mkhludnev commented on issue #14260:
URL: https://github.com/apache/lucene/issues/14260#issuecomment-2673974937
main passed as well
https://ci-builds.apache.org/job/Lucene/job/Lucene-NightlyTests-main/1579/changes
--
This is an automated message from the Apache Git Service.
To respond to
javanna opened a new pull request, #14270:
URL: https://github.com/apache/lucene/pull/14270
We currently have a gap in testing the completion postings format. While it
allows to load FSTs off heap, and the format has a constructor that takes the
fst load mode, at read time SPI goes through
javanna commented on code in PR #14270:
URL: https://github.com/apache/lucene/pull/14270#discussion_r1965557931
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/Completion101PostingsFormat.java:
##
@@ -25,17 +25,10 @@
* @lucene.experimental
*/
public cla
gf2121 merged PR #14269:
URL: https://github.com/apache/lucene/pull/14269
--
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.apac
javanna commented on code in PR #14270:
URL: https://github.com/apache/lucene/pull/14270#discussion_r1965557931
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/Completion101PostingsFormat.java:
##
@@ -25,17 +25,10 @@
* @lucene.experimental
*/
public cla
javanna commented on code in PR #14270:
URL: https://github.com/apache/lucene/pull/14270#discussion_r1965567430
##
lucene/suggest/src/test/org/apache/lucene/search/suggest/document/TestSuggestField.java:
##
@@ -949,10 +952,7 @@ static IndexWriterConfig iwcWithSuggestField(Analyz
javanna commented on code in PR #14270:
URL: https://github.com/apache/lucene/pull/14270#discussion_r1965564752
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/CompletionPostingsFormat.java:
##
@@ -104,6 +104,9 @@ public abstract class CompletionPostingsForm
javanna commented on code in PR #14270:
URL: https://github.com/apache/lucene/pull/14270#discussion_r1965565789
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/CompletionPostingsFormat.java:
##
@@ -122,18 +125,13 @@ public enum FSTLoadMode {
private fin
gf2121 opened a new pull request, #14272:
URL: https://github.com/apache/lucene/pull/14272
Use`DocIdSetIterator#range` instead of `DocBaseBitsetIterator` as
`DocIdSetIterator#range` implemented `#intoBitset`.
--
This is an automated message from the Apache Git Service.
To respond to the m
jpountz opened a new pull request, #14273:
URL: https://github.com/apache/lucene/pull/14273
This attempts to generalize the `IndexSearcher#count` optimization from PR
#12415 to histogram facets by introducing specialization for counting the
number of matching docs in a range of doc IDs.
sgup432 commented on issue #14222:
URL: https://github.com/apache/lucene/issues/14222#issuecomment-2675501604
I re-ran test with 1mb cache, and assuming 50 lucene segments
Numbers are even better!
## Performance Comparison: v1 vs v2
| Benchmark
benwtrent commented on code in PR #14274:
URL: https://github.com/apache/lucene/pull/14274#discussion_r1966141126
##
lucene/core/src/test/org/apache/lucene/search/BaseKnnVectorQueryTestCase.java:
##
@@ -31,23 +31,7 @@
import org.apache.lucene.document.IntPoint;
import org.apac
benwtrent commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2675053049
> Sorry I didn't understand this.
@msokolov I mean that during search, we would by default have more neighbors
to look at. Ensuring diversity eagerly means that its possible
carlosdelest opened a new pull request, #14274:
URL: https://github.com/apache/lucene/pull/14274
When doing approximate knn search, it's possible that the approximate search
returns less than k results. In case there is a filter, we know the filter cost
so we can check if there are actually
javanna commented on code in PR #14271:
URL: https://github.com/apache/lucene/pull/14271#discussion_r1965909207
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/CompletionsTermsReader.java:
##
@@ -72,10 +72,8 @@ public final class CompletionsTermsReader imple
Vikasht34 commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2675065270
@benwtrent As far as I understand from your idea is to use Delaunay
Triangulation and skip connectComponents() ?
--
This is an automated message from the Apache Git Service.
