INTPYTHON-785 Only list authorized collections when listing collections (langchain-ai#226)

Fixes langchain-ai#225

---------

Co-authored-by: Rafael Odon <rafael.odon@serpro.gov.br>

Author: rafaelodon

.gitignore

@@ -9,3 +9,4 @@ __pycache__
 docs/langchain_mongodb
 docs/langgraph_checkpoint_mongodb
 docs/index.md
+.vscode

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

@@ -57,7 +57,9 @@ def __init__(
             )
         self._include_colls = set(include_collections or [])
         self._ignore_colls = set(ignore_collections or [])
-        self._all_colls = set(self._db.list_collection_names())
+        self._all_colls = set(
+            self._db.list_collection_names(authorizedCollections=True)
+        )
 
         self._sample_docs_in_coll_info = sample_docs_in_collection_info
         self._indexes_in_coll_info = indexes_in_collection_info

libs/langchain-mongodb/langchain_mongodb/cache.py

@@ -52,7 +52,9 @@ def __init__(
         self.__database_name = database_name
         self.__collection_name = collection_name
 
-        if self.__collection_name not in self.database.list_collection_names():
+        if self.__collection_name not in self.database.list_collection_names(
+            authorizedCollections=True
+        ):
             self.database.create_collection(self.__collection_name)
             # Create an index on key and llm_string
             self.collection.create_index([self.PROMPT, self.LLM])

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

@@ -142,7 +142,9 @@ def __init__(
                 driver=DRIVER_METADATA,
             )
             db = client[database_name]
-            if collection_name not in db.list_collection_names():
+            if collection_name not in db.list_collection_names(
+                authorizedCollections=True
+            ):
                 validator = {"$jsonSchema": self._schema} if validate else None
                 collection = client[database_name].create_collection(
                     collection_name,

libs/langchain-mongodb/langchain_mongodb/index.py

@@ -60,7 +60,9 @@ def create_vector_search_index(
     """
     logger.info("Creating Search Index %s on %s", index_name, collection.name)
 
-    if collection.name not in collection.database.list_collection_names():
+    if collection.name not in collection.database.list_collection_names(
+        authorizedCollections=True
+    ):
         collection.database.create_collection(collection.name)
 
     result = collection.create_search_index(
@@ -219,7 +221,9 @@ def create_fulltext_search_index(
     """
     logger.info("Creating Search Index %s on %s", index_name, collection.name)
 
-    if collection.name not in collection.database.list_collection_names():
+    if collection.name not in collection.database.list_collection_names(
+        authorizedCollections=True
+    ):
         collection.database.create_collection(collection.name)
 
     if isinstance(field, str):

libs/langchain-mongodb/tests/utils.py

@@ -238,7 +238,7 @@ class MockDatabase:
     def __init__(self, client=None):
         self.client = client or MockClient()
 
-    def list_collection_names(self) -> list[str]:
+    def list_collection_names(self, authorizedCollections: bool = True) -> list[str]:
         return ["test"]
 
     def __getitem__(self, key: str) -> Any:

libs/langgraph-store-mongodb/langgraph/store/mongodb/base.py

@@ -251,7 +251,9 @@ def from_conn_string(
                 driver=DRIVER_METADATA,
             )
             db = client[db_name]
-            if collection_name not in db.list_collection_names():
+            if collection_name not in db.list_collection_names(
+                authorizedCollections=True
+            ):
                 db.create_collection(collection_name)
             collection = client[db_name][collection_name]