Merge branch 'dev' into bug-KuzuAdpter

output format is same as the othere adpaters get_filtered_graph_data.

cognee/infrastructure/databases/graph/kuzu/adapter.py

@@ -1277,7 +1277,6 @@ class KuzuAdapter(GraphDBInterface):
             A tuple containing a list of filtered node properties and a list of filtered edge
             properties.
         """
-
         where_clauses = []
         params = {}
 
@@ -1288,16 +1287,50 @@ class KuzuAdapter(GraphDBInterface):
                 params[param_name] = values
 
         where_clause = " AND ".join(where_clauses)
-        nodes_query = f"MATCH (n:Node) WHERE {where_clause} RETURN properties(n)"
+        nodes_query = (
+            f"MATCH (n:Node) WHERE {where_clause} RETURN n.id, {{properties: n.properties}}"
+        )
         edges_query = f"""
         MATCH (n1:Node)-[r:EDGE]->(n2:Node)
         WHERE {where_clause.replace("n.", "n1.")} AND {where_clause.replace("n.", "n2.")}
-        RETURN properties(r)
+        RETURN n1.id, n2.id, r.relationship_name, r.properties
         """
         nodes, edges = await asyncio.gather(
             self.query(nodes_query, params), self.query(edges_query, params)
         )
-        return ([n[0] for n in nodes], [e[0] for e in edges])
+        formatted_nodes = []
+        for n in nodes:
+            if n[0]:
+                node_id = str(n[0])
+                props = n[1]
+                if props.get("properties"):
+                    try:
+                        additional_props = json.loads(props["properties"])
+                        props.update(additional_props)
+                        del props["properties"]
+                    except json.JSONDecodeError:
+                        logger.warning(f"Failed to parse properties JSON for node {node_id}")
+                formatted_nodes.append((node_id, props))
+        if not formatted_nodes:
+            logger.warning("No nodes found in the database")
+            return [], []
+
+        formatted_edges = []
+        for e in edges:
+            if e and len(e) >= 3:
+                source_id = str(e[0])
+                target_id = str(e[1])
+                rel_type = str(e[2])
+                props = {}
+                if len(e) > 3 and e[3]:
+                    try:
+                        props = json.loads(e[3])
+                    except (json.JSONDecodeError, TypeError):
+                        logger.warning(
+                            f"Failed to parse edge properties for {source_id}->{target_id}"
+                        )
+                formatted_edges.append((source_id, target_id, rel_type, props))
+        return formatted_nodes, formatted_edges
 
     async def get_graph_metrics(self, include_optional=False) -> Dict[str, Any]:
         """

cognee/modules/data/methods/create_authorized_dataset.py

@@ -6,6 +6,15 @@ from .create_dataset import create_dataset
 
 
 async def create_authorized_dataset(dataset_name: str, user: User) -> Dataset:
+    """
+        Create a new dataset and give all permissions on this dataset to the given user.
+    Args:
+        dataset_name: Name of the dataset.
+        user: The user object.
+
+    Returns:
+        Dataset: The new authorized dataset.
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/data/methods/get_authorized_dataset.py

@@ -15,7 +15,7 @@ async def get_authorized_dataset(
     Get a specific dataset with permissions for a user.
 
     Args:
-        user_id (UUID): user id
+        user: User object
         dataset_id (UUID): dataset id
         permission_type (str): permission type(read, write, delete, share), default is read
 

cognee/modules/data/methods/get_authorized_dataset_by_name.py

@@ -11,6 +11,17 @@ from ..models import Dataset
 async def get_authorized_dataset_by_name(
     dataset_name: str, user: User, permission_type: str
 ) -> Optional[Dataset]:
+    """
+    Get a specific dataset with the given name, with permissions for a given user.
+
+    Args:
+        dataset_name: Name of the dataset.
+        user: User object.
+        permission_type (str): permission type(read, write, delete, share), default is read
+
+    Returns:
+        Optional[Dataset]: dataset with permissions
+    """
     authorized_datasets = await get_authorized_existing_datasets([], permission_type, user)
 
     return next((dataset for dataset in authorized_datasets if dataset.name == dataset_name), None)

cognee/modules/graph/utils/retrieve_existing_edges.py

@@ -23,8 +23,6 @@ async def retrieve_existing_edges(
         chunk_graphs (list[KnowledgeGraph]): List of knowledge graphs corresponding to each
             data chunk. Each graph contains nodes (entities) and edges (relationships) that
             were extracted from the chunk content.
-        graph_engine (GraphDBInterface): Interface to the graph database that will be queried
-            to check for existing edges. Must implement the has_edges() method.
 
