iverase commented on PR #12460:
URL: https://github.com/apache/lucene/pull/12460#issuecomment-1735063044
I have been thinking a bit longer about this and I think this approach of
`DataInput` is not right. Instead we should try to return an API more similar
to `RandomAccessInput` as Uwe sugg
gf2121 opened a new pull request, #12591:
URL: https://github.com/apache/lucene/pull/12591
Inspired by the #91, This PR proposes to use a stable radix sorter to sort
update terms instead of tie comparator to doc id. As terms are appended in
order, the latest update of each term value should
iverase opened a new issue, #12592:
URL: https://github.com/apache/lucene/issues/12592
It is currently not possible to read all bytes from a RandomAccessInput
without previous knowledge of how many bytes were written. I would like to
propose that RandomAccessInput can provide the user the i
jpountz opened a new pull request, #12593:
URL: https://github.com/apache/lucene/pull/12593
MaxScoreBulkScorer computes windows based on the set of clauses that were
essential in the *previous* window. This usually works well as the set of
essential clauses tends to be stable over time, but
jpountz commented on PR #12593:
URL: https://github.com/apache/lucene/pull/12593#issuecomment-1735586604
No impact on wikibigall, which is expected as all queries have sets of
essential/non-essential clauses which are stable over time.
```
TaskQPS baseli
jpountz commented on PR #12564:
URL: https://github.com/apache/lucene/pull/12564#issuecomment-1735610903
Closing: benchmarks on wikibigall don't suggest it helps more than #12382.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHu
jpountz closed pull request #12564: Window-at-a-time scoring for conjunctions.
URL: https://github.com/apache/lucene/pull/12564
--
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.
s1monw opened a new pull request, #12595:
URL: https://github.com/apache/lucene/pull/12595
`IndexWriter#flushNextBuffer()` is a convenient way to control indexing
buffer sizes across multiple index writers. This change also flushes deletes if
necessary when `#flushNextBuffer()` is called ev
heemin32 opened a new issue, #12596:
URL: https://github.com/apache/lucene/issues/12596
### Description
https://github.com/apache/lucene/blob/main/lucene/test-framework/src/java/org/apache/lucene/tests/geo/ShapeTestUtil.java
surpriseMePolygon and createRegularPolygon returns in
jpountz commented on PR #12595:
URL: https://github.com/apache/lucene/pull/12595#issuecomment-1736121993
Thinking out loud: it looks like your change always flushes both the largest
pending writer and deletes. I wonder if we should try to make it more granular
and e.g. check whichever of de
sgup432 opened a new issue, #12597:
URL: https://github.com/apache/lucene/issues/12597
### Description
As of now, lucene LRU query cache and other OpenSearch caches uses CacheKey
as a primary key for their caches as it helps to determine any changes during
segment merges etc. We use
shubhamvishu commented on issue #12394:
URL: https://github.com/apache/lucene/issues/12394#issuecomment-1736155862
@jpountz I have raised a PR #12548 that adds the required APIs to DVS for
computing vector similarity scores. Thanks!
--
This is an automated message from the Apache Git Serv
iverase commented on PR #12594:
URL: https://github.com/apache/lucene/pull/12594#issuecomment-1736205488
Thanks @jpountz! What it is not clear to me is if adding a method to this
interface is considered a breaking change and it can only be introduced in a
major release, or if it can be back
gsmiller commented on issue #12558:
URL: https://github.com/apache/lucene/issues/12558#issuecomment-1736274592
Oh interesting. Thanks @Shradha26! Is drill-sideways the only means for
reproducing this (that we're aware of anyway)? That's tricky. Do we think it's
doing the right thing, or sho
gf2121 commented on code in PR #12573:
URL: https://github.com/apache/lucene/pull/12573#discussion_r1338004334
##
lucene/core/src/java/org/apache/lucene/index/BufferedUpdates.java:
##
@@ -139,15 +131,11 @@ public void addTerm(Term term, int docIDUpto) {
return;
}
-
15 matches
Mail list logo
