Lucene: Status=Open

[Save this search]

Author relation

Updated ago

Comment count

Reaction count

Review Requested

User

Last comment user

Type

Filters: Status (Open), Issue type, Author relation, Created, Updated, Updated ago, Comment count, Reaction count, Review Requested, Mentioned, Reviewed, Commented, User, Last comment user, Draft, Component, Type, Labels, Commits?, Reporter, Assignee

#16451: [ENH]: Optimize DocValues Bulk Decoding using Panama Vector API

8.2 hours ago 2 comments 0 votes 0 watches CH-Abhinav, jpountz, mikemccand, rajat315315, rmuir, uschindler

# Motivation Lucene heavily relies on DocValues for sorting, faceting, and range filtering. For byte-aligned packed formats (8-bit, 16-bit, and 32-bit), the bulk decoding process ( ... Currently, the default implementation (DefaultDocValuesBulkDecodeSupport) performs this decoding ...

CH-Abhinav 8.2 hours ago:	@rmuir @uschindler well this issue's author used a fare bit of AI without understanding what he ... Like his test queries and benchmarks do not touch DocValues. But the Idea of using Panama's ZERO_EXTEND isn't bad at all.
rmuir 15.6 days ago:	this is all hallucinated nonsense. > Fuzzy Terms & Edit Distances: Operations like Respell show a statistically significant +24.1% ...

#15113: Move long[] group varint to backward-codecs

12.7 hours ago 6 comments 0 votes 0 watches Exporterhe, SAKSHIC, easyice, jpountz, nishanth-01, subhramit, uschindler

DescriptionThis is a follow-up of #15104. The benchmark was accidentally calling legacy code that writes the output of a varint group into a ... This wouldn't have happened if this code - which is only used by backward codecs - was in lucene/ ...

subhramit 12.7 hours ago:	Hi, this seems to have been abandoned. May I take it? @uschindler, @easyice / @jpountz
SAKSHIC 10.8 months ago:	@jpountz Gentle reminder to review https://github.com/apache/lucene/pull/15186

#16431 PR: Enable Block-Max WAND dynamic pruning for FunctionScoreQuery via DocValuesSkipper (+422% QPS)

18.1 hours ago 45 comments 0 votes 0 watches github-actions[bot], jpountz, mikemccand, msokolov, rajat315315, romseygeek, uschindler

Resolves: 16429 # Description Currently, FunctionScoreQuery hardcodes getMaxScore(int upTo) ... Lucene is forced to evaluate DoubleValuesSource and script bytecode on **every single matching ... Additionally, Lucene's default DocValuesSkipper block interval is 4,096 documents (DEFAULT_SKIP_INDE ...

uschindler 18.1 hours ago:	Sorry this got stale. Will verify test cases.
github-actions[bot] 1 day ago:	This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the dev@lucene.apache.org list. Thank you for your contribution!

#16439 PR: Optimize Non-Overlapping LongRangeFacetCounts via Batching & Fast O(1) Precomputed Lookup

1 day ago 1 comments 0 votes 0 watches github-actions[bot], rajat315315

Resolves: 16436 --- # 🚀 Performance Improvement Summary This PR introduces two ... **Doc Value Batching**: Buffers up to 512 values from the doc values iterator before feeding them ... This reduces the virtual/interface method call overhead per document and improves cache/instruction ...

github-actions[bot] 1 day ago:

This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the dev@lucene.apache.org list. Thank you for your contribution!

#16442 PR: Fix addMultiTermClauses: preserve per-field SHOULD when a clause isn't a bare TermQuery (#16441)

1 day ago 1 comments 0 votes 0 watches OlivierJaquemet, github-actions[bot]

Fixes 16441 **Problem:** QueryParserBaseaddMultiTermClauses decides whether to pass a nested ... This check is too narrow: a BoostQuery-wrapped term (or any other non-TermQuery leaf — PrefixQuery, ... Example: with a per-field boost configured and AND_OPERATOR, MultiFieldQueryParser.parse( ...

github-actions[bot] 1 day ago:

This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the dev@lucene.apache.org list. Thank you for your contribution!

#16446 PR: MultiFieldQueryParser: default operator is not honored between term positions (#16443)

1 day ago 1 comments 0 votes 0 watches OlivierJaquemet, github-actions[bot]

Fixes 16443 **Problem:** MultiFieldQueryParser's javadoc promises that under AND_OPERATOR, "all ... This only holds when each term reaches the parser as a separate grammar-level token (real ... It doesn't hold when several terms are produced by the *analyzer alone* from a single piece of ...

github-actions[bot] 1 day ago:

This PR has not had activity in the past 2 weeks, labeling it as stale. If the PR is waiting for review, notify the dev@lucene.apache.org list. Thank you for your contribution!

#16414 PR: perf: optimize StateSet.equals to eliminate heap allocations during a…

1.5 days ago 8 comments 0 votes 0 watches github-actions[bot], rajat315315, rmuir, uschindler

Closes: 16413 📊 Benchmark & GC Allocation Metrics Evaluated using ThreadMXBean.getThreadAllo ... ** | | **Avg Heap Allocated per determinize()** | **18.29 MB** (19,179,268 bytes) | **3.35 MB** ... ** | | **GC Collection Cycles** | **11 GC cycles** | **2 GC ...

uschindler 1.5 days ago:	I have not yet closely looked into that but we can indeed enforce correct behavior with sealed and ... Basic Ann the common base class of the only two final implementations should be marked sealed with ... Then we can convert the part after the initial instance of check (that ensures classes are ...
rmuir 1.9 days ago:	also if things are really calling StateSet.equals(FrozenIntSet) then possibly also FrozenIntSet.equa ...

«
1
2
3
4
5
»

2,635 issues

[15.7 msec search, 16.6 msec total]