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

#16029: Data-blind scalar quantization

26.3 minutes ago 10 comments 0 votes 0 watches Pranshu-S, kaivalnp, mccullocht, mikemccand, msokolov, romseygeek, shubhamvishu, xande

Description Add options for data-blind scalar quantization. Data blind quantization will allow users to drop the raw float vectors which will be a significant ... There are 3 parts to this change: - [x] Allow users to disable centering and drop the raw float ...

Pranshu-S 26.4 minutes ago:	Hey @mccullocht , curious as to why this was reverted (https://github.com/apache/lucene/commit/ ... Do we have plans of adding it back soon? I was coming up with changes to introduce quantise residuals on the data blind path.
mccullocht 1.7 months ago:	I merged the data blind changes (thanks @kaivalnp) with some expanded scope: you can choose whether ... We disallow dropping the float vectors for asymmetric quantization schemes (1 and 2 bit docs) ... The next step here is to optionally quantize residuals.

#16868 PR: Use the single-valued leaf cutter for single-valued segments in LongRangeFacetCutter

33.7 minutes ago 8 comments 0 votes 0 watches epotyom, github-actions[bot], slow-J

If DocValues.unwrapSingleton returns a segment's values, use the single-valued leaf cutter (instead ... Follow-up to 16268. It was originally part of that, but was spun off as a follow-up.

github-actions[bot] 7.5 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!
slow-J 22.9 days ago:	Thanks for the review Egor!

#16691 PR: Optimize DocValues bulk decoding using Panama widening casts (#16677)

2.1 hours ago 11 comments 0 votes 0 watches CH-Abhinav, costin, github-actions[bot], rmuir

Description This PR introduces SIMD bulk decoding for DocValues (8-bit, 16-bit, and 32-bit ... Currently, default scalar loops handle these widening. This PR tries to use Panama Vector API's shape-widening operations (convertShape with ...

CH-Abhinav 2.1 hours ago:	Hi @costin saw your previously worked on this files recently. Can you see this PR once and review it whenever your free. It is relatively small PR.
CH-Abhinav 1.8 days ago:	Hi @rmuir it would be great if you can review this PR since you started reviewing it. Suggest changes if any.We can try to integrate this in Lucene 11. Thanks :)

#16843 PR: Adding optimal Binary Fuse Filter to replace Bloom Filter

2.1 hours ago 2 comments 0 votes 0 watches CH-Abhinav, gsmiller, mikemccand

DescriptionCloses 16780. This PR implements native, zero-dependency **Binary Fuse Filters** (Graf & Lemire, 2022) in Lucene, ... Binary Fuse filters are an optimal fit for Apache Lucene because Lucene segments are immutable ...

CH-Abhinav 2.1 hours ago:	Hey @mikemccand @gsmiller can you guys review the PR once whenever your free. Lemire (one of the authors of Binary fuse) also helped a bit regarding this PR. Any suggestions or feedbacks I will fix them.
CH-Abhinav 20.9 days ago:	@mikemccand can you review the PR and suggest feedbacks whenever you have time? I think this will sit perfectly to lucene's architecture since we have immutable segments.

#17020 PR: Reduce CachingTokenFilter memory by not copying stale bytes

2.3 hours ago 0 comments 0 votes 0 watches github-actions[bot], reugn

CharTermAttributeImplclone copies builder, which holds the bytes of the last term passed to ... That term can be much longer than the current one, and can come from an earlier analysis that ... CachingTokenFilter calls captureState() for every token, so each cached token keeps its own copy of ...

#17001 PR: Determinize in SimplePatternTokenizer String constructors

7.8 hours ago 1 comments 0 votes 0 watches github-actions[bot], john-wagster, rmuir

Fixes 16999. Since RegExptoAutomaton() no longer determinizes, the String constructors of SimplePatternTokenizer ... This determinizes with determinizeWorkLimit, mirroring SimplePatternSplitTokenizer.

rmuir 7.8 hours ago:

personally I think this heads in a different direction than where we've been going with fixing this ... We've been trying to remove determinize() calls in lucene, which i think is the right way to go. For this one, IMO ideally: * Fix any issues so the logic works correctly with NFAs as well as DFAs

#16998 PR: GITHUB#16997 Negative Query Optimizations for Lucene Monitor Presearcher

9.5 hours ago 21 comments 0 votes 0 watches github-actions[bot], kotman12, romseygeek

Description Hey all, I'm a Solr committer and also a frequent user of the lucene-monitor module ... I've had this idea for a while which was greatly inspired by the very clever way that ... The same approach can actually be extended to more optimally handle negative queries.

kotman12 9.5 hours ago:	> Thanks for opening @kotman12, this looks very cool. It is obviously a lot to review at once so I will take some time over it, if that's OK. Do you have any numbers on how it performs for your usecases?
kotman12 10.2 hours ago:	A propos the dual-interface attribute idea, looks like ConcatenateGraphFilter does something _ ...

«
1
2
3
4
5
»

2,737 issues

[21.8 msec search, 22.8 msec total]