Add hierarchical structure to the graph index #402
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
…where the node is present
vels to candidates pool in layer 0
… no dupes are submitted
…multiple maxDegrees with addHierarchy=False
# Conflicts: # jvector-base/src/main/java/io/github/jbellis/jvector/graph/disk/CommonHeader.java # jvector-base/src/main/java/io/github/jbellis/jvector/graph/disk/OnDiskGraphIndex.java # jvector-base/src/main/java/io/github/jbellis/jvector/graph/disk/OnDiskGraphIndexWriter.java # jvector-base/src/main/java/io/github/jbellis/jvector/quantization/BQVectors.java # jvector-examples/src/main/java/io/github/jbellis/jvector/example/Bench.java # jvector-examples/src/main/java/io/github/jbellis/jvector/example/Grid.java # jvector-examples/src/main/java/io/github/jbellis/jvector/example/IPCService.java # jvector-tests/src/test/java/io/github/jbellis/jvector/graph/Test2DThreshold.java # jvector-tests/src/test/java/io/github/jbellis/jvector/graph/disk/TestOnDiskGraphIndex.java
…om the same level. If the entrypoint has been deleted, the new one can be in level 0 if that's the new top level
…e live nodes in the bottom layer.
…t is smaller than layers.size() - 1
|
A point worth noting is that we recently updated the file format for graphs to V4 in main (#400). Since this PR should be merged fairly quickly after that one and V4 has not been included in a release yet, I opted for not bumping up the file version to V5 but to just update V4 instead. This was to avoid an unnecessary proliferation of versions that will complicate backwards compatibility in the future. |
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Outdated suggestions cannot be applied.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.
This PR introduces a hybrid HNSW-DiskANN graph. From HNSW, we take the idea of using multiple layers. This adds robustness to particularly hard datasets. Each layer is a Vamana graph. The upper layers reside in memory while the base layer resides on disk (in DiskANN style). It also enable using a single layer. In that case, it is a plain Vamana graph.