Archives
/
langchain
mirror of https://github.com/hwchase17/langchain
2
0
Fork 1
Code Issues Packages Projects Releases Wiki Activity

community[minor]: Add hybrid search to Cassandra VectorStore (#20286)

Browse Source 
Only supported by Astra DB at the moment.
**Twitter handle:** cbornet_
pull/20012/head^2
Christophe Bornet 3 months ago committed by GitHub
parent d2d01370bc
commit 8f0b5687a3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 64 additions and 17 deletions
Show Stats Download Patch File Download Diff File

81
libs/community/langchain_community/vectorstores/cassandra.py
Unescape Escape View File

@ -57,17 +57,12 @@ class Cassandra(VectorStore):
keyspace: Cassandra key space.
table_name: Cassandra table.
ttl_seconds: Optional time-to-live for the added texts.
body_index_options: Optional options used to create the body index.
Eg. body_index_options = [cassio.table.cql.STANDARD_ANALYZER]
"""
_embedding_dimension: Union[int, None]
@staticmethod
def _filter_to_metadata(filter_dict: Optional[Dict[str, str]]) -> Dict[str, Any]:
if filter_dict is None:
return {}
else:
return filter_dict
def _get_embedding_dimension(self) -> int:
if self._embedding_dimension is None:
self._embedding_dimension = len(
@ -82,6 +77,8 @@ class Cassandra(VectorStore):
keyspace: str,
table_name: str,
ttl_seconds: Optional[int] = None,
*,
body_index_options: Optional[List[Tuple[str, Any]]] = None,
) -> None:
try:
from cassio.table import MetadataVectorCassandraTable
@ -99,6 +96,10 @@ class Cassandra(VectorStore):
#
self._embedding_dimension = None
#
kwargs = {}
if body_index_options is not None:
kwargs["body_index_options"] = body_index_options
self.table = MetadataVectorCassandraTable(
session=session,
keyspace=keyspace,
@ -106,6 +107,7 @@ class Cassandra(VectorStore):
vector_dimension=self._get_embedding_dimension(),
metadata_indexing="all",
primary_key_type="TEXT",
**kwargs,
)
@property
@ -212,22 +214,30 @@ class Cassandra(VectorStore):
embedding: List[float],
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
) -> List[Tuple[Document, float, str]]:
"""Return docs most similar to embedding vector.
Args:
embedding (str): Embedding to look up documents similar to.
k (int): Number of Documents to return. Defaults to 4.
embedding: Embedding to look up documents similar to.
k: Number of Documents to return. Defaults to 4.
filter: Filter on the metadata to apply.
body_search: Document textual search terms to apply.
Only supported by Astra DB at the moment.
Returns:
List of (Document, score, id), the most similar to the query vector.
"""
search_metadata = self._filter_to_metadata(filter)
kwargs: Dict[str, Any] = {}
if filter is not None:
kwargs["metadata"] = filter
if body_search is not None:
kwargs["body_search"] = body_search
#
hits = self.table.metric_ann_search(
vector=embedding,
n=k,
metric="cos",
metadata=search_metadata,
**kwargs,
)
# We stick to 'cos' distance as it can be normalized on a 0-1 axis
# (1=most relevant), as required by this class' contract.
@ -248,12 +258,14 @@ class Cassandra(VectorStore):
query: str,
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
) -> List[Tuple[Document, float, str]]:
embedding_vector = self.embedding.embed_query(query)
return self.similarity_search_with_score_id_by_vector(
embedding=embedding_vector,
k=k,
filter=filter,
body_search=body_search,
)
# id-unaware search facilities
@ -262,12 +274,16 @@ class Cassandra(VectorStore):
embedding: List[float],
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
) -> List[Tuple[Document, float]]:
"""Return docs most similar to embedding vector.
Args:
embedding (str): Embedding to look up documents similar to.
k (int): Number of Documents to return. Defaults to 4.
embedding: Embedding to look up documents similar to.
k: Number of Documents to return. Defaults to 4.
filter: Filter on the metadata to apply.
body_search: Document textual search terms to apply.
Only supported by Astra DB at the moment.
Returns:
List of (Document, score), the most similar to the query vector.
"""
@ -277,6 +293,7 @@ class Cassandra(VectorStore):
embedding=embedding,
k=k,
filter=filter,
body_search=body_search,
)
]
@ -285,6 +302,7 @@ class Cassandra(VectorStore):
query: str,
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
**kwargs: Any,
) -> List[Document]:
embedding_vector = self.embedding.embed_query(query)
@ -292,6 +310,7 @@ class Cassandra(VectorStore):
embedding_vector,
k,
filter=filter,
body_search=body_search,
)
def similarity_search_by_vector(
@ -299,6 +318,7 @@ class Cassandra(VectorStore):
embedding: List[float],
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
