KNNIterators should support with and without filters #2155
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
requested review from
heemin32,
navneet1v,
vamshin,
jmazanec15,
naveentatikonda,
junqiu-lei,
martin-gaievski,
ryanbogan and
luyuncheng
as code owners
navneet1v
reviewed
Sep 27, 2024
Comment on lines
30
to
31
| public ByteVectorIdsKNNIterator( | ||
| final BitSet filterIdsBitSet, |
There was a problem hiding this comment.
should we make filterIdsBitSet as Optional? or atleast annotate as @nullable to provide a signal that this parameter can be null.
Update VectorIterator and NesterVector Iterator to iterate even if there is no filters provided to iterator. Currently this is used by exact search to score either topk docs or all docs when filter is provided by users. However, in future we will be allowing exact search even if there are no filters. Hence, decouple filter and make it option to support both cases. Signed-off-by: Vijayan Balasubramanian <[email protected]>
Signed-off-by: Vijayan Balasubramanian <[email protected]>
navneet1v
reviewed
Sep 27, 2024
src/main/java/org/opensearch/knn/index/query/iterators/VectorIdsKNNIterator.java
Outdated
Show resolved
Hide resolved
VijayanB
force-pushed
the
refactor-iterator-new
branch
from
7b53e0c to
71dba07
Compare
heemin32
reviewed
Sep 27, 2024
| this.docId = getNextDocId(); | ||
| } | ||
|
|
||
| ByteVectorIdsKNNIterator(final byte[] queryVector, final KNNBinaryVectorValues binaryVectorValues, final SpaceType spaceType) |
There was a problem hiding this comment.
Suggested change
| ByteVectorIdsKNNIterator(final byte[] queryVector, final KNNBinaryVectorValues binaryVectorValues, final SpaceType spaceType) | |
| public ByteVectorIdsKNNIterator(final byte[] queryVector, final KNNBinaryVectorValues binaryVectorValues, final SpaceType spaceType) |
| super(filterIdsArray, queryVector, binaryVectorValues, spaceType); | ||
| this.parentBitSet = parentBitSet; | ||
| } | ||
|
|
||
| NestedByteVectorIdsKNNIterator( |
There was a problem hiding this comment.
Suggested change
| NestedByteVectorIdsKNNIterator( | |
| public NestedByteVectorIdsKNNIterator( |
| private final BitSet parentBitSet; | ||
|
|
||
| NestedFilteredIdsKNNIterator( | ||
| final BitSet filterIdsArray, | ||
| NestedVectorIdsKNNIterator( |
There was a problem hiding this comment.
Suggested change
| NestedVectorIdsKNNIterator( | |
| public NestedVectorIdsKNNIterator( |
| this(filterIdsArray, queryVector, knnFloatVectorValues, spaceType, parentBitSet, null, null); | ||
| } | ||
|
|
||
| public NestedFilteredIdsKNNIterator( | ||
| NestedVectorIdsKNNIterator( |
There was a problem hiding this comment.
Suggested change
| NestedVectorIdsKNNIterator( | |
| public NestedVectorIdsKNNIterator( |
src/main/java/org/opensearch/knn/index/query/iterators/VectorIdsKNNIterator.java
Outdated
Show resolved
Hide resolved
src/main/java/org/opensearch/knn/index/query/iterators/VectorIdsKNNIterator.java
Outdated
Show resolved
Hide resolved
| @@ -55,10 +58,13 @@ public int nextDoc() throws IOException { | |||
| if (docId == DocIdSetIterator.NO_MORE_DOCS) { | |||
| return DocIdSetIterator.NO_MORE_DOCS; | |||
| } | |||
| int doc = binaryVectorValues.advance(docId); | |||
| if (bitSetIterator != null) { | |||
| binaryVectorValues.advance(docId); | |||
There was a problem hiding this comment.
Can we move this line inside getNextDocId()? See line#84 for example.
|
|
||
| protected int getNextDocId() throws IOException { | ||
| if (bitSetIterator != null) { | ||
| return this.bitSetIterator.nextDoc(); |
There was a problem hiding this comment.
Let's remove this for consistency. this is mostly needed only inside constructor to distinguish the variable from parameter.
Suggested change
| return this.bitSetIterator.nextDoc(); | |
| int docId = bitSetIterator.nextDoc(); | |
| binaryVectorValues.advance(docId); | |
| return docId; |
There was a problem hiding this comment.
@heemin32 It is missing NO_DOCUMENTS check. Please check my response here #2155 (comment)
There was a problem hiding this comment.
I am good with having NO_DOCUMENTS check here if needed.
Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
force-pushed
the
refactor-iterator-new
branch
from
71dba07 to
a9ab6cd
Compare
Signed-off-by: Vijayan Balasubramanian <[email protected]>
VijayanB
force-pushed
the
refactor-iterator-new
branch
from
a9ab6cd to
71adc4f
Compare
Signed-off-by: Vijayan Balasubramanian <[email protected]>
heemin32
approved these changes
Sep 28, 2024
navneet1v
approved these changes
Sep 28, 2024
opensearch-trigger-bot bot
pushed a commit
that referenced
this pull request
Sep 28, 2024
* Rename class names to represent both and filter and non filter use cases * Iterator should support with filters Update VectorIterator and NesterVector Iterator to iterate even if there is no filters provided to iterator. Currently this is used by exact search to score either topk docs or all docs when filter is provided by users. However, in future we will be allowing exact search even if there are no filters. Hence, decouple filter and make it option to support both cases. --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 6f6dd56)
VijayanB
added a commit
that referenced
this pull request
Sep 28, 2024
* Rename class names to represent both and filter and non filter use cases * Iterator should support with filters Update VectorIterator and NesterVector Iterator to iterate even if there is no filters provided to iterator. Currently this is used by exact search to score either topk docs or all docs when filter is provided by users. However, in future we will be allowing exact search even if there are no filters. Hence, decouple filter and make it option to support both cases. --------- Signed-off-by: Vijayan Balasubramanian <[email protected]> (cherry picked from commit 6f6dd56) Co-authored-by: Vijayan Balasubramanian <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Update VectorIterator, ByteIterator, NestedVectorIterator, NestedByteIterator to iterate even if there are no filters provided.
Currently this is used by exact search to score either topK docs or all docs when filter is provided by users.
However, in future we will be allowing exact search even if there are no filters. Hence, decouple filter
and make it option to support both cases.
Related Issues
Pre-requisite for #1942
Check List
--signoff.By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.