Merge 8b0b552da8 into 81eb03d230

common/data_source/teams_connector.py

@@ -1,27 +1,32 @@
 """Microsoft Teams connector"""
 
 from typing import Any
-
+import logging
 import msal
 from office365.graph_client import GraphClient
 from office365.runtime.client_request_exception import ClientRequestException
+from common.data_source.utils import run_with_timeout
 
 from common.data_source.exceptions import (
     ConnectorValidationError,
     InsufficientPermissionsError,
-    UnexpectedValidationError, ConnectorMissingCredentialError
+    UnexpectedValidationError, 
+    ConnectorMissingCredentialError
 )
 from common.data_source.interfaces import (
     SecondsSinceUnixEpoch,
-    SlimConnectorWithPermSync, CheckpointedConnectorWithPermSync
+    SlimConnectorWithPermSync, 
+    CheckpointedConnectorWithPermSync
 )
 from common.data_source.models import (
     ConnectorCheckpoint
 )
 
-_SLIM_DOC_BATCH_SIZE = 5000
 
 
+_SLIM_DOC_BATCH_SIZE = 5000
+_MAX_WORKERS = 10
+
 class TeamsCheckpoint(ConnectorCheckpoint):
     """Teams-specific checkpoint"""
     todo_team_ids: list[str] | None = None
@@ -30,9 +35,12 @@ class TeamsCheckpoint(ConnectorCheckpoint):
 class TeamsConnector(CheckpointedConnectorWithPermSync, SlimConnectorWithPermSync):
     """Microsoft Teams connector for accessing Teams messages and channels"""
 
-    def __init__(self, batch_size: int = _SLIM_DOC_BATCH_SIZE) -> None:
+    def __init__(self, teams_lst: list[str] = None, max_workers: int = _MAX_WORKERS) -> None:
-        self.batch_size = batch_size
+        self.teams_lst = teams_lst
-        self.teams_client = None
+        self.max_workers = max_workers
+        self.teams_client: GraphClient | None = None
+        self.msal_app: msal.ConfidentialClientApplication | None = None
+
 
     def load_credentials(self, credentials: dict[str, Any]) -> dict[str, Any] | None:
         """Load Microsoft Teams credentials"""
@@ -45,20 +53,25 @@ class TeamsConnector(CheckpointedConnectorWithPermSync, SlimConnectorWithPermSyn
                 raise ConnectorMissingCredentialError("Microsoft Teams credentials are incomplete")
             
             # Create MSAL confidential client
-            app = msal.ConfidentialClientApplication(
+            self.msal_app = msal.ConfidentialClientApplication(
                 client_id=client_id,
                 client_credential=client_secret,
                 authority=f"https://login.microsoftonline.com/{tenant_id}"
             )
             
-            # Get access token
+            def _acquire_token_callback() -> dict[str, Any]:
-            result = app.acquire_token_for_client(scopes=["https://graph.microsoft.com/.default"])
+                if self.msal_app is None:
+                    raise RuntimeError("Failed to create MSAL ConfidentialClientApplication")
             
-            if "access_token" not in result:
+                # Get access token
-                raise ConnectorMissingCredentialError("Failed to acquire Microsoft Teams access token")
+                token = self.msal_app.acquire_token_for_client(scopes=["https://graph.microsoft.com/.default"])
+                if not isinstance(token, dict) or "access_token" not in token:
+                    raise RuntimeError("Failed to acquire token for Microsoft Graph API")
+            
+                return token
 
             # Create Graph client for Teams
-            self.teams_client = GraphClient(result["access_token"])
+            self.teams_client = GraphClient(token_callback=_acquire_token_callback)
             
             return None
         except Exception as e:
@@ -69,16 +82,67 @@ class TeamsConnector(CheckpointedConnectorWithPermSync, SlimConnectorWithPermSyn
         if not self.teams_client:
             raise ConnectorMissingCredentialError("Microsoft Teams")
         
