "c";
assert_eq!(mut_iter.next(), None);
}
// into_iter
let mut iter = map.into_iter();
let (k, v) = iter.next().unwrap();
assert_eq!(k.as_slice(), &key1);
assert_eq!(v, "a");
assert_eq!(v, "c");
assert_eq!(iter.next(), None);
}
}

125
src/aggregation/bucket/composite/mod.rs
View File

@@ -338,89 +338,76 @@ impl ToTypePaginationOrder for CompositeKey {
}
}
/// After key is a string that encodes the intermediate composite key as "<type>:<value>"
/// A wrapper type for CompositeIntermediateKey that serializes/deserializes
/// to/from the "<type>:<value>" format.
/// A wrapper type for CompositeIntermediateKey that serializes to ES-compatible
/// raw values (strings as strings, numbers as numbers, etc.) and deserializes
/// from both raw ES format and the legacy "<type>:<value>" format.
#[derive(Clone, Debug, PartialEq)]
pub struct AfterKey(pub CompositeIntermediateKey);
impl Serialize for AfterKey {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where S: serde::Serializer {
let s = match &self.0 {
CompositeIntermediateKey::Bool(b) => format!("bool:{}", b),
CompositeIntermediateKey::Str(s) => format!("str:{}", s),
CompositeIntermediateKey::I64(i) => format!("i64:{}", i),
CompositeIntermediateKey::U64(u) => format!("u64:{}", u),
CompositeIntermediateKey::F64(f) => format!("f64:{}", f),
CompositeIntermediateKey::IpAddr(ip) => format!("ip:{}", ip),
CompositeIntermediateKey::DateTime(dt) => format!("dt:{}", dt),
CompositeIntermediateKey::Null => "null:".to_string(),
};
serializer.serialize_str(&s)
match &self.0 {
CompositeIntermediateKey::Bool(b) => serializer.serialize_bool(*b),
CompositeIntermediateKey::Str(s) => serializer.serialize_str(s),
CompositeIntermediateKey::I64(i) => serializer.serialize_i64(*i),
CompositeIntermediateKey::U64(u) => serializer.serialize_u64(*u),
CompositeIntermediateKey::F64(f) => serializer.serialize_f64(*f),
CompositeIntermediateKey::IpAddr(ip) => serializer.serialize_str(&ip.to_string()),
CompositeIntermediateKey::DateTime(dt) => serializer.serialize_i64(*dt),
CompositeIntermediateKey::Null => serializer.serialize_none(),
}
}
}
impl<'de> Deserialize<'de> for AfterKey {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where D: serde::Deserializer<'de> {
let s = String::deserialize(deserializer)?;
let parts: Vec<&str> = s.splitn(2, ':').collect();
use serde::de;
if parts.len() != 2 {
return Err(serde::de::Error::custom("invalid after key format"));
struct AfterKeyVisitor;
impl<'de> de::Visitor<'de> for AfterKeyVisitor {
type Value = AfterKey;
fn expecting(&self, formatter: &mut std::fmt::Formatter) -> std::fmt::Result {
formatter.write_str("a string, number, boolean, or null")
}
fn visit_bool<E: de::Error>(self, v: bool) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::Bool(v)))
}
fn visit_i64<E: de::Error>(self, v: i64) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::I64(v)))
}
fn visit_u64<E: de::Error>(self, v: u64) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::U64(v)))
}
fn visit_f64<E: de::Error>(self, v: f64) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::F64(v)))
}
fn visit_str<E: de::Error>(self, v: &str) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::Str(v.to_string())))
}
fn visit_string<E: de::Error>(self, v: String) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::Str(v)))
}
fn visit_none<E: de::Error>(self) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::Null))
}
fn visit_unit<E: de::Error>(self) -> Result<AfterKey, E> {
Ok(AfterKey(CompositeIntermediateKey::Null))
}
}
let key = match parts[0] {
"bool" => {
let b = parts[1].parse::<bool>().map_err(|e| {
serde::de::Error::custom(format!("failed to parse bool: {}", e))
})?;
CompositeIntermediateKey::Bool(b)
}
"str" => CompositeIntermediateKey::Str(parts[1].to_string()),
"i64" => {
let i = parts[1]
.parse::<i64>()
.map_err(|e| serde::de::Error::custom(format!("failed to parse i64: {}", e)))?;
CompositeIntermediateKey::I64(i)
}
"u64" => {
let u = parts[1]
.parse::<u64>()
.map_err(|e| serde::de::Error::custom(format!("failed to parse u64: {}", e)))?;
CompositeIntermediateKey::U64(u)
}
"f64" => {
let f = parts[1]
.parse::<f64>()
.map_err(|e| serde::de::Error::custom(format!("failed to parse f64: {}", e)))?;
if f.is_nan() {
return Err(serde::de::Error::custom(
"NaN is not supported in after key",
));
}
CompositeIntermediateKey::F64(f)
}
"ip" => {
let ip = IpAddr::from_str(parts[1]).map_err(|e: AddrParseError| {
serde::de::Error::custom(format!("failed to parse ip: {}", e))
})?;
CompositeIntermediateKey::IpAddr(ip.into_ipv6_addr())
}
"dt" => {
let dt = parts[1].parse::<i64>().map_err(|e| {
serde::de::Error::custom(format!("failed to parse datetime: {}", e))
})?;
CompositeIntermediateKey::DateTime(dt)
}
"null" => CompositeIntermediateKey::Null,
_ => {
return Err(serde::de::Error::custom("invalid after key type"));
}
};
Ok(AfterKey(key))
deserializer.deserialize_any(AfterKeyVisitor)
}
}
@@ -820,7 +807,7 @@ mod tests {
{"key": {"myterm": "apple"}, "doc_count": 1}
])
);
assert!(res["fruity_aggreg"].get("after_key").is_none());
assert!(res["my_composite"].get("after_key").is_none());
Ok(())
}

