Single KNN Field Optimisation #530
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
* made test_search pass * Updated CUDA version to match mainline
|
This PR replaces the POC PR (comments from the POC PR may still be relevant): #509 |
farshidz
requested changes
Jul 14, 2023
pandu-k
commented
Jul 15, 2023
| return to_be_sanitised | ||
|
|
||
|
|
||
| def add_chunks_prefix_to_filter_string_fields(filter_string: Optional[str], simple_properties: typing.Iterable) -> str: |
There was a problem hiding this comment.
if filter_string MUST NOT be None, it should be reflected in the signature:
Omit Optional[] and just leave filter_string: str
There was a problem hiding this comment.
filter string can be None, actually this is what is passed when the user does not input a filter string. simple_properties is the one that cannot be None.
| @@ -81,6 +81,47 @@ | |||
| logger = get_logger(__name__) | |||
|
|
|||
|
|
|||
| def _get_dimension_from_model_properties(model_properties: dict): | |||
There was a problem hiding this comment.
note for future work: this, and the other get_model_properties() function can be moved to its file, for example: models/model_properties.py
| ) | ||
|
|
||
|
|
||
| def _add_knn_field(ix_settings: dict): |
There was a problem hiding this comment.
note for future work: this can be moved, alongside create_index, into a dedicated file (like create_index.py)
farshidz
approved these changes
Jul 16, 2023
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Outdated suggestions cannot be applied.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.
What kind of change does this PR introduce? (Bug fix, feature, docs update, ...)
Optimisation
What is the current behavior? (You can also link to an open issue here)
Every field in a document has a separate OpenSearch knn vector field, thus a different HNSW graph (eg. 5 fields create 5 HNSW graphs). Increases RPS at scale for indexes with multiple tensor fields.
What is the new behavior (if this is a feature change)?
Each index will now have exactly 1 OpenSearch knn vector field:
__vector_marqo_knn_field.dimensionsinmodel_propertiesfilter_stringandsearchable_attributeswill be combined into 1 filter for OpenSearch for result retrieval\\. This is for simpler and cleaner Lucene DSL integration.Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)
Yes. Indexes created with a version prior to this cannot be searched with this version and vice versa. Old indexes need to be reindexed.
Have unit tests been run against this PR? (Has there also been any additional testing?)
Yes. Also manual testing on: boosting, score modifiers
Related API Test changes (link commit/PR here)
TO FOLLOW
Related Python client changes (link commit/PR here)
None
Related documentation changes (link commit/PR here)
TO FOLLOW
Please check if the PR fulfills these requirements