+        # Check for special characters in team names
+        has_special_chars = self._has_odata_incompatible_chars(self.teams_lst)
+        if has_special_chars:
+            logging.info(
+                "Some requested team names contain special characters (&, (, )) that require "
+                "client-side filtering during data retrieval."
+            )
+
+        timeout = 10
         try:
-            # Test connection by getting teams
+            logging.info(
-            teams = self.teams_client.teams.get().execute_query()
+                f"Requested team count: {len(self.teams_lst) if self.teams_lst else 0}, "
-            if not teams:
+                f"Has special chars: {has_special_chars}"
-                raise ConnectorValidationError("Failed to access Microsoft Teams")
+            )
+
+            validation_query = self.teams_client.teams.get().top(1)
+            run_with_timeout(
+                timeout=timeout,
+                func=lambda: validation_query.execute_query()
+            )
+            
+            logging.info("Microsoft Teams connector settings validated successfully.")
+
+        except TimeoutError as e:
+            raise ConnectorValidationError(
+                f"Timeout while validating Teams access (waited {timeout}s). "
+                f"This may indicate network issues or authentication problems. "
+                f"Error: {e}"
+            )
         except ClientRequestException as e:
-            if "401" in str(e) or "403" in str(e):
+            if not e.response:
-                raise InsufficientPermissionsError("Invalid credentials or insufficient permissions")
+                raise RuntimeError(f"No response provided in {e=}")
-            else:
+            status_code = e.response.status_code
-                raise UnexpectedValidationError(f"Microsoft Teams validation error: {e}")
+            if status_code == 401:
+                raise ConnectorValidationError(
+                    "Invalid or expired Microsoft Teams credentials. (401 Unauthorized)"
+                )
+            elif status_code == 403:
+                raise InsufficientPermissionsError(
+                    "Microsoft Teams connector lacks necessary permissions. (403 Forbidden)"
+                )
+            raise UnexpectedValidationError(
+                f"Unexpected error during Teams validation: {e} (Status code: {status_code})"
+            )
+        except Exception as e:
+            error_str = str(e).lower()
+            if (
+                "unauthorized" in error_str
+                or "401" in error_str
+                or "invalid_grant" in error_str
+            ):
+                raise ConnectorValidationError(
+                    "Invalid or expired Microsoft Teams credentials."
+                )
+            elif "forbidden" in error_str or "403" in error_str:
+                raise InsufficientPermissionsError(
+                    "App lacks required permissions to read from Microsoft Teams."
+                )
+            raise ConnectorValidationError(
+                f"Unexpected error during Teams validation: {e}"
+            )
 
     def poll_source(self, start: SecondsSinceUnixEpoch, end: SecondsSinceUnixEpoch) -> Any:
         """Poll Microsoft Teams for recent messages"""
@@ -113,3 +177,27 @@ class TeamsConnector(CheckpointedConnectorWithPermSync, SlimConnectorWithPermSyn
         """Retrieve all simplified documents with permission sync"""
         # Simplified implementation
         return []
+    
+    def load_from_checkpoint_with_perm_sync(self, start, end, checkpoint):
+        return super().load_from_checkpoint_with_perm_sync(start, end, checkpoint)
+    
+
+    def _has_odata_incompatible_chars(self, team_names: list[str] | None) -> bool:
+        """Check if any team name contains characters that break Microsoft Graph OData filters.
+
+        The Microsoft Graph Teams API has limited OData support. Characters like
+        &, (, and ) cause parsing errors and require client-side filtering instead.
+        """
+        if not team_names:
+            return False
+        return any(char in name for name in team_names for char in ["&", "(", ")"])
+
+if __name__ == "__main__":
+    connector = TeamsConnector()
+    creds = {
+        "tenant_id": "",
+        "client_id": "",
+        "client_secret": ""
+    }
+    connector.load_credentials(creds)
+    connector.validate_connector_settings()

common/data_source/utils.py

@@ -1146,3 +1146,4 @@ def parallel_yield(gens: list[Iterator[R]], max_workers: int = 10) -> Iterator[R
                     future_to_index[executor.submit(_next_or_none, ind, gens[ind])] = next_ind
                     next_ind += 1
                 del future_to_index[future]
+