This PR aims to fix#2047 by doing the following things:
- Add a distance_type parameter to the sync query builders of Python
SDK.
- Make metric an alias to distance_type.
* `createTable()` now saves embeddings in the schema metadata.
Previously, it would drop them. (`createEmptyTable()` was already tested
and worked.)
* `mergeInsert()` now uses embeddings.
Fixes#2066
Fixes#2031
When we do hybrid search, we normalize the scores. We do this
calculation in-place, because the Rerankers expect the `_distance` and
`_score` columns to be the normalized ones. So I've changed the logic so
that we restore the original distance and scores by matching on row ids.
When calling `replace_field_metadata` we pass in an iter of tuples
`(u32, HashMap<String, String>)`.
That `u32` needs to be the field id from the lance schema
7f60aa0a87/rust/lance-core/src/datatypes/field.rs (L123)
This can sometimes be different than the index of the field in the arrow
schema (e.g. if fields have been dropped).
This PR adds docs that try to clarify what that argument should be, as
well as corrects the usage in the test (which was improperly passing the
index of the arrow schema).
This includes several improvements and fixes to the Python Async query
builders:
1. The API reference docs show all the methods for each builder
2. The hybrid query builder now has all the same setter methods as the
vector search one, so you can now set things like `.distance_type()` on
a hybrid query.
3. Re-rankers are now properly hooked up and tested for FTS and vector
search. Previously the re-rankers were accidentally bypassed in unit
tests, because the builders overrode `.to_arrow()`, but the unit test
called `.to_batches()` which was only defined in the base class. Now all
builders implement `.to_batches()` and leave `.to_arrow()` to the base
class.
4. The `AsyncQueryBase` and `AsyncVectoryQueryBase` setter methods now
return `Self`, which provides the appropriate subclass as the type hint
return value. Previously, `AsyncQueryBase` had them all hard-coded to
`AsyncQuery`, which was unfortunate. (This required bringing in
`typing-extensions` for older Python version, but I think it's worth
it.)
related to #2014
this fixes:
- linear reranker may lost some results if the merging consumes all
vector results earlier than fts results
- linear reranker inverts the fts score but only vector distance can be
inverted
---------
Signed-off-by: BubbleCal <bubble-cal@outlook.com>
