INTPYTHON-752 Integrate pymongo-vectorsearch-utils

Author: aclark4life

libs/langchain-mongodb/langchain_mongodb/agent_toolkit/database.py

@@ -64,7 +64,7 @@ def __init__(
         self._sample_docs_in_coll_info = sample_docs_in_collection_info
         self._indexes_in_coll_info = indexes_in_collection_info
 
-        _append_client_metadata(self._client)
+        _append_client_metadata(self._client, DRIVER_METADATA)
 
     @classmethod
     def from_connection_string(

libs/langchain-mongodb/langchain_mongodb/chat_message_histories.py

@@ -112,7 +112,7 @@ def __init__(
             if connection_string:
                 raise ValueError("Must provide connection_string or client, not both")
             self.client = client
-            _append_client_metadata(self.client)
+            _append_client_metadata(self.client, DRIVER_METADATA)
         elif connection_string:
             try:
                 self.client = MongoClient(

libs/langchain-mongodb/langchain_mongodb/docstores.py

@@ -37,7 +37,7 @@ def __init__(self, collection: Collection, text_key: str = "page_content") -> No
         self.collection = collection
         self._text_key = text_key
 
-        _append_client_metadata(self.collection.database.client)
+        _append_client_metadata(self.collection.database.client, DRIVER_METADATA)
 
     @classmethod
     def from_connection_string(
@@ -99,12 +99,13 @@ def mset(
             batch_size: Number of documents to insert at a time.
                 Tuning this may help with performance and sidestep MongoDB limits.
         """
-        keys, docs = zip(*key_value_pairs)
+        keys, docs = zip(*key_value_pairs, strict=False)
         n_docs = len(docs)
         start = 0
         for end in range(batch_size, n_docs + batch_size, batch_size):
             texts, metadatas = zip(
-                *[(doc.page_content, doc.metadata) for doc in docs[start:end]]
+                *[(doc.page_content, doc.metadata) for doc in docs[start:end]],
+                strict=False,
             )
             self.insert_many(texts=texts, metadatas=metadatas, ids=keys[start:end])  # type: ignore
             start = end
@@ -149,6 +150,7 @@ def insert_many(
         in the batch that do not have conflicting _ids will still be inserted.
         """
         to_insert = [
-            {"_id": i, self._text_key: t, **m} for i, t, m in zip(ids, texts, metadatas)
+            {"_id": i, self._text_key: t, **m}
+            for i, t, m in zip(ids, texts, metadatas, strict=False)
         ]
         self.collection.insert_many(to_insert)  # type: ignore

libs/langchain-mongodb/langchain_mongodb/graphrag/graph.py

@@ -186,7 +186,7 @@ def __init__(
         self.collection = collection
 
         # append_metadata was added in PyMongo 4.14.0, but is a valid database name on earlier versions
-        _append_client_metadata(collection.database.client)
+        _append_client_metadata(collection.database.client, DRIVER_METADATA)
 
         self.entity_extraction_model = entity_extraction_model
         self.entity_prompt = (

libs/langchain-mongodb/langchain_mongodb/index.py

@@ -7,6 +7,13 @@
 from pymongo.collection import Collection
 from pymongo.operations import SearchIndexModel
 
+# Don't break imports for modules that expect these functions
+# to be in this module.
+from pymongo_search_utils import (  # noqa: F401
+    create_vector_search_index,
+    update_vector_search_index,
+)
+
 logger = logging.getLogger(__file__)
 
 
@@ -34,60 +41,6 @@ def _vector_search_index_definition(
     return definition
 
 
-def create_vector_search_index(
-    collection: Collection,
-    index_name: str,
-    dimensions: int,
-    path: str,
-    similarity: str,
-    filters: Optional[List[str]] = None,
-    *,
-    wait_until_complete: Optional[float] = None,
-    **kwargs: Any,
-) -> None:
-    """Experimental Utility function to create a vector search index
-
-    Args:
-        collection (Collection): MongoDB Collection
-        index_name (str): Name of Index
-        dimensions (int): Number of dimensions in embedding
-        path (str): field with vector embedding
-        similarity (str): The similarity score used for the index
-        filters (List[str]): Fields/paths to index to allow filtering in $vectorSearch
-        wait_until_complete (Optional[float]): If provided, number of seconds to wait
-            until search index is ready.
-        kwargs: Keyword arguments supplying any additional options to SearchIndexModel.
-    """
-    logger.info("Creating Search Index %s on %s", index_name, collection.name)
-
-    if collection.name not in collection.database.list_collection_names(
-        authorizedCollections=True
-    ):
-        collection.database.create_collection(collection.name)
-
-    result = collection.create_search_index(
-        SearchIndexModel(
-            definition=_vector_search_index_definition(
-                dimensions=dimensions,
-                path=path,
-                similarity=similarity,
-                filters=filters,
-                **kwargs,
-            ),
-            name=index_name,
-            type="vectorSearch",
-        )
-    )
-
-    if wait_until_complete:
-        _wait_for_predicate(
-            predicate=lambda: _is_index_ready(collection, index_name),
-            err=f"{index_name=} did not complete in {wait_until_complete}!",
-            timeout=wait_until_complete,
-        )
-    logger.info(result)
-
-
 def drop_vector_search_index(
     collection: Collection,
     index_name: str,
@@ -115,54 +68,6 @@ def drop_vector_search_index(
     logger.info("Vector Search index %s.%s dropped", collection.name, index_name)
 
