Fix file_id propagation through task chain

- Pass file_id as parameter through all task chains
- Update process_document, extract_metadata_with_gpt, embed_metadata_into_pdf, finalize_document_storage, send_to_all_destinations, rotate_pdf_pages, and process_with_azure_document_intelligence
- Remove unreliable LIKE queries, use explicit file_id passing instead

Co-authored-by: christianlouis <361235+christianlouis@users.noreply.github.com>

app/tasks/embed_metadata_into_pdf.py

@@ -46,7 +46,7 @@ def persist_metadata(metadata, final_pdf_path):
     return json_path
 
 @celery.task(base=BaseTaskWithRetry, bind=True)
-def embed_metadata_into_pdf(self, local_file_path: str, extracted_text: str, metadata: dict):
+def embed_metadata_into_pdf(self, local_file_path: str, extracted_text: str, metadata: dict, file_id: int = None):
     """
     Embeds extracted metadata into the PDF's standard metadata fields.
     The mapping is as follows:
@@ -62,11 +62,14 @@ def embed_metadata_into_pdf(self, local_file_path: str, extracted_text: str, met
     """
     task_id = self.request.id
     logger.info(f"[{task_id}] Starting metadata embedding for: {local_file_path}")
-    log_task_progress(task_id, "embed_metadata_into_pdf", "in_progress", f"Embedding metadata into {os.path.basename(local_file_path)}")
+    log_task_progress(task_id, "embed_metadata_into_pdf", "in_progress", f"Embedding metadata into {os.path.basename(local_file_path)}", file_id=file_id)
     
-    # Get file_id from database
+    # Get file_id from database if not provided
-    file_id = None
+    if file_id is None:
-    # Check for file existence; if not found, try the known shared tmp directory.
+        with SessionLocal() as db:
+            file_record = db.query(FileRecord).filter_by(local_filename=local_file_path).first()
+            if file_record:
+                file_id = file_record.id
     if not os.path.exists(local_file_path):
         alt_path = os.path.join(settings.workdir, "tmp", os.path.basename(local_file_path))
         if os.path.exists(alt_path):
@@ -76,10 +79,8 @@ def embed_metadata_into_pdf(self, local_file_path: str, extracted_text: str, met
             log_task_progress(task_id, "embed_metadata_into_pdf", "failure", "File not found")
             return {"error": "File not found"}
     
-    with SessionLocal() as db:
+    
-        file_record = db.query(FileRecord).filter_by(local_filename=local_file_path).first()
+    # Check for file existence; if not found, try the known shared tmp directory.
-        if file_record:
-            file_id = file_record.id
 
     # Work on a safe copy in a secure temporary directory
     original_file = local_file_path
@@ -149,7 +150,7 @@ def embed_metadata_into_pdf(self, local_file_path: str, extracted_text: str, met
         # Trigger the next step: final storage.
         logger.info(f"[{task_id}] Queueing final storage task")
         log_task_progress(task_id, "embed_metadata_into_pdf", "success", "Metadata embedded, queuing finalization", file_id=file_id)
-        finalize_document_storage.delay(original_file, final_file_path, metadata)
+        finalize_document_storage.delay(original_file, final_file_path, metadata, file_id)
 
         # After triggering final storage, delete the original file if it is in workdir/tmp.
         workdir_tmp = os.path.join(settings.workdir, "tmp")

app/tasks/extract_metadata_with_gpt.py

@@ -46,21 +46,21 @@ def extract_json_from_text(text):
     return None
 
 @celery.task(base=BaseTaskWithRetry, bind=True)
-def extract_metadata_with_gpt(self, filename: str, cleaned_text: str):
+def extract_metadata_with_gpt(self, filename: str, cleaned_text: str, file_id: int = None):
     """Uses OpenAI to classify document metadata."""
     task_id = self.request.id
     logger.info(f"[{task_id}] Starting metadata extraction for: {filename}")
-    log_task_progress(task_id, "extract_metadata_with_gpt", "in_progress", f"Extracting metadata for {filename}")
+    log_task_progress(task_id, "extract_metadata_with_gpt", "in_progress", f"Extracting metadata for {filename}", file_id=file_id)
     
-    # Get file_id from database
+    # Get file_id from database if not provided
-    file_id = None
+    if file_id is None:
-    tmp_dir = os.path.join(settings.workdir, "tmp")
+        tmp_dir = os.path.join(settings.workdir, "tmp")
-    file_path = os.path.join(tmp_dir, filename)
+        file_path = os.path.join(tmp_dir, filename)
-    if os.path.exists(file_path):
+        if os.path.exists(file_path):
-        with SessionLocal() as db:
+            with SessionLocal() as db:
-            file_record = db.query(FileRecord).filter_by(local_filename=file_path).first()
+                file_record = db.query(FileRecord).filter_by(local_filename=file_path).first()
-            if file_record:
+                if file_record:
-                file_id = file_record.id
+                    file_id = file_record.id
     
     prompt = f"""
 You are a specialized document analyzer trained to extract structured metadata from documents.
@@ -123,7 +123,7 @@ Return only valid JSON with no additional commentary.
         # Trigger the next step: embedding metadata into the PDF
         logger.info(f"[{task_id}] Queueing metadata embedding task")
         log_task_progress(task_id, "extract_metadata_with_gpt", "success", "Metadata extracted, queuing embed task", file_id=file_id)
-        embed_metadata_into_pdf.delay(filename, cleaned_text, metadata)
+        embed_metadata_into_pdf.delay(filename, cleaned_text, metadata, file_id)
 
         return {"s3_file": filename, "metadata": metadata}
 

app/tasks/finalize_document_storage.py

@@ -17,29 +17,29 @@ logger = logging.getLogger(__name__)
 
