[Save this search]

Status
All
   Fixed (7187)
  Closed (5271)
✓Open (2662)
   Won't Fix (545)
   Duplicate (297)
   Invalid (217)
   Not A Problem (195)
Issue type
✓All
  Issue (2289)
  PR (373)
Author relation
✓All
  None (2068)
  Member (669)
  Contributor (524)
  New contributor (70)
Created
✓All
  Past day (2)
  Past 2 days (7)
  Past 3 days (8)
  Past week (17)
  Past month (58)
  Past 3 months (122)
  Past 6 months (199)
  Past year (295)
Updated
✓All
  Past day (10)
  Past 2 days (18)
  Past 3 days (21)
  Past week (42)
  Past month (90)
  Past 3 months (199)
  Past 6 months (248)
  Past year (552)
Updated ago
✓All
  > 1 day ago (2652)
  > 2 days ago (2644)
  > 3 days ago (2641)
  > 1 week ago (2620)
  > 1 month ago (2572)
  > 3 months ago (2463)
  > 1 year ago (2110)
Comment count
✓All
  0 (591)
  1 (375)
  2 - 5 (832)
  6 - 10 (451)
  10 - 20 (315)
  > 20 (172)
Reaction count
✓All
  0 (2370)
  1 (175)
  2 - 5 (95)
  6 - 10 (16)
  10 - 20 (5)
  > 20 (1)
Review Requested
✓All
  jpountz (14)
  mikemccand (13)
  rmuir (6)
  msokolov (4)
  iverase (4)
  romseygeek (4)
  benwtrent (4)

See all 26...
Mentioned
✓All
  jpountz (100)
  mikemccand (85)
  rmuir (64)
  benwtrent (53)
  uschindler (42)
  msokolov (40)
  romseygeek (20)

See all 217...
Reviewed
✓All
  mikemccand (25)
  jpountz (23)
  rmuir (14)
  benwtrent (14)
  uschindler (12)
  dweiss (10)
  msokolov (10)

See all 64...
Commented
✓All
  asfimport (1368)
  github-actions[bot] (286)
  jpountz (137)
  mikemccand (125)
  rmuir (97)
  msokolov (86)
  benwtrent (83)

See all 317...
User
✓All
  asfimport (1810)
  github-actions[bot] (342)
  mikemccand (217)
  jpountz (208)
  rmuir (141)
  benwtrent (121)
  msokolov (107)

See all 481...
Last comment user
✓All
  asfimport (1319)
  github-actions[bot] (277)
  mikemccand (34)
  jpountz (31)
  rmuir (22)
  msokolov (17)
  benwtrent (17)

See all 168...
Draft
✓All
  No (309)
  Yes (64)
Component
✓All
  core (653)
  analysis (155)
  highlighter (46)
  facet (42)
  spatial (41)
  queryparser (30)
  sandbox (25)

See all 23...
Type
✓All
  enhancement (1208)
  bug (729)
  task (201)
  test (79)
  documentation (21)
Labels
✓All
  Stale (276)
  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 (2662)
Reporter
✓All
  rmuir (269)
  mikemccand (154)
  jpountz (124)
  dsmiley (67)
  uschindler (52)
  iverase (44)
  dweiss (42)

See all 770...
Assignee
✓All
  Unassigned (2419)
  mikemccand (35)
  uschindler (33)
  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

#16682 PR: Let the structure reading vectors say how they are read
5.4 hours ago  21 comments  0 votes  0 watches  ChrisHegarty, github-actions[bot], jimczi, thecoop
A format opening a file can attach hints to say what it holds and how it will be read, and a ... MMapDirectory turns them into madvise calls if an application asks it to. Today the vectors formats get two things wrong there.
    jimczi 5.4 hours ago:  Good catch, the old wording claimed more than we want. A file can be read over and over and still not be worth keeping in memory. What it's missing is reuse a cache can exploit, either because the reads land in no usable pattern ...
    jimczi 5.4 hours ago:  Done.

