[Save this search]

Status
All
   Fixed (7187)
  Closed (4804)
Open (2573)
   Won't Fix (545)
   Duplicate (297)
   Invalid (217)
   Not A Problem (195)
Issue type
All
  Issue (2258)
  PR (315)
Author relation
All
  None (2041)
  Member (634)
  Contributor (444)
  New contributor (59)
Created
All
  Past day (3)
  Past 2 days (8)
  Past 3 days (12)
  Past week (24)
  Past month (56)
  Past 3 months (96)
  Past 6 months (144)
  Past year (261)
Updated
All
  Past day (14)
  Past 2 days (31)
  Past 3 days (36)
  Past week (43)
  Past month (84)
  Past 3 months (131)
  Past 6 months (190)
  Past year (489)
Updated ago
All
  > 1 day ago (2559)
  > 2 days ago (2542)
  > 3 days ago (2537)
  > 1 week ago (2530)
  > 1 month ago (2489)
  > 3 months ago (2442)
  > 1 year ago (2084)
Comment count
All
  0 (569)
  1 (361)
  2 - 5 (798)
  6 - 10 (437)
  10 - 20 (304)
  > 20 (173)
Reaction count
All
  0 (2293)
  1 (167)
  2 - 5 (91)
  6 - 10 (16)
  10 - 20 (5)
  > 20 (1)
Review Requested
All
  jpountz (14)
  mikemccand (9)
  iverase (5)
  romseygeek (5)
  rmuir (4)
  benwtrent (4)
  dweiss (3)

See all 22...
Mentioned
All
  jpountz (94)
  mikemccand (76)
  benwtrent (55)
  rmuir (50)
  msokolov (40)
  uschindler (39)
  romseygeek (23)

See all 205...
Reviewed
All
  mikemccand (24)
  jpountz (23)
  benwtrent (15)
  rmuir (12)
  uschindler (12)
  msokolov (11)
  dweiss (10)

See all 63...
Commented
All
  asfimport (1369)
  github-actions[bot] (235)
  jpountz (140)
  mikemccand (122)
  rmuir (95)
  benwtrent (86)
  msokolov (81)

See all 293...
User
All
  asfimport (1811)
  github-actions[bot] (281)
  mikemccand (210)
  jpountz (202)
  rmuir (126)
  benwtrent (124)
  msokolov (103)

See all 446...
Last comment user
All
  asfimport (1326)
  github-actions[bot] (225)
  jpountz (32)
  mikemccand (31)
  rmuir (22)
  msokolov (17)
  benwtrent (17)

See all 160...
Draft
All
  No (260)
  Yes (55)
Component
All
  core (602)
  analysis (151)
  highlighter (46)
  spatial (41)
  facet (40)
  queryparser (26)
  test-framework (25)

See all 23...
Type
All
  enhancement (1182)
  bug (725)
  task (200)
  test (79)
  documentation (22)
Labels
All
  Stale (220)
  legacy-jira-fix-versio... (214)
  legacy-jira-fix-versio... (169)
  affects-version:4.0-ALPHA (81)
  tool:build (54)
  vector-based-search (48)
  affects-version:6.0 (37)

See all 157...
Commits?
All
  No (2573)
Reporter
All
  rmuir (270)
  mikemccand (155)
  jpountz (126)
  dsmiley (68)
  uschindler (52)
  iverase (45)
  romseygeek (42)

See all 746...
Assignee
All
  Unassigned (2328)
  mikemccand (35)
  uschindler (31)
  romseygeek (27)
  dsmiley (22)
  rmuir (15)
  jpountz (11)

See all 45...
  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

#16142 PR: Optimize disjunction docIDRunEnd
5.4 minutes ago  2 comments  0 votes  0 watches  costingithub-actions[bot]jainankitk
ReqExclBulkScorer uses the prohibited scorer's docIDRunEnd() to skip runs of excluded documents. For multi-clause MUST_NOT queries this often goes through DisjunctionDISIApproximation, whose ... This PR limits docIDRunEnd() to heap-led top clauses that are already positioned on the current doc ...
    costin 5.5 minutes ago:  Thanks @jainankitk Re 1: this is what test "testDocIDRunEndOnlyUsesLeadIterators" intends to ... I've renamed the method to make it more clear along with some comments. Re 2: I've expanded the docIDRunEnd() comment.
    jainankitk 5.9 hours ago:  Thanks for fixing this issue. Couple of non-blocking suggestions: 1. Maybe add a test when both heap and otherIterators are positioned on the current doc with the ...