8
src/aggregation/bucket/composite/numeric_types.rs
View File

@@ -1,4 +1,4 @@
/// This module helps comparing numerical values of different types (i64, u64
/// This modules helps comparing numerical values of different types (i64, u64
/// and f64).
pub(super) mod num_cmp {
use std::cmp::Ordering;
@@ -93,7 +93,7 @@ pub(super) mod num_cmp {
}
}
/// This module helps projecting numerical values to other numerical types.
/// This modules helps projecting numerical values to other numerical types.
/// When the target value space cannot exactly represent the source value, the
/// next representable value is returned (or AfterLast if the source value is
/// larger than the largest representable value).
@@ -138,9 +138,9 @@ pub(super) mod num_proj {
pub fn f64_to_i64(value: f64) -> ProjectedNumber<i64> {
if value < (i64::MIN as f64) {
ProjectedNumber::Next(i64::MIN)
return ProjectedNumber::Next(i64::MIN);
} else if value >= (i64::MAX as f64) {
ProjectedNumber::AfterLast
return ProjectedNumber::AfterLast;
} else if value.fract() == 0.0 {
ProjectedNumber::Exact(value as i64)
} else if value > 0.0 {

18
src/aggregation/bucket/term_agg.rs
View File

@@ -807,13 +807,11 @@ impl<TermMap: TermAggregationMap, C: SubAggCache> SegmentAggregationCollector
let req_data = &mut self.terms_req_data;
agg_data
.column_block_accessor
.fetch_block_with_missing_unique_per_doc(
docs,
&req_data.accessor,
req_data.missing_value_for_accessor,
);
agg_data.column_block_accessor.fetch_block_with_missing(
docs,
&req_data.accessor,
req_data.missing_value_for_accessor,
);
if let Some(sub_agg) = &mut self.sub_agg {
let term_buckets = &mut self.parent_buckets[parent_bucket_id as usize];
@@ -2349,7 +2347,7 @@ mod tests {
// text field
assert_eq!(res["my_texts"]["buckets"][0]["key"], "Hello Hello");
assert_eq!(res["my_texts"]["buckets"][0]["doc_count"], 4);
assert_eq!(res["my_texts"]["buckets"][0]["doc_count"], 5);
assert_eq!(res["my_texts"]["buckets"][1]["key"], "Empty");
assert_eq!(res["my_texts"]["buckets"][1]["doc_count"], 2);
assert_eq!(
@@ -2358,7 +2356,7 @@ mod tests {
);
// text field with number as missing fallback
assert_eq!(res["my_texts2"]["buckets"][0]["key"], "Hello Hello");
assert_eq!(res["my_texts2"]["buckets"][0]["doc_count"], 4);
assert_eq!(res["my_texts2"]["buckets"][0]["doc_count"], 5);
assert_eq!(res["my_texts2"]["buckets"][1]["key"], 1337.0);
assert_eq!(res["my_texts2"]["buckets"][1]["doc_count"], 2);
assert_eq!(
@@ -2372,7 +2370,7 @@ mod tests {
assert_eq!(res["my_ids"]["buckets"][0]["key"], 1337.0);
assert_eq!(res["my_ids"]["buckets"][0]["doc_count"], 4);
assert_eq!(res["my_ids"]["buckets"][1]["key"], 1.0);
assert_eq!(res["my_ids"]["buckets"][1]["doc_count"], 2);
assert_eq!(res["my_ids"]["buckets"][1]["doc_count"], 3);
assert_eq!(res["my_ids"]["buckets"][2]["key"], serde_json::Value::Null);
Ok(())