     Returns:
         dict[str, bool]: A mapping of edge keys to boolean values indicating existence.

cognee/modules/pipelines/layers/resolve_authorized_user_dataset.py

@@ -11,6 +11,19 @@ from cognee.modules.data.methods import (
 
 
 async def resolve_authorized_user_dataset(dataset_id: UUID, dataset_name: str, user: User):
+    """
+    Function handles creation and dataset authorization if dataset already exist for Cognee.
+    Verifies that provided user has necessary permission for provided Dataset.
+    If Dataset does not exist creates the Dataset and gives permission for the user creating the dataset.
+
+    Args:
+        dataset_id: Id of the dataset.
+        dataset_name: Name of the dataset.
+        user: Cognee User request is being processed for, if None default user will be used.
+
+    Returns:
+        Tuple[User, Dataset]: A tuple containing the user and the authorized dataset.
+    """
     if not user:
         user = await get_default_user()
 

cognee/modules/pipelines/layers/resolve_authorized_user_datasets.py

@@ -25,7 +25,7 @@ async def resolve_authorized_user_datasets(
         datasets: Dataset names or Dataset UUID (in case Datasets already exist)
 
     Returns:
-
+        Tuple[User, List[Dataset]]: A tuple containing the user and the list of authorized datasets.
     """
     # If no user is provided use default user
     if user is None:

cognee/modules/users/permissions/methods/authorized_give_permission_on_datasets.py

@@ -9,6 +9,18 @@ from uuid import UUID
 async def authorized_give_permission_on_datasets(
     principal_id: UUID, dataset_ids: Union[List[UUID], UUID], permission_name: str, owner_id: UUID
 ):
+    """
+        Give permission to certain datasets to a user.
+        The request owner must have the necessary permission to share the datasets.
+    Args:
+        principal_id: Id of user to whom datasets are shared
+        dataset_ids: Ids of datasets to share
+        permission_name: Name of permission to give
+        owner_id: Id of the request owner
+
+    Returns:
+        None
+    """
     # If only a single dataset UUID is provided transform it to a list
     if not isinstance(dataset_ids, list):
         dataset_ids = [dataset_ids]

cognee/modules/users/permissions/methods/check_permission_on_dataset.py

@@ -10,6 +10,17 @@ logger = get_logger()
 
 
 async def check_permission_on_dataset(user: User, permission_type: str, dataset_id: UUID):
+    """
+        Check if a user has a specific permission on a dataset.
+    Args:
+        user: User whose permission is checked
+        permission_type: Type of permission to check
+        dataset_id: Id of the dataset
+
+    Returns:
+        None
+
+    """
     if user is None:
         user = await get_default_user()
 

cognee/modules/users/permissions/methods/get_all_user_permission_datasets.py

@@ -11,6 +11,16 @@ logger = get_logger()
 
 
 async def get_all_user_permission_datasets(user: User, permission_type: str) -> list[Dataset]:
+    """
+        Return a list of datasets the user has permission for.
+        If the user is part of a tenant, return datasets his roles have permission for.
+    Args:
+        user
+        permission_type
+
+    Returns:
+        list[Dataset]: List of datasets user has permission for
+    """
     datasets = list()
     # Get all datasets User has explicit access to
     datasets.extend(await get_principal_datasets(user, permission_type))

cognee/modules/users/permissions/methods/get_document_ids_for_user.py

@@ -8,6 +8,16 @@ from ...models import ACL, Permission
 
 
 async def get_document_ids_for_user(user_id: UUID, datasets: list[str] = None) -> list[str]:
+    """
+        Return a list of documents ids for which the user has read permission.
+        If datasets are specified, return only documents from those datasets.
+    Args:
+        user_id: Id of the user
+        datasets: List of datasets
+
+    Returns:
+        list[str]: List of documents for which the user has read permission
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/get_principal.py

@@ -6,6 +6,15 @@ from ...models.Principal import Principal
 
 
 async def get_principal(principal_id: UUID):
+    """
+        Return information about a user based on their id
+    Args:
+        principal_id: Id of the user
+
+    Returns:
+        principal: Information about the user (principal)
+
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/get_principal_datasets.py

@@ -9,6 +9,17 @@ from ...models.ACL import ACL
 
 
 async def get_principal_datasets(principal: Principal, permission_type: str) -> list[Dataset]:
+    """
+        Return a list of datasets for which the user (principal) has a certain permission.
+    Args:
+        principal: Information about the user
+        permission_type: Type of permission
+
+    Returns:
+        list[Dataset]: List of datasets for which the user (principal)
+        has the permission (permission_type).
+
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/get_role.py

@@ -9,6 +9,16 @@ from ...models.Role import Role
 
 
 async def get_role(tenant_id: UUID, role_name: str):
+    """
+        Return the role with the name role_name of the given tenant.
+    Args:
+        tenant_id: Id of the given tenant
+        role_name: Name of the role
+
+    Returns
+        The role for the given tenant.
+
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/get_specific_user_permission_datasets.py

