Vector Search
7 results found
-
Improve Performance when filtering collection as part of Atlas Vector Search
Adding filters to the $vectorSearch aggregation step should improve performance of vectorSearch given that we are searching over a smaller subset of the collection. Especially if those filters are for fields that we have an index for.
When testing vectorSearch performance, this is not the case:
Our setup:
Index exists with type: vectorSearch, Index Fields: fieldA, fieldB, fieldC. Status is Active.Add this filter to the $vectorSearch aggregation step:
'filter': {
'$and': [
{
‘fieldA’: value1,
‘fieldB’: False,
}
]
},The original query times are as follows:
P50 Total Query Time: 2.0953004360198975
P95 Total Query Time: 3.579429221153259
P99 Total…5 votes -
Vector Search Index support in pymongo
Currently creating a Search Index of type "vector" is not possible with pymongo v4.6.2. https://www.mongodb.com/community/forums/t/create-vector-search-pymongo-latest-version-returns-command-not-found/267874/3?u=matthieu_mazzolini
4 votes -
Atlas Vector Search + Geospatial, please! 🙏
Would love to be able to do a geospatial query like $geoWithin together with Atlas Vector Search query!
3 votes -
Threshold-based similarity search
We should support threshold-based similarity search instead of solely topK search. This would allow us to more efficiently gather results by considering a relevancy threshold within the prefilter.
3 votes -
RAG and Semantic Search Tools in the Atlas Search Playground
Have ideas about what features would be helpful for RAG and Semantic Search in the Atlas Search Playground? Let us know!
2 votes -
Project out embedding fields by default in $vectorSearch
We should update the default behavior to project out embeddings, since they are typically not used in subsequent aggregation stages, with an additional query parameter to force include them.
2 votes -
Support search on vectors of integers
Right now we only support search on vectors of doubles, but we may want to support more types such as ints or booleans that are typically associated with quantized representations of embeddings.
2 votes
- Don't see your idea?