unstructured-ingest 1.0.4__tar.gz → 1.0.5__tar.gz

{unstructured_ingest-1.0.4 → unstructured_ingest-1.0.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: unstructured_ingest
-Version: 1.0.4
+Version: 1.0.5
 Summary: Local ETL data pipeline to get data RAG ready
 Author-email: Unstructured Technologies <devops@unstructuredai.io>
 License-Expression: Apache-2.0

unstructured_ingest-1.0.5/unstructured_ingest/__version__.py

@@ -0,0 +1 @@
+__version__ = "1.0.5"  # pragma: no cover

{unstructured_ingest-1.0.4 → unstructured_ingest-1.0.5}/unstructured_ingest/data_types/file_data.py

@@ -32,6 +32,28 @@ class FileDataSourceMetadata(BaseModel):
     permissions_data: Union[list[dict[str, Any]], dict[str, Any], None] = None
     filesize_bytes: Optional[int] = None
 
+    @field_validator("permissions_data", mode="before")
+    @classmethod
+    def coerce_permissions_data(cls, v: Any) -> Any:
+        if isinstance(v, dict):
+            # Temporarily convert dict to list for validation
+            return [v]
+        return v
+
+    @field_validator("permissions_data", mode="after")
+    @classmethod
+    def restore_dict_permissions_data(
+        cls, v: Optional[list[dict[str, Any]]]
+    ) -> Union[list[dict[str, Any]], dict[str, Any], None]:
+        if (
+            isinstance(v, list)
+            and len(v) == 1
+            and isinstance(v[0], dict)
+            and any(isinstance(val, dict) for val in v[0].values())
+        ):
+            return v[0]
+        return v
+
 
 class FileData(BaseModel):
     identifier: str

{unstructured_ingest-1.0.4 → unstructured_ingest-1.0.5}/unstructured_ingest/processes/connectors/confluence.py

@@ -352,6 +352,7 @@ class ConfluenceDownloader(Downloader):
     def _get_permissions_for_space(self, space_id: int) -> Optional[List[dict]]:
         if space_id in self._permissions_cache:
             self._permissions_cache.move_to_end(space_id)  # mark recent use
+            logger.debug(f"Retrieved cached permissions for space {space_id}")
             return self._permissions_cache[space_id]
         else:
             with self.connection_config.get_client() as client:
@@ -371,6 +372,7 @@ class ConfluenceDownloader(Downloader):
                         self._permissions_cache.popitem(last=False)  # LRU/FIFO eviction
                     self._permissions_cache[space_id] = space_permissions
 
+                    logger.debug(f"Retrieved permissions for space {space_id}")
                     return space_permissions
                 except Exception as e:
                     logger.debug(f"Could not retrieve permissions for space {space_id}: {e}")
@@ -387,6 +389,7 @@ class ConfluenceDownloader(Downloader):
                 logger.debug(f"Could not retrieve permissions for doc {doc_id}: {e}")
                 return None
 
+        logger.debug(f"normalized permissions generated: {parsed_permissions_dict}")
         return parsed_permissions_dict
 
     def run(self, file_data: FileData, **kwargs) -> download_responses:

{unstructured_ingest-1.0.4 → unstructured_ingest-1.0.5}/unstructured_ingest/processes/connectors/google_drive.py

@@ -409,8 +409,9 @@ class GoogleDriveIndexer(Indexer):
             d.metadata.record_locator["drive_id"]: object_id
         return data
 
-    def extract_permissions(self, permissions: list[dict]) -> dict:
+    def extract_permissions(self, permissions: Optional[list[dict]]) -> dict:
         if not permissions:
+            logger.debug("no permissions found")
             return {}
 
         # https://developers.google.com/workspace/drive/api/guides/ref-roles
@@ -442,6 +443,7 @@ class GoogleDriveIndexer(Indexer):
             for key in role_dict:
                 role_dict[key] = sorted(role_dict[key])
 
+        logger.debug(f"normalized permissions generated: {normalized_permissions}")
         return normalized_permissions
 
     def run(self, **kwargs: Any) -> Generator[FileData, None, None]:

unstructured_ingest-1.0.4/unstructured_ingest/__version__.py

@@ -1 +0,0 @@
-__version__ = "1.0.4"  # pragma: no cover