**kwargs: Any,
) -> List[Document]:
return [
@ -307,6 +327,7 @@ class Cassandra(VectorStore):
embedding,
k,
filter=filter,
body_search=body_search,
)
]
@ -315,12 +336,14 @@ class Cassandra(VectorStore):
query: str,
k: int = 4,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
) -> List[Tuple[Document, float]]:
embedding_vector = self.embedding.embed_query(query)
return self.similarity_search_with_score_by_vector(
embedding_vector,
k,
filter=filter,
body_search=body_search,
)
def max_marginal_relevance_search_by_vector(
@ -330,6 +353,7 @@ class Cassandra(VectorStore):
fetch_k: int = 20,
lambda_mult: float = 0.5,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
**kwargs: Any,
) -> List[Document]:
"""Return docs selected using the maximal marginal relevance.
@ -337,22 +361,31 @@ class Cassandra(VectorStore):
among selected documents.
Args:
embedding: Embedding to look up documents similar to.
k: Number of Documents to return.
k: Number of Documents to return. Defaults to 4.
fetch_k: Number of Documents to fetch to pass to MMR algorithm.
Defaults to 20.
lambda_mult: Number between 0 and 1 that determines the degree
of diversity among the results with 0 corresponding to maximum
diversity and 1 to minimum diversity.
Defaults to 0.5.
filter: Filter on the metadata to apply.
body_search: Document textual search terms to apply.
Only supported by Astra DB at the moment.
Returns:
List of Documents selected by maximal marginal relevance.
"""
search_metadata = self._filter_to_metadata(filter)
_kwargs: Dict[str, Any] = {}
if filter is not None:
_kwargs["metadata"] = filter
if body_search is not None:
_kwargs["body_search"] = body_search
prefetch_hits = list(
self.table.metric_ann_search(
vector=embedding,
n=fetch_k,
metric="cos",
metadata=search_metadata,
**_kwargs,
)
)
# let the mmr utility pick the *indices* in the above array
@ -382,6 +415,7 @@ class Cassandra(VectorStore):
fetch_k: int = 20,
lambda_mult: float = 0.5,
filter: Optional[Dict[str, str]] = None,
body_search: Optional[Union[str, List[str]]] = None,
**kwargs: Any,
) -> List[Document]:
"""Return docs selected using the maximal marginal relevance.
@ -389,12 +423,16 @@ class Cassandra(VectorStore):
among selected documents.
Args:
query: Text to look up documents similar to.
k: Number of Documents to return.
k: Number of Documents to return. Defaults to 4.
fetch_k: Number of Documents to fetch to pass to MMR algorithm.
Defaults to 20.
lambda_mult: Number between 0 and 1 that determines the degree
of diversity among the results with 0 corresponding to maximum
diversity and 1 to minimum diversity.
Defaults to 0.5.
filter: Filter on the metadata to apply.
body_search: Document textual search terms to apply.
Only supported by Astra DB at the moment.
Returns:
List of Documents selected by maximal marginal relevance.
"""
@ -405,6 +443,7 @@ class Cassandra(VectorStore):
fetch_k,
lambda_mult=lambda_mult,
filter=filter,
body_search=body_search,
)
@classmethod
@ -420,6 +459,7 @@ class Cassandra(VectorStore):
ids: Optional[List[str]] = None,
batch_size: int = 16,
ttl_seconds: Optional[int] = None,
body_index_options: Optional[List[Tuple[str, Any]]] = None,
**kwargs: Any,
) -> CVST:
"""Create a Cassandra vectorstore from raw texts.
@ -435,6 +475,8 @@ class Cassandra(VectorStore):
batch_size: Number of concurrent requests to send to the server.
Defaults to 16.
ttl_seconds: Optional time-to-live for the added texts.
body_index_options: Optional options used to create the body index.
Eg. body_index_options = [cassio.table.cql.STANDARD_ANALYZER]
Returns:
a Cassandra vectorstore.
@ -451,6 +493,7 @@ class Cassandra(VectorStore):
keyspace=keyspace,
table_name=table_name,
ttl_seconds=ttl_seconds,
body_index_options=body_index_options,
)
store.add_texts(
texts=texts, metadatas=metadatas, ids=ids, batch_size=batch_size
@ -469,6 +512,7 @@ class Cassandra(VectorStore):
ids: Optional[List[str]] = None,
batch_size: int = 16,
ttl_seconds: Optional[int] = None,
body_index_options: Optional[List[Tuple[str, Any]]] = None,
**kwargs: Any,
) -> CVST:
"""Create a Cassandra vectorstore from a document list.
@ -483,6 +527,8 @@ class Cassandra(VectorStore):
batch_size: Number of concurrent requests to send to the server.
Defaults to 16.
ttl_seconds: Optional time-to-live for the added documents.
body_index_options: Optional options used to create the body index.
Eg. body_index_options = [cassio.table.cql.STANDARD_ANALYZER]
Returns:
a Cassandra vectorstore.
@ -499,5 +545,6 @@ class Cassandra(VectorStore):
ids=ids,
batch_size=batch_size,
ttl_seconds=ttl_seconds,
body_index_options=body_index_options,
**kwargs,
)

Write Preview
Loading…
Cancel
Save