navneet1v opened a new pull request, #14223:
URL: https://github.com/apache/lucene/pull/14223
### Description
Fixed a flaky test TestKnnFloatVectorQuery.testFindFewer
This PR solves the issue: https://github.com/apache/lucene/issues/14175
Details on why the test was faili
stefanvodita commented on issue #14222:
URL: https://github.com/apache/lucene/issues/14222#issuecomment-2653320901
Sounds interesting, keen to see if we measure a performance improvement!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
iverase commented on code in PR #14213:
URL: https://github.com/apache/lucene/pull/14213#discussion_r1952672292
##
lucene/core/src/java/org/apache/lucene/index/StoredFieldVisitor.java:
##
@@ -41,15 +40,17 @@ public abstract class StoredFieldVisitor {
protected StoredFieldVisi
gf2121 commented on PR #14203:
URL: https://github.com/apache/lucene/pull/14203#issuecomment-2653948712
I refactor code to inner-loop. Result on wikimediumall AVX512
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct d
benwtrent commented on PR #14160:
URL: https://github.com/apache/lucene/pull/14160#issuecomment-2653702500
Barring any further discussion, I am gonna merge this soon. While the
filtering search still isn't perfect, this is a marked improvement.
--
This is an automated message from the Apa
iverase commented on code in PR #14213:
URL: https://github.com/apache/lucene/pull/14213#discussion_r1952663407
##
lucene/backward-codecs/src/java/org/apache/lucene/backward_codecs/lucene50/compressing/Lucene50CompressingStoredFieldsReader.java:
##
@@ -25,12 +25,7 @@
import org
iverase commented on PR #14213:
URL: https://github.com/apache/lucene/pull/14213#issuecomment-2653747350
+1 I like the idea of encapsulating the DataInput and length inside
`StoredFieldDataInput`, and yes, it gives symmetry between the read and the
write side.
--
This is an automated mes
gf2121 commented on PR #14203:
URL: https://github.com/apache/lucene/pull/14203#issuecomment-2653001052
Inner loop performance get better on the newest commit.
```
Mac M2
Benchmark(bpv) (countVariable) Mode CntScore
Error Units
BKDCodec
msfroh opened a new pull request, #14228:
URL: https://github.com/apache/lucene/pull/14228
### Description
On https://github.com/apache/lucene/pull/14221, I foolishly asserted that
the error-prone `NonFinalStaticField` checks would be "a lot less noisy". It
seems that PR and its mere
navneet1v commented on PR #14223:
URL: https://github.com/apache/lucene/pull/14223#issuecomment-2655527582
An unrelated test to this change is failing.
```
TestForTooMuchCloning > test FAILED
java.lang.AssertionError: too many calls to IndexInput.clone during
merging: 523
carlosdelest opened a new issue, #14225:
URL: https://github.com/apache/lucene/issues/14225
### Description
Merges should expose an estimation of the amount of heap memory that will be
used for the actual merge to be done.
This information should be exposed as part of the merge
rmuir commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654392511
I can reproduce it, thanks! My concern is around the `[]`, this is an empty
"character class", maybe unlucky result from the RNG? Otherwise when 'empty'
logic is wrong, usually you kn
benwtrent commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654393417
Thanks @rmuir !
I really don't know if it's a test data failure or something interesting
actually broke.
--
This is an automated message from the Apache Git Service.
To r
rmuir commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654396301
I see `[]` in the regexp which matches no possible characters, so it creates
Automata.makeEmpty(). Concatenation of empty with anything else returns empty,
so that's why the test fail
HoustonPutman commented on code in PR #13914:
URL: https://github.com/apache/lucene/pull/13914#discussion_r1953282099
##
lucene/facet/src/java/org/apache/lucene/facet/range/DynamicRangeUtil.java:
##
@@ -202,66 +208,83 @@ public SegmentOutput(int hitsLength) {
* is used t
benwtrent commented on PR #14226:
URL: https://github.com/apache/lucene/pull/14226#issuecomment-2654818796
This is indeed interesting. I eagerly wait some of the perf numbers ;).