138
src/aggregation/intermediate_agg_result.rs
View File

@@ -15,9 +15,8 @@ use serde::{Deserialize, Serialize};
use super::agg_req::{Aggregation, AggregationVariants, Aggregations};
use super::agg_result::{AggregationResult, BucketResult, MetricResult, RangeBucketEntry};
use super::bucket::{
composite_intermediate_key_ordering, cut_off_buckets, get_agg_name_and_property,
intermediate_histogram_buckets_to_final_buckets, CompositeAggregation, GetDocCount,
MissingOrder, Order, OrderTarget, RangeAggregation, TermsAggregation,
cut_off_buckets, get_agg_name_and_property, intermediate_histogram_buckets_to_final_buckets,
GetDocCount, Order, OrderTarget, RangeAggregation, TermsAggregation,
};
use super::metric::{
IntermediateAverage, IntermediateCount, IntermediateExtendedStats, IntermediateMax,
@@ -28,7 +27,10 @@ use super::{format_date, AggregationError, Key, SerializedKey};
use crate::aggregation::agg_result::{
AggregationResults, BucketEntries, BucketEntry, CompositeBucketEntry, FilterBucketResult,
};
use crate::aggregation::bucket::TermsAggregationInternal;
use crate::aggregation::bucket::{
composite_intermediate_key_ordering, CompositeAggregation, MissingOrder,
TermsAggregationInternal,
};
use crate::aggregation::metric::CardinalityCollector;
use crate::TantivyError;
@@ -247,6 +249,11 @@ pub(crate) fn empty_from_req(req: &Aggregation) -> IntermediateAggregationResult
is_date_agg: true,
})
}
Composite(_) => {
IntermediateAggregationResult::Bucket(IntermediateBucketResult::Composite {
buckets: Default::default(),
})
}
Average(_) => IntermediateAggregationResult::Metric(IntermediateMetricResult::Average(
IntermediateAverage::default(),
)),
@@ -281,11 +288,6 @@ pub(crate) fn empty_from_req(req: &Aggregation) -> IntermediateAggregationResult
doc_count: 0,
sub_aggregations: IntermediateAggregationResults::default(),
}),
Composite(_) => {
IntermediateAggregationResult::Bucket(IntermediateBucketResult::Composite {
buckets: IntermediateCompositeBucketResult::default(),
})
}
}
}
@@ -579,13 +581,13 @@ impl IntermediateBucketResult {
sub_aggregations: final_sub_aggregations,
}))
}
IntermediateBucketResult::Composite { buckets } => {
let composite_req = req
.agg
IntermediateBucketResult::Composite { buckets } => buckets.into_final_result(
req.agg
.as_composite()
.expect("unexpected aggregation, expected composite aggregation");
buckets.into_final_result(composite_req, req.sub_aggregation(), limits)
}
.expect("unexpected aggregation, expected composite aggregation"),
req.sub_aggregation(),
limits,
),
}
}
@@ -654,13 +656,13 @@ impl IntermediateBucketResult {
}
(
IntermediateBucketResult::Composite {
buckets: composite_left,
buckets: buckets_left,
},
IntermediateBucketResult::Composite {
buckets: composite_right,
buckets: buckets_right,
},
) => {
composite_left.merge_fruits(composite_right)?;
buckets_left.merge_fruits(buckets_right)?;
}
(IntermediateBucketResult::Range(_), _) => {
panic!("try merge on different types")
@@ -920,31 +922,6 @@ pub struct IntermediateTermBucketEntry {
pub sub_aggregation: IntermediateAggregationResults,
}
impl MergeFruits for IntermediateTermBucketEntry {
fn merge_fruits(&mut self, other: IntermediateTermBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation.merge_fruits(other.sub_aggregation)?;
Ok(())
}
}
impl MergeFruits for IntermediateRangeBucketEntry {
fn merge_fruits(&mut self, other: IntermediateRangeBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation_res
.merge_fruits(other.sub_aggregation_res)?;
Ok(())
}
}
impl MergeFruits for IntermediateHistogramBucketEntry {
fn merge_fruits(&mut self, other: IntermediateHistogramBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation.merge_fruits(other.sub_aggregation)?;
Ok(())
}
}
/// Entry for the composite bucket.