#16143 PR: Tighten constant-score bulk routing
1.1 hours ago  9 comments  0 votes  0 watches  costingithub-actions[bot]jainankitkromseygeek
Follow-up based on the comments from 16141. With ConstantScoreBulkScorer now in main, this routes the remaining constant-score iterator paths ... It also makes the batch-collection scorer-positioning rule explicit, which is the contract 16141 ...
    costin 1.1 hours ago:  Done
    jainankitk 6.1 hours ago:  This seems like a bug fix in addition to other routing changes. Should the CHANGES entry reflect that?

#16144 PR: Document batch collection score contract
1.5 hours ago  7 comments  0 votes  0 watches  costingithub-actions[bot]jainankitk
Split from 16143. Makes the batch-collection scorer-positioning rule explicit, which is the contract 16141 had to ... The constant-score bulk routing work exposed a small contract gap around batch collection: collect( ...
    costin 1.5 hours ago:  A blanket guard would make valid scorables fail because the contract issue is about iterator-backed ... I've added a short comment explaining this.
    costin 1.5 hours ago:  Updated the javadoc.

#16133 PR: FilterLeafCollector should delegate competitiveIterator()
7.5 hours ago  7 comments  0 votes  0 watches  github-actions[bot]javannamrkm4ntr
Description FilterLeafCollector delegates setScorer(), collect(), and finish() to the wrapped ... This causes the default implementation from the LeafCollector interface (return null) to be used, ... One notable case is ProfilerCollectorWrapper in the sandbox module.
    javanna 7.5 hours ago:  Could we check that the competitive iterator returned is meaningful, and provides the expected ...
    javanna 7.5 hours ago:  This is a good test that's specific to the profiler collector, thanks for adding it The fix is ... Otherwise the fix could be specific to the profilter collector wrapper.

#16146 PR: Optimize cached-filter conjunction batching
7.8 hours ago  3 comments  0 votes  0 watches  costingithub-actions[bot]iprithv
This builds on 16141 by allowing DefaultBulkScorer to use the DocIdStream path for dense no-score ... The useful target is cached-filter conjunctions, especially when the cached filter is dense enough ... Those cases show the clearest wins, up to 2.38x in this short run.
    costin 7.8 hours ago:  @iprithv You're right, thanks for catching this! The direct-forwarding default on FilterLeafCollector was too optimistic: it preserves bulk ... That is a correctness regression relative to the old LeafCollector defaults that I overlooked this ...
    iprithv 18.2 hours ago:  FilterLeafCollector is meant to let subclasses override collect(int) and handle each doc. but with ... I see javadoc saying subclasses should override the new methods too, but that’s not really safe. ... a safer default would be to break the batch calls into per-doc calls, like: `java id="safex1" @ ...

#16145 PR: MemorySegmentIndexInput: always prefetch on RANDOM mode
8.2 hours ago  12 comments  0 votes  0 watches  github-actions[bot]iprithvjimczimichaeljmarshallnavneet1v
Description The power-of-two back off in MemorySegmentIndexInput.prefetch() assumes consecutive ... That assumption breaks for ReadAdvice.RANDOM, where each request accesses unpredictable pages — a ... HNSW graph traversal touches a different path per query).
    jimczi 8.2 hours ago:  This could be avoided too in the random case since we don't use this information?
    navneet1v 8.4 hours ago:  @michaeljmarshall did we see any improvements after adding this change?

#15979 PR: Add a de-duplicating flat vector format
14 hours ago  11 comments  0 votes  0 watches  github-actions[bot]kaivalnpmikemccandmsokolov
Description Closes 14758 Add a new de-duplicating vector format that only stores unique ... De-duplication is done for vectors across all docs and fields indexed by the format. Disclaimer: This was mostly written by an AI, with me refining the implementation through prompts - ...
    kaivalnp 1.6 days ago:  Re-did the changes on top of main, benchmark details same as https://github.com/apache/lucene/pull/ ... Common: ` Results: NOTE: nDoc = 500000 for all runs; skipping column NOTE: searchType = KNN ...
    github-actions[bot] 14.6 days 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!

[21.4 msec search, 22.3 msec total]