odidev opened a new issue, #14006:
URL: https://github.com/apache/lucene/issues/14006
### Description
## Description
For aarch64 platform, I have built " Lucene " from source.
## Steps
Following
[STEPS](https://github.com/apache/lucene?tab=readme-ov-file#bu
mkhludnev commented on PR #14005:
URL: https://github.com/apache/lucene/pull/14005#issuecomment-2490162870
Hi @paulk-asert!
Thanks for your contribution.
Please create backport commit targeting
https://github.com/apache/lucene/tree/branch_10x
--
This is an automated message fro
mkhludnev merged PR #14005:
URL: https://github.com/apache/lucene/pull/14005
--
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
vigyasharma commented on code in PR #13951:
URL: https://github.com/apache/lucene/pull/13951#discussion_r1851338572
##
lucene/core/src/java/org/apache/lucene/index/IndexWriterRAMManager.java:
##
@@ -0,0 +1,217 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
paulk-asert opened a new pull request, #14005:
URL: https://github.com/apache/lucene/pull/14005
Fix typo
### Description
--
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 spe
jpountz commented on issue #14004:
URL: https://github.com/apache/lucene/issues/14004#issuecomment-2489618305
I confirmed that #266 seems to help for this case. It does find merges to
run with the above example (100kB, 300kB, 800kB, 2MB, 5MB, 12MB, 30MB, 70MB,
150MB, 400MB). And when I run
krickert commented on PR #13525:
URL: https://github.com/apache/lucene/pull/13525#issuecomment-2489546204
> Chunk-Based Highlighting – Interesting. With getAllVectorValues(), we can
find all vector values with similarity above a separate sim-threshold for
highlights?
Not sure. But i
shatejas commented on PR #13985:
URL: https://github.com/apache/lucene/pull/13985#issuecomment-2489459115
> The `org.apache.lucene.index.TestConcurrentMergeScheduler.testNoWaitClose`
test hits a new assert that I added - sorry. I need to look to see if it is a
test issue or more of a design
vigyasharma commented on PR #13525:
URL: https://github.com/apache/lucene/pull/13525#issuecomment-2489342443
Thank you for sharing these use-cases @krickert !
1. **Aggregate Scoring** – I think we can do this today by joining the child
doc hits with their parents and calculating score
jpountz commented on code in PR #266:
URL: https://github.com/apache/lucene/pull/266#discussion_r1850802831
##
lucene/core/src/java/org/apache/lucene/index/TieredMergePolicy.java:
##
@@ -495,36 +495,36 @@ private MergeSpecification doFindMerges(
for (int startIdx = 0; sta
jpountz commented on code in PR #266:
URL: https://github.com/apache/lucene/pull/266#discussion_r1850802366
##
lucene/core/src/java/org/apache/lucene/index/TieredMergePolicy.java:
##
@@ -495,36 +495,36 @@ private MergeSpecification doFindMerges(
for (int startIdx = 0; sta
jpountz commented on issue #14004:
URL: https://github.com/apache/lucene/issues/14004#issuecomment-2489139597
Interestingly, it looks like this PR
https://github.com/apache/lucene/pull/266 would do what I'm looking for, though
it aimed at solving a different problem (!)
--
This is an aut
jpountz opened a new issue, #14004:
URL: https://github.com/apache/lucene/issues/14004
### Description
I have been experimenting with configuring `TieredMergePolicy` to keep the
segment count very low:
- segsPerTier = 2
- floorSegmentSize = 512MB
This typically helps if
ChrisHegarty commented on PR #13998:
URL: https://github.com/apache/lucene/pull/13998#issuecomment-2489087426
> This works for me. Maybe implement this API on our in-memory index inputs
to return true, e.g. `ByteBuffersIndexInput`?
yeah, I think that this prob makes sense. Lemme satis
ChrisHegarty commented on code in PR #13998:
URL: https://github.com/apache/lucene/pull/13998#discussion_r1850657822
##
lucene/core/src/java/org/apache/lucene/store/IndexInput.java:
##
@@ -226,4 +227,17 @@ public String toString() {
* @param length the number of bytes to pre
ChrisHegarty commented on code in PR #13998:
URL: https://github.com/apache/lucene/pull/13998#discussion_r1850653616
##
lucene/core/src/java21/org/apache/lucene/store/MemorySegmentIndexInput.java:
##
@@ -406,6 +406,15 @@ void advise(long offset, long length,
IOConsumer advice)
ChrisHegarty commented on PR #13985:
URL: https://github.com/apache/lucene/pull/13985#issuecomment-2488990896
The `org.apache.lucene.index.TestConcurrentMergeScheduler.testNoWaitClose`
test hits a new assert that I added - sorry. I need to look to see if it is a
test issue or more of a desi
jpountz commented on PR #14003:
URL: https://github.com/apache/lucene/pull/14003#issuecomment-2488789866
The speedup is not as good, but still significant:
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct diff p-valu
jpountz opened a new pull request, #14003:
URL: https://github.com/apache/lucene/pull/14003
WANDScorer implements block-max WAND and needs to recompute score upper
bounds whenever it moves to a different block. Thus it's important for these
blocks to be large enough to avoid re-computing sc
benwtrent commented on issue #13564:
URL: https://github.com/apache/lucene/issues/13564#issuecomment-2488523095
@dungba88 I left a comment on your POC. It seems to me the best abstraction
is a new query that doesn't inherit from AbstractKnnVectorQuery. Instead its
just a new query that requ
krickert commented on PR #13525:
URL: https://github.com/apache/lucene/pull/13525#issuecomment-2488410934
> And we can use getAllVectorValues() for scoring with max or avg of all
vectors in the doc at query time.
Your proposal to implement `getAllVectorValues()` for scoring documents
jpountz commented on PR #14000:
URL: https://github.com/apache/lucene/pull/14000#issuecomment-2487985918
This changed top hits in nightly benchmarks, which caused a failure. I'm
reverting and will look into it.
--
This is an automated message from the Apache Git Service.
To respond to the
vigyasharma commented on PR #13525:
URL: https://github.com/apache/lucene/pull/13525#issuecomment-2487597269
_...contd. from above – thoughts on supporting independent multi-vectors
specified via `NONE` multi-vector aggregation..._
__
The `Knn{Float|Byte}Vector` fields will accept
23 matches
Mail list logo