pub type IntermediateCompositeBucketEntry = IntermediateTermBucketEntry;
@@ -990,11 +967,41 @@ impl std::hash::Hash for CompositeIntermediateKey {
/// Composite aggregation page.
#[derive(Default, Clone, Debug, PartialEq, Serialize, Deserialize)]
pub struct IntermediateCompositeBucketResult {
#[serde(
serialize_with = "serialize_composite_entries",
deserialize_with = "deserialize_composite_entries"
)]
pub(crate) entries: FxHashMap<Vec<CompositeIntermediateKey>, IntermediateCompositeBucketEntry>,
pub(crate) target_size: u32,
pub(crate) orders: Vec<(Order, MissingOrder)>,
}
fn serialize_composite_entries<S>(
entries: &FxHashMap<Vec<CompositeIntermediateKey>, IntermediateCompositeBucketEntry>,
serializer: S,
) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
{
use serde::ser::SerializeSeq;
let mut seq = serializer.serialize_seq(Some(entries.len()))?;
for (k, v) in entries {
seq.serialize_element(&(k, v))?;
}
seq.end()
}
fn deserialize_composite_entries<'de, D>(
deserializer: D,
) -> Result<FxHashMap<Vec<CompositeIntermediateKey>, IntermediateCompositeBucketEntry>, D::Error>
where
D: serde::Deserializer<'de>,
{
let vec: Vec<(Vec<CompositeIntermediateKey>, IntermediateCompositeBucketEntry)> =
serde::Deserialize::deserialize(deserializer)?;
Ok(vec.into_iter().collect())
}
impl IntermediateCompositeBucketResult {
pub(crate) fn into_final_result(
self,
@@ -1050,12 +1057,16 @@ impl IntermediateCompositeBucketResult {
fn merge_fruits(&mut self, other: IntermediateCompositeBucketResult) -> crate::Result<()> {
merge_maps(&mut self.entries, other.entries)?;
if self.entries.len() as u32 > 2 * self.target_size {
// 2x factor used to avoid trimming too often (expensive operation)
// an optimal threshold could probably be figured out
self.trim()?;
}
Ok(())
}
/// Trim the composite buckets to the target size, according to the ordering.
///
/// Returns an error if the ordering comparison fails.
pub(crate) fn trim(&mut self) -> crate::Result<()> {
if self.entries.len() as u32 <= self.target_size {
return Ok(());
@@ -1085,19 +1096,20 @@ fn trim_composite_buckets(
let mut entries: Vec<_> = entries.into_iter().collect();
let mut sort_error: Option<TantivyError> = None;
entries.sort_by(|(left_key, _), (right_key, _)| {
// Only attempt sorting if we haven't encountered an error yet
if sort_error.is_some() {
return Ordering::Equal;
return Ordering::Equal; // Return a default, we'll handle the error after sorting
}
for idx in 0..orders.len() {
for i in 0..orders.len() {
match composite_intermediate_key_ordering(
&left_key[idx],
&right_key[idx],
orders[idx].0,
orders[idx].1,
&left_key[i],
&right_key[i],
orders[i].0,
orders[i].1,
) {
Ok(ordering) if ordering != Ordering::Equal => return ordering,
Ok(_) => continue,
Ok(_) => continue, // Equal, try next key
Err(err) => {
sort_error = Some(err);
break;
@@ -1107,6 +1119,7 @@ fn trim_composite_buckets(
Ordering::Equal
});
// If we encountered an error during sorting, return it now
if let Some(err) = sort_error {
return Err(err);
}
@@ -1115,6 +1128,31 @@ fn trim_composite_buckets(
Ok(entries)
}
impl MergeFruits for IntermediateTermBucketEntry {
fn merge_fruits(&mut self, other: IntermediateTermBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation.merge_fruits(other.sub_aggregation)?;
Ok(())
}
}
impl MergeFruits for IntermediateRangeBucketEntry {
fn merge_fruits(&mut self, other: IntermediateRangeBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation_res
.merge_fruits(other.sub_aggregation_res)?;
Ok(())
}
}
impl MergeFruits for IntermediateHistogramBucketEntry {
fn merge_fruits(&mut self, other: IntermediateHistogramBucketEntry) -> crate::Result<()> {
self.doc_count += other.doc_count;
self.sub_aggregation.merge_fruits(other.sub_aggregation)?;
Ok(())
}
}
#[cfg(test)]
mod tests {
use std::collections::HashMap;