 
-def update_vector_search_index(
-    collection: Collection,
-    index_name: str,
-    dimensions: int,
-    path: str,
-    similarity: str,
-    filters: Optional[List[str]] = None,
-    *,
-    wait_until_complete: Optional[float] = None,
-    **kwargs: Any,
-) -> None:
-    """Update a search index.
-
-    Replace the existing index definition with the provided definition.
-
-    Args:
-        collection (Collection): MongoDB Collection
-        index_name (str): Name of Index
-        dimensions (int): Number of dimensions in embedding
-        path (str): field with vector embedding
-        similarity (str): The similarity score used for the index.
-        filters (List[str]): Fields/paths to index to allow filtering in $vectorSearch
-        wait_until_complete (Optional[float]): If provided, number of seconds to wait
-            until search index is ready.
-        kwargs: Keyword arguments supplying any additional options to SearchIndexModel.
-    """
-    logger.info(
-        "Updating Search Index %s from Collection: %s", index_name, collection.name
-    )
-    collection.update_search_index(
-        name=index_name,
-        definition=_vector_search_index_definition(
-            dimensions=dimensions,
-            path=path,
-            similarity=similarity,
-            filters=filters,
-            **kwargs,
-        ),
-    )
-    if wait_until_complete:
-        _wait_for_predicate(
-            predicate=lambda: _is_index_ready(collection, index_name),
-            err=f"Index {index_name} update did not complete in {wait_until_complete}!",
-            timeout=wait_until_complete,
-        )
-    logger.info("Update succeeded")
-
-
 def _is_index_ready(collection: Collection, index_name: str) -> bool:
     """Check for the index name in the list of available search indexes to see if the
     specified index is of status READY

libs/langchain-mongodb/langchain_mongodb/indexes.py

@@ -36,7 +36,7 @@ def __init__(self, collection: Collection) -> None:
         super().__init__(namespace=namespace)
         self._collection = collection
 
-        _append_client_metadata(self._collection.database.client)
+        _append_client_metadata(self._collection.database.client, DRIVER_METADATA)
 
     @classmethod
     def from_connection_string(
@@ -85,7 +85,7 @@ def update(
         if len(keys) != len(group_ids):
             raise ValueError("Number of keys does not match number of group_ids")
 
-        for key, group_id in zip(keys, group_ids):
+        for key, group_id in zip(keys, group_ids, strict=False):
             self._collection.find_one_and_update(
                 {"namespace": self.namespace, "key": key},
                 {"$set": {"group_id": group_id, "updated_at": self.get_time()}},

libs/langchain-mongodb/langchain_mongodb/loaders.py

@@ -54,7 +54,7 @@ def __init__(
         self.include_db_collection_in_metadata = include_db_collection_in_metadata
 
         # append_metadata was added in PyMongo 4.14.0, but is a valid database name on earlier versions
-        _append_client_metadata(self.db.client)
+        _append_client_metadata(self.db.client, DRIVER_METADATA)
 
     @classmethod
     def from_connection_string(

libs/langchain-mongodb/langchain_mongodb/retrievers/full_text_search.py

@@ -8,7 +8,11 @@
 from pymongo.collection import Collection
 
 from langchain_mongodb.pipelines import text_search_stage
-from langchain_mongodb.utils import _append_client_metadata, make_serializable
+from langchain_mongodb.utils import (
+    DRIVER_METADATA,
+    _append_client_metadata,
+    make_serializable,
+)
 
 
 class MongoDBAtlasFullTextSearchRetriever(BaseRetriever):
@@ -64,7 +68,7 @@ def _get_relevant_documents(
         )
 
         if not self._added_metadata:
-            _append_client_metadata(self.collection.database.client)
+            _append_client_metadata(self.collection.database.client, DRIVER_METADATA)
             self._added_metadata = True
 
         # Execution

libs/langchain-mongodb/langchain_mongodb/utils.py

@@ -24,22 +24,21 @@
 from typing import Any, Dict, List, Union
 
 import numpy as np
-from pymongo import MongoClient
 from pymongo.driver_info import DriverInfo
 
+# Don't break imports for modules that expect this function
+# to be in this module.
+from pymongo_search_utils import (
+    append_client_metadata as _append_client_metadata,  # noqa: F401
+)
+
 logger = logging.getLogger(__name__)
 
 Matrix = Union[List[List[float]], List[np.ndarray], np.ndarray]
 
 DRIVER_METADATA = DriverInfo(name="Langchain", version=version("langchain-mongodb"))
 
 
-def _append_client_metadata(client: MongoClient) -> None:
-    # append_metadata was added in PyMongo 4.14.0, but is a valid database name on earlier versions
-    if callable(client.append_metadata):
-        client.append_metadata(DRIVER_METADATA)
-
-
 def cosine_similarity(X: Matrix, Y: Matrix) -> np.ndarray:
     """Row-wise cosine similarity between two equal-width matrices."""
     if len(X) == 0 or len(Y) == 0: