sam-herman commented on issue #12615:
URL: https://github.com/apache/lucene/issues/12615#issuecomment-2611333530
I am actually in the process of extending Lucene Codec for JVector DiskANN
integration.
Note this work is part of
https://github.com/opensearch-project/k-NN/issues/2386 I can
jhinch-at-atlassian-com commented on issue #14161:
URL: https://github.com/apache/lucene/issues/14161#issuecomment-2611152321
Having `BytesRefBuilder#toString` delegate to its underlying buffer or
calling `BytesRefBuilder#get` within `PointInSetQuery` both seem like
reasonable options.
--
mayya-sharipova commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610983018
@benwtrent Thanks for raising this, this indeed happens because of
MultiLeafKnnCollector and search threads exchanging info of the globally
collected results. Because it is not d
aliciavargas opened a new issue, #14168:
URL: https://github.com/apache/lucene/issues/14168
### Description
The Lucene docs specify that wildcard search is only supported for single
terms but not phrases
([link](https://lucene.apache.org/core/8_10_0/queryparser/org/apache/lucene/quer
benwtrent commented on code in PR #14160:
URL: https://github.com/apache/lucene/pull/14160#discussion_r1927632268
##
lucene/core/src/java/org/apache/lucene/util/hnsw/FilteredHnswGraphSearcher.java:
##
@@ -0,0 +1,332 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) und
benchaplin commented on code in PR #14160:
URL: https://github.com/apache/lucene/pull/14160#discussion_r1927620854
##
lucene/core/src/java/org/apache/lucene/util/hnsw/FilteredHnswGraphSearcher.java:
##
@@ -0,0 +1,332 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) un
benchaplin commented on code in PR #14160:
URL: https://github.com/apache/lucene/pull/14160#discussion_r1927609309
##
lucene/core/src/java/org/apache/lucene/util/hnsw/FilteredHnswGraphSearcher.java:
##
@@ -0,0 +1,332 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) un
benchaplin commented on code in PR #14160:
URL: https://github.com/apache/lucene/pull/14160#discussion_r1927609309
##
lucene/core/src/java/org/apache/lucene/util/hnsw/FilteredHnswGraphSearcher.java:
##
@@ -0,0 +1,332 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) un
benwtrent commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610476022
OK, if I change to never use `MultiLeafKnnCollector`, the multi-threaded
consistency test passes. But with using that collector, it will fail a couple
times over 10k+ repeats.
--
Th
mayya-sharipova commented on PR #14154:
URL: https://github.com/apache/lucene/pull/14154#issuecomment-2610406769
@jpountz @benwtrent I've addressed your comments in the last commit, please
continue to review
--
This is an automated message from the Apache Git Service.
To respond to the me
benwtrent commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610393806
OK, I cleaned it all up, and have two separate tests, one for multi-threaded
one for single threaded.
The multi-threaded one is the only one that fails periodically, which
expla
javanna commented on issue #13499:
URL: https://github.com/apache/lucene/issues/13499#issuecomment-2610294122
@parastooGit you need to create collector managers instead of collectors.
There is not static create method any longer, you need to create the collector
managers using their constru
gsmiller commented on issue #14161:
URL: https://github.com/apache/lucene/issues/14161#issuecomment-2610240868
Oh gross. Good catch! It seems like the desire in this exception message is
to print out the `previous` bytes ref in the same was as `current`. I wonder if
we should implement `Byt
benwtrent commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610150959
> I think our comments relate to the observation that the test does not
reproducibly fail with the same seed
🤦 for sure. Let me see if I can shore it up.
--
This is an automat
msokolov commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610145295
I think our comments relate to the observation that the test does not
reproducibly fail with the same seed
--
This is an automated message from the Apache Git Service.
To respond to t
benwtrent commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610139152
@msokolov @mikemccand maybe the consistency I am testing isn't clear.
First: Index a bunch of vectors
Second: do a single query on a static index to get the top-k
Repeat-N:
msokolov commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610126329
As for the reproducibility problem, that may be caused by concurrent HNSW
merging, which is nondeterministic.
--
This is an automated message from the Apache Git Service.
To respond
mikemccand commented on PR #14167:
URL: https://github.com/apache/lucene/pull/14167#issuecomment-2610089955
> Frustratingly, the seeded failures do not seem to be repeatable.
Hmm that is bad ... it means there is a test bug or test infra bug (separate
from the scary bug this test is
benwtrent opened a new pull request, #14167:
URL: https://github.com/apache/lucene/pull/14167
Inspired by some weird behavior I have seen, adding a consistency test.
I found that indeed, this fails over some seeds.
Frustratingly, the seeded failures do not seem to be repeatable
parastooGit commented on issue #13499:
URL: https://github.com/apache/lucene/issues/13499#issuecomment-2610054733
what is the replacement for create method?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above
mayya-sharipova commented on PR #14154:
URL: https://github.com/apache/lucene/pull/14154#issuecomment-2610017660
@benwtrent Thanks for the review, I am not happy with the design either,
will see how I can incorporate your feedback.
> I don't like that CompletionAnalyzer needs to trac
jpountz opened a new pull request, #14165:
URL: https://github.com/apache/lucene/pull/14165
`maxMergeAtOnce` increases merge amplification by running multiple merges
when it could run a single merge, without giving significant benefits in
exchange. We removed this parameter for forced merge
jpountz commented on code in PR #14164:
URL: https://github.com/apache/lucene/pull/14164#discussion_r1926977463
##
lucene/core/src/java/org/apache/lucene/codecs/lucene101/Lucene101PostingsReader.java:
##
@@ -388,6 +388,7 @@ private enum DeltaEncoding {
final boolean needsOf
jpountz merged PR #14155:
URL: https://github.com/apache/lucene/pull/14155
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apa
jpountz commented on PR #14163:
URL: https://github.com/apache/lucene/pull/14163#issuecomment-2609834797
I don't think we should merge this change, but it's good that you were able
to use it to confirm that merging would reclaim these deleted docs.
Can you add your data about this iss
RS146BIJAY commented on issue #13387:
URL: https://github.com/apache/lucene/issues/13387#issuecomment-2609475069
Make sense. I think we can extend MultiReader functionality to use it as a
combined view if we can support couple of read side features of IndexWriter
like opening a reader from
gf2121 opened a new pull request, #14164:
URL: https://github.com/apache/lucene/pull/14164
The `docBufferUpTo` variable is mainly maintained to obtain the
corresponding value of freq/pos buffer. We can avoid the maintaining when only
docs needed.
Result on `wikimediumall`:
```
27 matches
Mail list logo