28
src/aggregation/metric/percentiles.rs
View File

@@ -331,7 +331,7 @@ mod tests {
use crate::aggregation::AggregationCollector;
use crate::query::AllQuery;
use crate::schema::{Schema, FAST};
use crate::{assert_nearly_equals, Index};
use crate::Index;
#[test]
fn test_aggregation_percentiles_empty_index() -> crate::Result<()> {
@@ -614,17 +614,13 @@ mod tests {
let res = exec_request_with_query(agg_req, &index, None)?;
assert_eq!(res["range_with_stats"]["buckets"][0]["doc_count"], 3);
assert_nearly_equals!(
res["range_with_stats"]["buckets"][0]["percentiles"]["values"]["1.0"]
.as_f64()
.unwrap(),
5.0028295751107414
assert_eq!(
res["range_with_stats"]["buckets"][0]["percentiles"]["values"]["1.0"],
5.002829575110705
);
assert_nearly_equals!(
res["range_with_stats"]["buckets"][0]["percentiles"]["values"]["99.0"]
.as_f64()
.unwrap(),
10.07469668951144
assert_eq!(
res["range_with_stats"]["buckets"][0]["percentiles"]["values"]["99.0"],
10.07469668951133
);
Ok(())
@@ -669,14 +665,8 @@ mod tests {
let res = exec_request_with_query(agg_req, &index, None)?;
assert_nearly_equals!(
res["percentiles"]["values"]["1.0"].as_f64().unwrap(),
5.0028295751107414
);
assert_nearly_equals!(
res["percentiles"]["values"]["99.0"].as_f64().unwrap(),
10.07469668951144
);
assert_eq!(res["percentiles"]["values"]["1.0"], 5.002829575110705);
assert_eq!(res["percentiles"]["values"]["99.0"], 10.07469668951133);
Ok(())
}

1
src/directory/composite_file.rs
View File

@@ -167,7 +167,6 @@ impl CompositeFile {
.map(|byte_range| self.data.slice(byte_range.clone()))
}
/// Returns the space usage per field in this composite file.
pub fn space_usage(&self, schema: &Schema) -> PerFieldSpaceUsage {
let mut fields = Vec::new();
for (&field_addr, byte_range) in &self.offsets_index {