But, this re-entering the graph makes me think that collectors will need to
track visitation. This way the
msokolov commented on PR #14226:
URL: https://github.com/apache/lucene/pull/14226#issuecomment-2654820602
Hmm github's test run failed with:
```
gradlew :lucene:core:test --tests
"org.apache.lucene.search.TestByteVectorSimilarityQuery.testFallbackToExact"
-Ptests.jvms=1 -Ptests.jv
msokolov commented on PR #14226:
URL: https://github.com/apache/lucene/pull/14226#issuecomment-2654826667
> But, this re-entering the graph makes me think that collectors will need
to track visitation. This way the same vector path isn't visited multiple
times. Its possible that once you en
rmuir commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654406157
i think a test against old parser should show it. maybe it indirectly did
the `peek()`s and `match()`s differently in the case of no content within a
`[]`, and didnt try to parse it
HoustonPutman commented on code in PR #13914:
URL: https://github.com/apache/lucene/pull/13914#discussion_r1953123437
##
lucene/facet/src/java/org/apache/lucene/facet/range/DynamicRangeUtil.java:
##
@@ -202,66 +208,83 @@ public SegmentOutput(int hitsLength) {
* is used t
benwtrent opened a new issue, #14224:
URL: https://github.com/apache/lucene/issues/14224
### Description
I bet the failure is due to the recent development around `Automaton`
stuffs.
Gitbisect led me to fe42efc5918
```
TestOperations > testGetRandomAcceptedString FAI
benwtrent commented on issue #14214:
URL: https://github.com/apache/lucene/issues/14214#issuecomment-2654755040
So, verifying the "fewDistinct" slowness, here is how connect components
works in this adverse case:
```
1> HNSW 1 [2025-02-12T20:14:45.641640Z;
TEST-TestKnnFloatVector
msokolov opened a new pull request, #14226:
URL: https://github.com/apache/lucene/pull/14226
### Description
This is a WIP patch to work out an idea for Knn hit collection that is
deterministic and efficient in the sense that the number of hits collected per
leaf scales with the size
rmuir commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2655012824
This one was simpler, it is just a case where you should be getting a
parsing error: https://github.com/apache/lucene/pull/14227
--
This is an automated message from the Apache Git
rmuir opened a new pull request, #14227:
URL: https://github.com/apache/lucene/pull/14227
string: `?+½]+]+Ř*+[\]ᖴﴁ.`
expected: before #14193
```
java.lang.IllegalArgumentException: expected ']' at position 17
```
actual: after #14193
```
REGEXP_CONCATENATION
benwtrent commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654345727
Here is the regex string that creates the empty operations:
```
regex str: (\?){1,}\½(\]){1,}\(\]){1,}((\Ř)*){1,}[]"ᖴﴁ".
```
--
This is an automated message from t
dweiss merged PR #14216:
URL: https://github.com/apache/lucene/pull/14216
--
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
dweiss merged PR #14221:
URL: https://github.com/apache/lucene/pull/14221
--
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
HoustonPutman commented on code in PR #13914:
URL: https://github.com/apache/lucene/pull/13914#discussion_r1953286478
##
lucene/facet/src/java/org/apache/lucene/facet/range/DynamicRangeUtil.java:
##
@@ -202,66 +208,83 @@ public SegmentOutput(int hitsLength) {
* is used t
rmuir commented on issue #14224:
URL: https://github.com/apache/lucene/issues/14224#issuecomment-2654370971
@benwtrent I can look tonight
--
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 speci
msfroh commented on PR #14221:
URL: https://github.com/apache/lucene/pull/14221#issuecomment-2654462496
> #14216 may help with this specifically `NonFinalStaticField` which was off
since it was super noisy. https://errorprone.info/bugpattern/NonFinalStaticField
Once #14216 and this PR
dweiss commented on PR #14221:
URL: https://github.com/apache/lucene/pull/14221#issuecomment-2654625492
Done. Thank you!
--
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
cgejian commented on PR #14163:
URL: https://github.com/apache/lucene/pull/14163#issuecomment-2655275009
@mikemccand
This is my first time submitting a PR to the Lucene project, Could you
please tell me if this PR can be merged? Looking forward to your reply, thank
you.
--
This is an
33 matches
Mail list logo
