AsyncAzureCosmosDBNoSqlSemanticCache(
self,
embedding: Embeddings,
*,
vector_embedding_policy: Dict[str,| Name | Type | Description |
|---|---|---|
embedding* | Embeddings | CosmosDB Embedding. |
vector_embedding_policy* | Dict[str, Any] | CosmosDB vector embedding policy. |
indexing_policy* | Dict[str, Any] | |
cosmos_container_properties* | Dict[str, Any] | |
cosmos_database_properties* | Dict[str, Any] | |
vector_search_fields* | Dict[str, Any] | |
database_name | str | Default: 'CosmosNoSqlCacheDB' |
container_name | str | Default: 'CosmosNoSqlCacheContainer' |
search_type | str | Default: 'vector' |
create_container | bool | Default: True |
score_threshold | float | Default: 0.5 |
Async cache that uses Cosmos DB NoSQL backend.
CosmosDB indexing policy.
CosmosDB container properties.
CosmosDB database properties.
Vector Search Fields for the container.
CosmosDB database name.
CosmosDB container name.
CosmosDB search type.
Create the container if it doesn't exist.
Threshold for a cache hit. For cosine and dot product, this is the minimum similarity score — results below it are treated as cache misses. For Euclidean, this is the maximum distance — results above it are misses.