4
src/directory/mod.rs
View File

@@ -21,7 +21,7 @@ use std::path::PathBuf;
pub use common::file_slice::{FileHandle, FileSlice};
pub use common::{AntiCallToken, OwnedBytes, TerminatingWrite};
pub use self::composite_file::{CompositeFile, CompositeWrite};
pub(crate) use self::composite_file::{CompositeFile, CompositeWrite};
pub use self::directory::{Directory, DirectoryClone, DirectoryLock};
pub use self::directory_lock::{Lock, INDEX_WRITER_LOCK, META_LOCK};
pub use self::ram_directory::RamDirectory;
@@ -52,7 +52,7 @@ pub use self::mmap_directory::MmapDirectory;
///
/// `WritePtr` are required to implement both Write
/// and Seek.
pub type WritePtr = BufWriter<Box<dyn TerminatingWrite + Send + Sync>>;
pub type WritePtr = BufWriter<Box<dyn TerminatingWrite>>;
#[cfg(test)]
mod tests;

19
src/indexer/log_merge_policy.rs
View File

@@ -94,7 +94,7 @@ impl MergePolicy for LogMergePolicy {
fn compute_merge_candidates(&self, segments: &[SegmentMeta]) -> Vec<MergeCandidate> {
let size_sorted_segments = segments
.iter()
.filter(|seg| (seg.num_docs() as usize) <= self.max_docs_before_merge)
.filter(|seg| seg.num_docs() <= (self.max_docs_before_merge as u32))
.sorted_by_key(|seg| std::cmp::Reverse(seg.max_doc()))
.collect::<Vec<&SegmentMeta>>();
@@ -372,21 +372,4 @@ mod tests {
assert_eq!(merge_candidates[0].0.len(), 1);
assert_eq!(merge_candidates[0].0[0], test_input[1].id());
}
#[test]
fn test_max_docs_before_merge_large_value() {
// Regression test: (max_docs_before_merge as u32) truncates values > u32::MAX.
// Casting num_docs() to usize instead avoids the truncation.
let mut policy = LogMergePolicy::default();
policy.set_min_num_segments(2);
policy.set_max_docs_before_merge(5_000_000_000usize);
let test_input = vec![
create_random_segment_meta(100_000),
create_random_segment_meta(100_000),
];
let result = policy.compute_merge_candidates(&test_input);
// Both segments should be eligible (100_000 < 5_000_000_000)
assert_eq!(result.len(), 1);
assert_eq!(result[0].0.len(), 2);
}
}

23
src/indexer/segment_updater.rs
View File

@@ -403,8 +403,7 @@ impl SegmentUpdater {
// from the different drives.
//
// Segment 1 from disk 1, Segment 1 from disk 2, etc.
committed_segment_metas
.sort_by_key(|segment_meta| std::cmp::Reverse(segment_meta.max_doc()));
committed_segment_metas.sort_by_key(|segment_meta| -(segment_meta.max_doc() as i32));
let index_meta = IndexMeta {
index_settings: index.settings().clone(),
segments: committed_segment_metas,
@@ -649,6 +648,9 @@ impl SegmentUpdater {
merge_operation.segment_ids(),
advance_deletes_err
);
assert!(!cfg!(test), "Merge failed.");
// ... cancel merge
// `merge_operations` are tracked. As it is dropped, the
// the segment_ids will be available again for merge.
return Err(advance_deletes_err);
@@ -703,7 +705,6 @@ mod tests {
use crate::collector::TopDocs;
use crate::directory::RamDirectory;
use crate::fastfield::AliveBitSet;
use crate::index::{SegmentId, SegmentMetaInventory};
use crate::indexer::merge_policy::tests::MergeWheneverPossible;
use crate::indexer::merger::IndexMerger;
use crate::indexer::segment_updater::merge_filtered_segments;
@@ -711,22 +712,6 @@ mod tests {
use crate::schema::*;
use crate::{Directory, DocAddress, Index, Segment};
#[test]
fn test_segment_sort_large_max_doc() {
// Regression test: -(max_doc as i32) overflows for max_doc >= 2^31.
// Using std::cmp::Reverse avoids this.
let inventory = SegmentMetaInventory::default();
let mut metas = [
inventory.new_segment_meta(SegmentId::generate_random(), 100),
inventory.new_segment_meta(SegmentId::generate_random(), (1u32 << 31) - 1),
inventory.new_segment_meta(SegmentId::generate_random(), 50_000),
];
metas.sort_by_key(|m| std::cmp::Reverse(m.max_doc()));
assert_eq!(metas[0].max_doc(), (1u32 << 31) - 1);
assert_eq!(metas[1].max_doc(), 50_000);
assert_eq!(metas[2].max_doc(), 100);
}
#[test]
fn test_delete_during_merge() -> crate::Result<()> {
let mut schema_builder = Schema::builder();

