feat: Added API for deletion of processed files

Author: Prajwal-Microsoft

src/ContentProcessorAPI/app/libs/cosmos_db/helper.py

@@ -102,6 +102,7 @@ def update_document_by_query(self, query: Dict[str, Any], update: Dict[str, Any]
         result = self.container.update_one(query, {"$set": update})
         return result
 
-    def delete_document(self, item_id: str):
-        result = self.container.delete_one({"Id": item_id})
+    def delete_document(self, item_id: str, field_name: str = None):
+        field_name = field_name or "Id"  # Use "Id" if field_name is empty or None
+        result = self.container.delete_one({field_name: item_id})
         return result

src/ContentProcessorAPI/app/libs/storage_blob/helper.py

@@ -90,3 +90,24 @@ def delete_blob_and_cleanup(self, blob_name, container_name=None):
             # Delete the (virtual) folder in the Container
             blob_client = container_client.get_blob_client(container_name)
             blob_client.delete_blob()
+
+    def delete_folder(self, folder_name, container_name=None):
+        container_client = self._get_container_client(container_name)
+        
+        # List all blobs inside the folder
+        blobs_to_delete = container_client.list_blobs(name_starts_with=folder_name + "/")
+
+        # Delete each blob
+        for blob in blobs_to_delete:
+            blob_client = container_client.get_blob_client(blob.name)
+            blob_client.delete_blob()
+        
+        blobs_to_delete = container_client.list_blobs()
+        if not blobs_to_delete:
+
+            # Get Parent Container
+            container_client = self._get_container_client()
+
+            # Delete the (virtual) folder in the Container
+            blob_client = container_client.get_blob_client(folder_name)
+            blob_client.delete_blob()

src/ContentProcessorAPI/app/routers/contentprocessor.py

@@ -6,7 +6,7 @@
 import urllib.parse
 import uuid
 
-from fastapi import APIRouter, Body, Depends, File, UploadFile
+from fastapi import APIRouter, Body, Depends, File, HTTPException, UploadFile
 from fastapi.responses import JSONResponse, StreamingResponse
 from pymongo.results import UpdateResult
 
@@ -28,6 +28,7 @@
     ContentProcess,
     ContentProcessorRequest,
     ContentResultUpdate,
+    ContentResultDelete,
     Paging,
     ProcessFile,
     Status,
@@ -490,3 +491,31 @@ async def get_original_file(
         return StreamingResponse(
             file_stream, media_type=content_type_string, headers=headers
         )
+
+@router.delete(
+    "/processed/{process_id}",
+    response_model=ContentResultDelete,
+    summary="Delete the processed content result",
+    description="""
+            Returns the deleted record for a given process ID.
+            """,
+)
+async def delete_processed_file(
+    process_id: str, app_config: AppConfiguration = Depends(get_app_config)
+    ) -> ContentResultDelete:
+    try:
+        deleted_file =  CosmosContentProcess(process_id=process_id).delete_processed_file(
+            connection_string=app_config.app_cosmos_connstr,
+            database_name=app_config.app_cosmos_database,
+            collection_name=app_config.app_cosmos_container_process,
+            storage_connection_string=app_config.app_storage_blob_url,
+            container_name=app_config.app_cps_processes,
+        )
+
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+    return ContentResultDelete(
+        status="Success" if deleted_file else "Failed",
+        process_id=deleted_file.process_id if deleted_file else "",
+        message="" if deleted_file else "This record no longer exists. Please refresh the page."
+    )

src/ContentProcessorAPI/app/routers/models/contentprocessor/content_process.py

@@ -189,6 +189,41 @@ def get_status_from_cosmos(
             return ContentProcess(**existing_process[0])
         else:
             return None
+    
+    def delete_processed_file(
+        self,
+        connection_string: str,
+        database_name: str,
+        collection_name: str,
+        storage_connection_string: str,
+        container_name: str,
+    ):
+        """
+        Delete the processed file from Cosmos DB & Storage account.
+        """
+        mongo_helper = CosmosMongDBHelper(
+            connection_string=connection_string,
+            db_name=database_name,
+            container_name=collection_name,
+            indexes=[("process_id", 1)],
+        )
+
+        blob_helper = StorageBlobHelper(
+            account_url=storage_connection_string, container_name=container_name
+        )
+
+        # Check if the process_id already exists in the database
+        existing_process = mongo_helper.find_document(
+            query={"process_id": self.process_id}
+        )
+
+        blob_helper.delete_folder(folder_name=self.process_id)
+
+        if existing_process:
+            mongo_helper.delete_document(item_id=self.process_id, field_name="process_id")
+            return ContentProcess(**existing_process[0])
+        else:
+            return None
 
     def update_process_result(
         self,

src/ContentProcessorAPI/app/routers/models/contentprocessor/model.py

@@ -64,6 +64,11 @@ class ContentResultUpdate(BaseModel):
     process_id: str
     modified_result: dict
 
+class ContentResultDelete(BaseModel):
+    process_id: str
+    status: str
+    message: str
+
 
 class ContentCommentUpdate(BaseModel):
     process_id: str