#16704 PR: Cover both early exit modes in TestHnswQueueSaturationCollector
5.6 hours ago  0 comments  0 votes  0 watches  github-actions[bot], salvatorecampagna
# TL;DR TestHnswQueueSaturationCollectortestEarlyExitRelation fails on rare seeds because it ... The test is split so each mode is covered deterministically. # Summary HnswQueueSaturationCollector ... That asymmetry is deliberate: AbstractKnnVectorQuery reads the relation to decide whether to fall ...

#16693 PR: Prevent unintended multi-segment flushes in TestLucene104ScalarQuantizedVectorFormat (#16692)
9.3 hours ago  18 comments  0 votes  0 watches  CH-Abhinav, dweiss, github-actions[bot], mccullocht
Resolves 16692 Description Follow-up to 16030: fixes test flakiness in TestLucene104ScalarQuantiz ... Under randomized test runs (e.g., seeds D9588E29B239F98E, ACA59AF04AC0B9BF, 29219F4CDF5F2122), ... IndexWriter randomly flushing multiple segments because of low randomized RAM buffers, breaking ...
    dweiss 9.3 hours ago:  Often it's hard to predict what a change will do with so many components - this is what this ... No worries.
    CH-Abhinav 11.3 hours ago:  Yeah let's wait for trevor . I guess this issue won't have happened if review was done correctly 😕 In the meantime should ... And should I close this PR?

#16526 PR: [New feature] Add an adaptive filtered strategy PathSeer for Lucene filtered HNSW search
13.7 hours ago  6 comments  0 votes  0 watches  benwtrent, github-actions[bot], jpountz, lizhiyuell, mikemccand, msokolov
Description This PR introduces PathSeer, an adaptive filtered HNSW search strategy, for ... . # Motivation Existing filtered HNSW strategies can perform very differently across workloads.
    lizhiyuell 13.9 hours ago:  > The idea makes sense. I don't have time for a deep review :(.
    benwtrent 7.4 days ago:  The idea makes sense. I don't have time for a deep review :(. I would prefer that this is an evolution of the HnswFiltered directly, instead of yet another way ...

#16656 PR: Parallel O_DIRECT rerank reads for larger-than-RAM KNN search
14.6 hours ago  3 comments  0 votes  0 watches  github-actions[bot], goankur, jimczi
Title: **(Implemented with AI, with Human in the loop)** Parallel O_DIRECT full-precision ... RescoreTopNQuery scores one document at a time, so a shortlist of M candidates becomes M dependent ... Batching per input is not enough either.
    jimczi 14.6 hours ago:  > I already built parts of it in my private branch so please take a look and advise the changes ... Yep that s exactly this. Currently the api is modelled after the HNSW use case that prefetches N neighbours at a time.
    goankur 21.7 hours ago:  Thanks jimczi - You were right, and the baseline was the problem. Measured: across a full 10,000-query run the JVM issues **23 madvise calls in total**, versus **~ ... Everything below is re-measured on that basis.

#16692: Flaky test: TestLucene104ScalarQuantizedVectorsFormat.testDataBlindNoRawFloatVectors fails when ...
14.9 hours ago  0 comments  0 votes  0 watches  CH-Abhinav, dweiss
Description TestLucene104ScalarQuantizedVectorsFormat.testDataBlindNoRawFloatVectors occasionally ... ContextIntroduced in 16030 ("Implement data-blind scalar quantization") where testDataBlindNoRawFloa ... `text java.lang.IllegalArgumentException: StandardDirectoryReader(segments_1:6 _0(11.0.0):c22 _1( ...

#16564 PR: Optimize automaton intersection when all remaining byte suffixes match.
20.6 hours ago  2 comments  0 votes  0 watches  github-actions[bot], mikemccand, vsop-479
This PR is a simplified version of 13072. When IntersectTermsEnum reaches an accepting automaton state that can match all remaining byte ... This PR only computes match-all-suffix states for the simple case where an accepting state has a ...
    github-actions[bot] 20.6 hours 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!
    vsop-479 28.4 days ago:  @mikemccand Please take a look when you get a chance. Performance on wikimediumall: ` TaskQPS baseline StdDevQPS my_modified_version StdDev ...

[16.1 msec search, 17.2 msec total]