4
src/lib.rs
View File

@@ -169,10 +169,8 @@ mod macros;
mod future_result;
// Re-exports
pub use columnar;
pub use common::{ByteCount, DateTime};
pub use query_grammar;
pub use time;
pub use {columnar, query_grammar, time};
pub use crate::error::TantivyError;
pub use crate::future_result::FutureResult;

3
src/postings/mod.rs
View File

@@ -14,8 +14,7 @@ mod postings;
mod postings_writer;
mod recorder;
mod segment_postings;
/// Serializer module for the inverted index
pub mod serializer;
mod serializer;
mod skip;
mod term_info;

51
src/postings/serializer.rs
View File

@@ -11,7 +11,7 @@ use crate::positions::PositionSerializer;
use crate::postings::compression::{BlockEncoder, VIntEncoder, COMPRESSION_BLOCK_SIZE};
use crate::postings::skip::SkipSerializer;
use crate::query::Bm25Weight;
use crate::schema::{Field, FieldEntry, IndexRecordOption, Schema};
use crate::schema::{Field, FieldEntry, FieldType, IndexRecordOption, Schema};
use crate::termdict::TermDictionaryBuilder;
use crate::{DocId, Score};
@@ -80,12 +80,9 @@ impl InvertedIndexSerializer {
let term_dictionary_write = self.terms_write.for_field(field);
let postings_write = self.postings_write.for_field(field);
let positions_write = self.positions_write.for_field(field);
let index_record_option = field_entry
.field_type()
.index_record_option()
.unwrap_or(IndexRecordOption::Basic);
let field_type: FieldType = (*field_entry.field_type()).clone();
FieldSerializer::create(
index_record_option,
&field_type,
total_num_tokens,
term_dictionary_write,
postings_write,
@@ -105,27 +102,29 @@ impl InvertedIndexSerializer {
/// The field serializer is in charge of
/// the serialization of a specific field.
pub struct FieldSerializer<'a, W: Write = WritePtr> {
term_dictionary_builder: TermDictionaryBuilder<&'a mut CountingWriter<W>>,
pub struct FieldSerializer<'a> {
term_dictionary_builder: TermDictionaryBuilder<&'a mut CountingWriter<WritePtr>>,
postings_serializer: PostingsSerializer,
positions_serializer_opt: Option<PositionSerializer<&'a mut CountingWriter<W>>>,
positions_serializer_opt: Option<PositionSerializer<&'a mut CountingWriter<WritePtr>>>,
current_term_info: TermInfo,
term_open: bool,
postings_write: &'a mut CountingWriter<W>,
postings_write: &'a mut CountingWriter<WritePtr>,
postings_start_offset: u64,
}
impl<'a, W: Write> FieldSerializer<'a, W> {
/// Creates a new `FieldSerializer` for the given field type.
pub fn create(
index_record_option: IndexRecordOption,
impl<'a> FieldSerializer<'a> {
fn create(
field_type: &FieldType,
total_num_tokens: u64,
term_dictionary_write: &'a mut CountingWriter<W>,
postings_write: &'a mut CountingWriter<W>,
positions_write: &'a mut CountingWriter<W>,
term_dictionary_write: &'a mut CountingWriter<WritePtr>,
postings_write: &'a mut CountingWriter<WritePtr>,
positions_write: &'a mut CountingWriter<WritePtr>,
fieldnorm_reader: Option<FieldNormReader>,
) -> io::Result<FieldSerializer<'a, W>> {
) -> io::Result<FieldSerializer<'a>> {
total_num_tokens.serialize(postings_write)?;
let index_record_option = field_type
.index_record_option()
.unwrap_or(IndexRecordOption::Basic);
let term_dictionary_builder = TermDictionaryBuilder::create(term_dictionary_write)?;
let average_fieldnorm = fieldnorm_reader
.as_ref()
@@ -193,11 +192,6 @@ impl<'a, W: Write> FieldSerializer<'a, W> {
Ok(())
}
/// Starts the postings for a new term without recording term frequencies.
pub fn new_term_without_freq(&mut self, term: &[u8]) -> io::Result<()> {
self.new_term(term, 0, false)
}
/// Serialize the information that a document contains for the current term:
/// its term frequency, and the position deltas.
///
@@ -303,7 +297,6 @@ impl Block {
}
}
/// Serializer for postings lists.
pub struct PostingsSerializer {
last_doc_id_encoded: u32,
@@ -323,9 +316,6 @@ pub struct PostingsSerializer {
}
impl PostingsSerializer {
/// Creates a new `PostingsSerializer`.
/// * avg_fieldnorm - average field norm for the field being serialized.
/// * mode - indexing options for the field being serialized.
pub fn new(
avg_fieldnorm: Score,
mode: IndexRecordOption,
@@ -348,8 +338,6 @@ impl PostingsSerializer {
}
}
/// Starts the serialization for a new term.
/// * term_doc_freq - the number of documents containing the term.
pub fn new_term(&mut self, term_doc_freq: u32, record_term_freq: bool) {
self.bm25_weight = None;
@@ -389,7 +377,6 @@ impl PostingsSerializer {
self.postings_write.extend(block_encoded);
}
if self.term_has_freq {
// encode the term frequencies
let (num_bits, block_encoded): (u8, &[u8]) = self
.block_encoder
.compress_block_unsorted(self.block.term_freqs(), true);
@@ -430,9 +417,6 @@ impl PostingsSerializer {
self.block.clear();
}
/// Register that the given document contains the current term.
/// * doc_id - the document id.
/// * term_freq - the term frequency within the document.
pub fn write_doc(&mut self, doc_id: DocId, term_freq: u32) {
self.block.append_doc(doc_id, term_freq);
if self.block.is_full() {
@@ -440,7 +424,6 @@ impl PostingsSerializer {
}
}
/// Finish the serialization for this term.
pub fn close_term(
&mut self,
doc_freq: u32,