 
 @celery.task(base=BaseTaskWithRetry, bind=True)
-def finalize_document_storage(self, original_file: str, processed_file: str, metadata: dict):
+def finalize_document_storage(self, original_file: str, processed_file: str, metadata: dict, file_id: int = None):
     """
     Final storage step after embedding metadata.
     We will now call 'send_to_all_destinations' to push the final PDF to Dropbox/Nextcloud/Paperless.
     """
     task_id = self.request.id
     logger.info(f"[{task_id}] Finalizing document storage for {processed_file}")
-    log_task_progress(task_id, "finalize_document_storage", "in_progress", f"Finalizing: {os.path.basename(processed_file)}")
+    log_task_progress(task_id, "finalize_document_storage", "in_progress", f"Finalizing: {os.path.basename(processed_file)}", file_id=file_id)
     
-    # Get file_id from database
+    # Get file_id from database if not provided
-    file_id = None
+    if file_id is None:
-    with SessionLocal() as db:
+        with SessionLocal() as db:
-        # Try to find by the processed file path first
+            # Try to find by the original file path
-        file_record = db.query(FileRecord).filter(
+            file_record = db.query(FileRecord).filter(
-            FileRecord.local_filename.like(f"%{os.path.basename(original_file)}%")
+                FileRecord.local_filename.like(f"%{os.path.basename(original_file)}%")
-        ).first()
+            ).first()
-        if file_record:
+            if file_record:
-            file_id = file_record.id
+                file_id = file_record.id
 
     # 2) Enqueue uploads to all destinations (Dropbox, Nextcloud, Paperless)
     logger.info(f"[{task_id}] Queueing uploads to all destinations")
     log_task_progress(task_id, "finalize_document_storage", "success", "Queuing uploads to destinations", file_id=file_id)
-    send_to_all_destinations.delay(processed_file)
+    send_to_all_destinations.delay(processed_file, True, file_id)
 
     return {
         "status": "Completed",

app/tasks/process_document.py

@@ -130,12 +130,12 @@ def process_document(self, original_local_file: str):
         # Call metadata extraction directly
         logger.info(f"[{task_id}] Queueing metadata extraction")
         log_task_progress(task_id, "process_document", "success", "Queued for metadata extraction", file_id=new_record.id)
-        extract_metadata_with_gpt.delay(new_filename, extracted_text)
+        extract_metadata_with_gpt.delay(new_filename, extracted_text, new_record.id)
         return {"file": new_local_path, "status": "Text extracted locally", "file_id": new_record.id}
 
     # 3. If no embedded text, queue Azure Document Intelligence processing
     logger.info(f"[{task_id}] No embedded text found. Queueing Azure Document Intelligence processing")
     log_task_progress(task_id, "check_text", "success", "No embedded text, queuing OCR", file_id=new_record.id)
     log_task_progress(task_id, "process_document", "success", "Queued for OCR processing", file_id=new_record.id)
-    process_with_azure_document_intelligence.delay(new_filename)
+    process_with_azure_document_intelligence.delay(new_filename, new_record.id)
     return {"file": new_local_path, "status": "Queued for OCR", "file_id": new_record.id}

app/tasks/process_with_azure_document_intelligence.py

@@ -76,7 +76,7 @@ def check_page_rotation(result, filename):
     return rotation_data
 
 @celery.task(base=BaseTaskWithRetry)
-def process_with_azure_document_intelligence(filename: str):
+def process_with_azure_document_intelligence(filename: str, file_id: int = None):
     """
     Processes a PDF document using Azure Document Intelligence and overlays OCR text onto
     the local temporary file (stored under <workdir>/tmp).
@@ -88,6 +88,10 @@ def process_with_azure_document_intelligence(filename: str):
       3. Saves the OCR-processed PDF locally in the same location as before.
       4. Checks for page rotation and triggers page rotation if needed.
       5. Triggers downstream metadata extraction.
+      
+    Args:
+        filename: Name of the file to process
+        file_id: Optional file ID to pass through to subsequent tasks
     """
     try:
         tmp_file_path = os.path.join(settings.workdir, "tmp", filename)
@@ -139,7 +143,7 @@ def process_with_azure_document_intelligence(filename: str):
         logger.info(f"Extracted text for {filename}: {len(extracted_text)} characters")
 
         # Trigger page rotation task if rotation is detected, otherwise proceed to metadata extraction
-        rotate_pdf_pages.delay(filename, extracted_text, rotation_data)
+        rotate_pdf_pages.delay(filename, extracted_text, rotation_data, file_id)
 
         return {"file": filename, "searchable_pdf": searchable_pdf_path, "cleaned_text": extracted_text}
     except Exception as e:

app/tasks/rotate_pdf_pages.py

@@ -47,7 +47,7 @@ def determine_rotation_angle(detected_angle):
     return rotation_value
 
 @celery.task(base=BaseTaskWithRetry)
-def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
+def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None, file_id: int = None):
     """
     Rotates pages in a PDF document based on detected rotation angles.
     