@@ -15,9 +15,9 @@ async def get_specific_user_permission_datasets(
         Return a list of datasets user has given permission for. If a list of datasets is provided,
         verify for which datasets user has appropriate permission for and return list of datasets he has permission for.
     Args:
-        user_id:
-        permission_type:
-        dataset_ids:
+        user_id: Id of the user.
+        permission_type: Type of the permission.
+        dataset_ids: Ids of the provided datasets
 
     Returns:
         list[Dataset]: List of datasets user has permission for

cognee/modules/users/permissions/methods/get_tenant.py

@@ -8,6 +8,15 @@ from ...models.Tenant import Tenant
 
 
 async def get_tenant(tenant_id: UUID):
+    """
+        Return information about the tenant based on the given id.
+    Args:
+        tenant_id: Id of the given tenant
+
+    Returns
+        Information about the given tenant.
+
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/give_default_permission_to_role.py

@@ -16,6 +16,15 @@ from cognee.modules.users.models import (
 
 
 async def give_default_permission_to_role(role_id: UUID, permission_name: str):
+    """
+        Give the permission with given name to the role with the given id as a default permission.
+    Args:
+        role_id: Id of the role
+        permission_name: Name of the permission
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/permissions/methods/give_default_permission_to_tenant.py

@@ -16,6 +16,15 @@ from cognee.modules.users.models import (
 
 
 async def give_default_permission_to_tenant(tenant_id: UUID, permission_name: str):
+    """
+        Give the permission with given name to the tenant with the given id as a default permission.
+    Args:
+        tenant_id: Id of the tenant
+        permission_name: Name of the permission
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         tenant = (

cognee/modules/users/permissions/methods/give_default_permission_to_user.py

@@ -16,6 +16,15 @@ from cognee.modules.users.models import (
 
 
 async def give_default_permission_to_user(user_id: UUID, permission_name: str):
+    """
+        Give the permission with given name to the user with the given id as a default permission.
+    Args:
+        user_id: Id of the tenant
+        permission_name: Name of the permission
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         user = (await session.execute(select(User).where(User.id == user_id))).scalars().first()

cognee/modules/users/permissions/methods/give_permission_on_dataset.py

@@ -24,6 +24,16 @@ async def give_permission_on_dataset(
     dataset_id: UUID,
     permission_name: str,
 ):
+    """
+        Give a specific permission on a dataset to a user.
+    Args:
+        principal: User who is being given the permission on the dataset
+        dataset_id: Id of the dataset
+        permission_name: Name of permission to give
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
 
     async with db_engine.get_async_session() as session:

cognee/modules/users/roles/methods/add_user_to_role.py

@@ -21,6 +21,17 @@ from cognee.modules.users.models import (
 
 
 async def add_user_to_role(user_id: UUID, role_id: UUID, owner_id: UUID):
+    """
+        Add a user with the given id to the role with the given id.
+    Args:
+        user_id: Id of the user.
+        role_id: Id of the role.
+        owner_id: Id of the request owner.
+
+    Returns:
+        None
+
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         user = (await session.execute(select(User).where(User.id == user_id))).scalars().first()

cognee/modules/users/roles/methods/create_role.py

@@ -16,6 +16,16 @@ async def create_role(
     role_name: str,
     owner_id: UUID,
 ):
+    """
+        Create a new role with the given name, if the request owner with the given id
+        has the necessary permission.
+    Args:
+        role_name: Name of the new role.
+        owner_id: Id of the request owner.
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         user = await get_user(owner_id)

cognee/modules/users/tenants/methods/add_user_to_tenant.py

@@ -13,6 +13,18 @@ from cognee.modules.users.exceptions import (
 
 
 async def add_user_to_tenant(user_id: UUID, tenant_id: UUID, owner_id: UUID):
+    """
+        Add a user with the given id to the tenant with the given id.
+        This can only be successful if the request owner with the given id is the tenant owner.
+    Args:
+        user_id: Id of the user.
+        tenant_id: Id of the tenant.
+        owner_id: Id of the request owner.
+
+    Returns:
+        None
+
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         user = await get_user(user_id)

cognee/modules/users/tenants/methods/create_tenant.py

@@ -8,6 +8,16 @@ from cognee.modules.users.methods import get_user
 
 
 async def create_tenant(tenant_name: str, user_id: UUID):
+    """
+        Create a new tenant with the given name, for the user with the given id.
+        This user is the owner of the tenant.
+    Args:
+        tenant_name: Name of the new tenant.
+        user_id: Id of the user.
+
+    Returns:
+        None
+    """
     db_engine = get_relational_engine()
     async with db_engine.get_async_session() as session:
         try: