Allowing users to mutate MultiTermQuery can give rise to odd bugs, for example
in wrapper queries such as BooleanQuery which lazily calculate their hashcodes
and then cache the result. This commit deprecates the setRewriteMethod()
method on MultiTermQuery, in preparation for removing it entirely, and adds
constructor parameters to the various MTQ implementations as a preferred
way to set the rewrite method.
This computes a pop count on a sample of the longs that back the bitset.
Quick benchmarks suggest that this runs 5x-10x faster than
`FixedBitSet#cardinality` depending on the length of the bitset.
In the effort or replacing usages of IndexSearcher#search(Query, Collector) with IndexSearcher#search(Query, CollectorManager), this commit replaces many test usages of TopScoreDocCollector with its corresponding CollectorManager created by calling TopScoreDocCollector#createSharedManager.
`KnnVectorQuery` currently uses the index reader's hashcode to make sure that
the query it builds runs on the right reader. We had added
`IndexContextReader#id` a while back for a similar purpose with `TermStates`,
let's reuse it?
Since doc IDs with a vector are loaded as an int[] in memory, this changes the
on-disk format of vectors to align with the in-memory representation by using
ints instead of vints to represent doc IDs. This might make vectors a bit
larger on disk, but also a bit faster to open.
I made the same change to how we encode nodes on levels for the same reason.
The original PR that added kNN filtering support overlooked non-default codecs.
This follow-up ensures that other codecs work with the new filtering logic:
* Make sure to check the visited nodes limit in `SimpleTextKnnVectorsReader`
and `Lucene90HnswVectorsReader`
* Add a test `BaseKnnVectorsFormatTestCase` to cover this case
* Fix failures in `TestKnnVectorQuery#testRandomWithFilter`, whose assumptions
don't hold when SimpleText is used
This PR also clarifies the limit checking logic for
`Lucene91HnswVectorsReader`. Now we always check the limit before visiting a
new node, whereas before we only checked it in an outer loop.
This PR adds support for a query filter in KnnVectorQuery. First, we gather the
query results for each leaf as a bit set. Then the HNSW search skips over the
non-matching documents (using the same approach as for live docs). To prevent
HNSW search from visiting too many documents when the filter is very selective,
we short-circuit if HNSW has already visited more than the number of documents
that match the filter, and execute an exact search instead. This bounds the
number of visited documents at roughly 2x the cost of just running the exact
filter, while in most cases HNSW completes successfully and does a lot better.
Co-authored-by: Joel Bernstein <jbernste@apache.org>
Since all documents are required to use the same features (LUCENE-9334) we can
rewrite DocValuesFieldExistsQuery to a MatchAllDocsQuery whenever terms or
points have a docCount that is equal to maxDoc.
Better encoding of doc Ids in Lucene91HnswVectorsFormat
for a dense case where all docs have vectors.
Currently we write doc Ids of all documents that have vectors
not very efficiently.
This improve their encoding by for a case when all documents
have vectors, we don't write document IDs, but just write a
single short value – a dense marker.
Implement Weight#count for PointRangeQuery to provide a faster way to calculate
the number of matching range docs when each doc has at-most one point and the
points are 1-dimensional.
The recently introduced testCount verifies that the Weight#count optimization kicks in. When SimpleText codec is used, `DocValues#unwrapSingleton` returns null which disables the optimization and makes the test fail.
Alan Woodward
Adrien Grand
Peter Gromov
Lu Xugang
Anand
Luca Cavanna
Greg Miller
Robert Muir
gf2121
Julie Tibshirani
Ignacio Vera
Tomoko Uchida
Vigya Sharma
spike.liu
Mayya Sharipova
Gautam Worah
Patrick Zhai
Uwe Schindler
Dawid Weiss