To
javanna merged PR #14271:
URL: https://github.com/apache/lucene/pull/14271
--
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
javanna commented on PR #14271:
URL: https://github.com/apache/lucene/pull/14271#issuecomment-2675212515
Thanks all!
--
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 unsu
msokolov commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2675147250
> I mean that during search, we would by default have more neighbors to look
at.
I see, yes, that's true.
--
This is an automated message from the Apache Git Service.
To
ChrisHegarty opened a new pull request, #14275:
URL: https://github.com/apache/lucene/pull/14275
This commit adds support load per-iteration replacement of NamedSPI. The
primary motivation for this change is to support deterministic SPI loading when
deploying Lucene as a module.
Whe
gf2121 opened a new pull request, #14269:
URL: https://github.com/apache/lucene/pull/14269
Implement `#intoBitset` for `DocIdSetIterator#all` and
`DocIdSetIterator#range`.
For reference, `DocIdSetIterator#all` used in queries hit all docs.
`DocIdSetIterator#range` used in docvalue qu
msokolov commented on code in PR #14226:
URL: https://github.com/apache/lucene/pull/14226#discussion_r1965419004
##
lucene/core/src/java/org/apache/lucene/search/OptimisticKnnVectorQuery.java:
##
@@ -0,0 +1,205 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under on
msokolov commented on issue #14266:
URL: https://github.com/apache/lucene/issues/14266#issuecomment-2674480203
These random test failures in an approximate world are hard. There are no
guarantees! In similar cases in the past I have handcrafted data to exercise
the case in question: ie get
benwtrent merged PR #14265:
URL: https://github.com/apache/lucene/pull/14265
--
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.a
msokolov commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2674487026
In the original diversity impl we had allowed the neighbors array to fill
without regard to any diversity criterion, and only started imposing it once
the array was full. This mea
benwtrent commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2674533124
> This means we have to sort at that point IIRC, but it might be a better,
more robust choice?
I am not sure if its "better" I would assume it makes search slower on well
d
msokolov commented on code in PR #14226:
URL: https://github.com/apache/lucene/pull/14226#discussion_r1965415548
##
lucene/core/src/java/org/apache/lucene/search/OptimisticKnnVectorQuery.java:
##
@@ -0,0 +1,205 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under on
jpountz commented on PR #13470:
URL: https://github.com/apache/lucene/pull/13470#issuecomment-2674572334
I plan on merging this PR soon if there are no objections.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
UR
msokolov commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2674568528
> I am not sure if its "better" I would assume it makes search slower on
well distributed and distinguished vectors :/
Sorry I didn't understand this. In the normal case, we'
benwtrent commented on issue #14266:
URL: https://github.com/apache/lucene/issues/14266#issuecomment-2674536319
> These random test failures in an approximate world are hard. There are no
guarantees!
No doubt!
I was super surprised by this failure 😂 . The probabilities are ast
jpountz commented on code in PR #14271:
URL: https://github.com/apache/lucene/pull/14271#discussion_r1965729900
##
lucene/suggest/src/java/org/apache/lucene/search/suggest/document/CompletionsTermsReader.java:
##
@@ -72,10 +72,8 @@ public final class CompletionsTermsReader imple
jpountz commented on PR #14273:
URL: https://github.com/apache/lucene/pull/14273#issuecomment-2674888106
@epotyom You may be interested in taking a look.
--
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
jpountz merged PR #14204:
URL: https://github.com/apache/lucene/pull/14204
--
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
gf2121 commented on code in PR #14269:
URL: https://github.com/apache/lucene/pull/14269#discussion_r1965325680
##
lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java:
##
@@ -87,6 +87,18 @@ public int advance(int target) throws IOException {
public long cos
gf2121 commented on code in PR #14269:
URL: https://github.com/apache/lucene/pull/14269#discussion_r1965326705
##
lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java:
##
@@ -224,7 +248,7 @@ protected final int slowAdvance(int target) throws
IOException {
*
jpountz commented on code in PR #14269:
URL: https://github.com/apache/lucene/pull/14269#discussion_r1965299369
##
lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java:
##
@@ -224,7 +248,7 @@ protected final int slowAdvance(int target) throws
IOException {
*
36 matches
Mail list logo
