Stefan Vodita

Support deletions in rearrange (#11814)

4

### Description Changes made to support deletes in `IndexRearranger`: 1. Add `isDeleted(LeafReader reader, int idx)` in document selectors, which returns true if the document found at index `idx` in the...

GH#11601: Add ability to compute reader states after refresh

5

This PR is only meant as a starting point for a conversation and not to be merged as is. It is a proposal to let users retrieve new `SortedSedDocValuesReaderState` objects...

### Description `IndexRearranger` applies deletes found in the original index, so the rearranged index will not contain any deleted docs. While deleted docs don't show up in search results, they...

Demonstrate ordinal map building on refresh

1

We're considering [building ordinal maps eagerly](https://github.com/apache/lucene/issues/11601). This can already be done to some extent using the current API. This PR shows how that should be done. I'm don't know if...

Reduce duplication in taxonomy facets; always do counts

10

### Note This is a large change, refactoring most of the taxonomy facets code and changing internal behavior, without changing the API. There are specific API changes this sets us...

TestFSTs.testRandomWords reproducible NPE

2

### Description `./gradlew :lucene:core:test --tests "org.apache.lucene.util.fst.TestFSTs.testRandomWords" -Ptests.seed=5E5E16BC74CD0647` ``` org.apache.lucene.util.fst.TestFSTs > testRandomWords FAILED java.lang.NullPointerException: Cannot invoke "org.apache.lucene.util.fst.FST.getMetadata()" because "fst" is null at __randomizedtesting.SeedInfo.seed([5E5E16BC74CD0647:20CFAD83BDC2A982]:0) at org.apache.lucene.tests.util.fst.FSTTester.doTest(FSTTester.java:291) at org.apache.lucene.util.fst.TestFSTs.doTest(TestFSTs.java:217) at org.apache.lucene.util.fst.TestFSTs.testRandomWords(TestFSTs.java:306) at org.apache.lucene.util.fst.TestFSTs.testRandomWords(TestFSTs.java:281)...

Stop double-checking priority queue inserts

2

### Description There are several occurrences in our code where we check that an incoming value can be inserted in a priority queue and then call `insertWithOverflow`, which re-does the...

### Description #12337 introduced the possibility of indexing data about facet labels directly into the taxonomy. We should write a [facets demo](https://github.com/apache/lucene/tree/main/lucene/demo/src/java/org/apache/lucene/demo/facet) for this functionality and mention it in the...

Compute multiple float aggregations in one go

6

Usually facets maintain a one-dimensional array indexed by ordinal which keeps the values they're supposed to compute. The change here is simple in principle - use a two-dimensional array, indexed...

Update int array growth calls

2

`LSBRadicSorter.sort` doesn't need the buffer to preserve the data that was written to it for a previous sort. `TaskSequence` doesn't need to grow arrays beyond the number of iterations it...