@@ -55,6 +55,7 @@ def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
         filename: The name of the file to rotate
         extracted_text: The extracted text from the document
         rotation_data: Optional rotation data dictionary {page_index: angle}
+        file_id: Optional file ID to pass through to subsequent tasks
     """
     try:
         pdf_path = os.path.join(settings.workdir, "tmp", filename)
@@ -64,7 +65,7 @@ def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
         # Skip rotation if no rotation data provided
         if not rotation_data:
             logger.info(f"No rotation data provided for {filename}, proceeding with metadata extraction")
-            extract_metadata_with_gpt.delay(filename, extracted_text)
+            extract_metadata_with_gpt.delay(filename, extracted_text, file_id)
             return {"file": filename, "status": "no_rotation_needed"}
 
         # Standardize rotation_data keys to integers
@@ -77,7 +78,7 @@ def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
 
         if not any(abs(angle) > 0 for angle in normalized_rotation_data.values()):
             logger.info(f"No significant rotations detected in {filename}, proceeding with metadata extraction")
-            extract_metadata_with_gpt.delay(filename, extracted_text)
+            extract_metadata_with_gpt.delay(filename, extracted_text, file_id)
             return {"file": filename, "status": "no_rotation_needed"}
         
         logger.info(f"Rotating {len(normalized_rotation_data)} pages in {filename}")
@@ -117,7 +118,7 @@ def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
             logger.info(f"Detected rotations in {filename} but no rotations were actually applied (angles too small or not multiples of 90°)")
         
         # Continue with metadata extraction
-        extract_metadata_with_gpt.delay(filename, extracted_text)
+        extract_metadata_with_gpt.delay(filename, extracted_text, file_id)
         
         return {
             "file": filename, 
@@ -129,5 +130,5 @@ def rotate_pdf_pages(filename: str, extracted_text: str, rotation_data=None):
     except Exception as e:
         logger.error(f"Error rotating PDF {filename}: {e}")
         # Continue with metadata extraction despite rotation failure
-        extract_metadata_with_gpt.delay(filename, extracted_text)
+        extract_metadata_with_gpt.delay(filename, extracted_text, file_id)
         return {"file": filename, "status": "rotation_failed", "error": str(e)}

app/tasks/send_to_all.py

@@ -108,7 +108,7 @@ def get_configured_services_from_validator():
     return result
 
 @celery.task(base=BaseTaskWithRetry, bind=True)
-def send_to_all_destinations(self, file_path: str, use_validator=True):
+def send_to_all_destinations(self, file_path: str, use_validator=True, file_id: int = None):
     """
     Distribute a file to all configured storage destinations.
     
@@ -116,25 +116,26 @@ def send_to_all_destinations(self, file_path: str, use_validator=True):
         file_path: Path to the file to distribute
         use_validator: Whether to use the config validator to determine enabled services
                        (if False, falls back to individual checks)
+        file_id: Optional file ID to associate with logs
     """
     task_id = self.request.id
     
     if not os.path.exists(file_path):
         logger.error(f"[{task_id}] File not found: {file_path}")
-        log_task_progress(task_id, "send_to_all_destinations", "failure", "File not found")
+        log_task_progress(task_id, "send_to_all_destinations", "failure", "File not found", file_id=file_id)
         raise FileNotFoundError(f"File not found: {file_path}")
     
     logger.info(f"[{task_id}] Sending {file_path} to all configured destinations")
-    log_task_progress(task_id, "send_to_all_destinations", "in_progress", f"Distributing: {os.path.basename(file_path)}")
+    log_task_progress(task_id, "send_to_all_destinations", "in_progress", f"Distributing: {os.path.basename(file_path)}", file_id=file_id)
     
-    # Get file_id from database
+    # Get file_id from database if not provided
-    file_id = None
+    if file_id is None:
-    with SessionLocal() as db:
+        with SessionLocal() as db:
-        file_record = db.query(FileRecord).filter(
+            file_record = db.query(FileRecord).filter(
-            FileRecord.local_filename.like(f"%{os.path.basename(file_path)}%")
+                FileRecord.local_filename.like(f"%{os.path.basename(file_path)}%")
-        ).first()
+            ).first()
-        if file_record:
+            if file_record:
-            file_id = file_record.id
+                file_id = file_record.id
     
     results = {}