6
src/postings/skip.rs
View File

@@ -14,11 +14,7 @@ use crate::{DocId, Score, TERMINATED};
// (requiring a 6th bit), but the biggest doc_id we can want to encode is TERMINATED-1, which can
// be represented on 31b without delta encoding.
fn encode_bitwidth(bitwidth: u8, delta_1: bool) -> u8 {
assert!(
bitwidth < 32,
"bitwidth needs to be less than 32, but got {}",
bitwidth
);
assert!(bitwidth < 32);
bitwidth | ((delta_1 as u8) << 6)
}

3
src/termdict/fst_termdict/term_info_store.rs
View File

@@ -48,7 +48,8 @@ impl BinarySerializable for TermInfoBlockMeta {
}
impl FixedSize for TermInfoBlockMeta {
const SIZE_IN_BYTES: usize = u64::SIZE_IN_BYTES + TermInfo::SIZE_IN_BYTES + 3;
const SIZE_IN_BYTES: usize =
u64::SIZE_IN_BYTES + TermInfo::SIZE_IN_BYTES + 3 * u8::SIZE_IN_BYTES;
}
impl TermInfoBlockMeta {

5
sstable/src/lib.rs
View File

@@ -302,9 +302,8 @@ where
|| self.previous_key[keep_len] < key[keep_len];
assert!(
increasing_keys,
"Keys should be increasing. ({:?} > {:?})",
String::from_utf8_lossy(&self.previous_key),
String::from_utf8_lossy(key),
"Keys should be increasing. ({:?} > {key:?})",
self.previous_key
);
self.previous_key.resize(key.len(), 0u8);
self.previous_key[keep_len..].copy_from_slice(&key[keep_len..]);

2
sstable/src/sstable_index_v3.rs
View File

@@ -553,7 +553,7 @@ impl FixedSize for BlockAddrBlockMetadata {
const SIZE_IN_BYTES: usize = u64::SIZE_IN_BYTES
+ BlockStartAddr::SIZE_IN_BYTES
+ 2 * u32::SIZE_IN_BYTES
+ 2
+ 2 * u8::SIZE_IN_BYTES
+ u16::SIZE_IN_BYTES;
}