vsop-479 commented on PR #11722:
URL: https://github.com/apache/lucene/pull/11722#issuecomment-1260490249
@jpountz Thranks for your review and suggestion. I have added a CHANGES
entry and assert term value code.
Please have a review.
--
This is an automated message from the Apache Git
uschindler commented on issue #11827:
URL: https://github.com/apache/lucene/issues/11827#issuecomment-1260501918
I fully agree, some checks should be done.
But here are a few bits that came into my mind:
- The @mikemccand benchmarks are running against main branch only. So the
first ch
jpountz commented on code in PR #11722:
URL: https://github.com/apache/lucene/pull/11722#discussion_r982050404
##
lucene/test-framework/src/java/org/apache/lucene/tests/index/BasePostingsFormatTestCase.java:
##
@@ -367,6 +367,53 @@ public void testGhosts() throws Exception {
Mahdi-Seeker commented on issue #10177:
URL: https://github.com/apache/lucene/issues/10177#issuecomment-1260514435
Hi guys
Thanks for your great job on Lucene and specially this ANN search!
Any progress on this issue? We're trying to use vector search, but HNSW
seems to take too mu
iverase commented on issue #11824:
URL: https://github.com/apache/lucene/issues/11824#issuecomment-1260667934
Fix seems to bring performance back to previous levels:
https://user-images.githubusercontent.com/29038686/192749125-0ac0b341-b2cb-4395-991c-9f676322592a.png";>
--
This
jpountz commented on code in PR #1039:
URL: https://github.com/apache/lucene/pull/1039#discussion_r982215123
##
lucene/core/src/test/org/apache/lucene/search/TestWANDScorer.java:
##
@@ -947,4 +988,82 @@ public long cost() {
};
}
}
+
+ private static class WANDSco
mikemccand commented on issue #11827:
URL: https://github.com/apache/lucene/issues/11827#issuecomment-1260778045
> yup. Possibly too if Mike M is bored he could implement an alarming system
:) or export the data somehow so we could bolt one on the side?
Actually I rather like the alar
mikemccand commented on issue #11824:
URL: https://github.com/apache/lucene/issues/11824#issuecomment-1260787630
Thanks for catching this @iverase and the quick fix, and the follow-on issue
to better detect such regressions before release: #11827
--
This is an automated message from the A
mikemccand commented on issue #11827:
URL: https://github.com/apache/lucene/issues/11827#issuecomment-1260788095
This was a spinoff from #11824.
--
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 th
jpountz opened a new issue, #11829:
URL: https://github.com/apache/lucene/issues/11829
### Description
The test seems to be creating invalid polygons.
```
03:40:22 org.apache.lucene.document.TestShapeDocValues >
testXYPolygonCentroid FAILED
03:40:22 WARNING: The Security
nknize commented on issue #11824:
URL: https://github.com/apache/lucene/issues/11824#issuecomment-1260938633
What's annoying is how incredibly trappy this override logic is. That a
method call literally moving from `createWeight` to `getScorerSupplier` results
in a 72.2% regression even sli
mikemccand commented on code in PR #11780:
URL: https://github.com/apache/lucene/pull/11780#discussion_r982553986
##
lucene/core/src/java/org/apache/lucene/search/ReferenceManager.java:
##
@@ -219,6 +219,36 @@ public final boolean maybeRefresh() throws IOException {
return
stefanvodita commented on code in PR #11780:
URL: https://github.com/apache/lucene/pull/11780#discussion_r982630324
##
lucene/core/src/java/org/apache/lucene/search/ReferenceManager.java:
##
@@ -219,6 +219,36 @@ public final boolean maybeRefresh() throws IOException {
retur
gsmiller commented on PR #11828:
URL: https://github.com/apache/lucene/pull/11828#issuecomment-1261195636
> I assume we already have tests that cover this case?
Good question. I'm going to go tweak our tests. We added tests that cover
the completely dense case (i.e., all docs in a seg
gsmiller commented on PR #11803:
URL: https://github.com/apache/lucene/pull/11803#issuecomment-1261198752
@zhaih I reexamined our test coverage and think we're in good shape already
actually. We've got good coverage for covering drill-sideways correctness with
multiple dimensions, etc. (inc
zhaih commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982692228
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
mdmarshmallow commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982759362
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundat
mdmarshmallow commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982764818
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundat
zhaih commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982825433
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
zhaih commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982825989
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
zhaih commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982826316
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
gsmiller commented on PR #11803:
URL: https://github.com/apache/lucene/pull/11803#issuecomment-1261458849
@zhaih that's a good point and valid concern. I dug into the existing tests
and it looks like we have lots of coverage _except_ that the majority of the
coverage is using basic, single-
mdmarshmallow commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982891007
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundat
uschindler commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982891128
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation
mdmarshmallow commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982902771
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundat
mdmarshmallow commented on PR #11796:
URL: https://github.com/apache/lucene/pull/11796#issuecomment-1261519604
@jpountz , in response to this:
> I'm considering exposing write amplification separately for flushes (as
flushedBytes / totalIndexSize), merges (as (totalIndexSize + mergedB
uschindler commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982909584
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation
uschindler commented on issue #11761:
URL: https://github.com/apache/lucene/issues/11761#issuecomment-1261531792
I was also doing consulting for an huge Elasticsearch user and they also had
this problem of keeping deletes as low as possible and the 20% limit was way
too high. 20% looks like
zhaih commented on code in PR #11796:
URL: https://github.com/apache/lucene/pull/11796#discussion_r982921333
##
lucene/misc/src/java/org/apache/lucene/misc/store/ByteWritesTrackingDirectoryWrapper.java:
##
@@ -0,0 +1,118 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
zhaih commented on PR #11803:
URL: https://github.com/apache/lucene/pull/11803#issuecomment-1261541919
@gsmiller Thank you for checking and continuous effort!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL abo
gsmiller commented on PR #11803:
URL: https://github.com/apache/lucene/pull/11803#issuecomment-1261579341
@zhaih well, thank you for keeping me honest with testing. I think I've
already found an insidious, potential bug with some beefier tests.
--
This is an automated message from the Ap
jtibshirani opened a new issue, #11830:
URL: https://github.com/apache/lucene/issues/11830
### Description
HNSW search is most efficient when all vector data fits in page cache. So
good to keep the size of vector files as small as possible.
We currently write all HNSW graph con
32 matches
Mail list logo
