Update app.py

app.py

@@ -183,8 +183,8 @@ def download_state_from_api() -> Dict[str, Any]:
             state_data["file_states"] = {}
         if "next_download_index" not in state_data:
             state_data["next_download_index"] = 0
-            
-        log_message(f"✅ Successfully downloaded state from API.", "INFO")
+        
+        log_message(f"✅ Downloaded state: next_download_index={state_data['next_download_index']}, processed_files={len([f for f,s in state_data['file_states'].items() if s=='processed'])}", "INFO")
         return state_data
         
     except requests.exceptions.RequestException as e:
@@ -214,11 +214,11 @@ def upload_state_to_api(state: Dict[str, Any]) -> bool:
             return True
             
     except requests.exceptions.HTTPError as e:
-            if hasattr(e, 'response') and e.response.status_code == 409:
-                log_message(f"⚠️ State file already exists on server (409 Conflict) - Lock failed.", "WARNING")
-                return False
-            log_message(f"❌ Failed to upload state file to API ({url}): {str(e)}", "ERROR")
-            return False
+        if hasattr(e, 'response') and e.response.status_code == 409:
+            log_message(f"⚠️ State file already exists on server (409 Conflict) - Treating as success.", "INFO")
+            return True
+        log_message(f"❌ Failed to upload state file to API ({url}): {str(e)}", "ERROR")
+        return False
     except requests.exceptions.RequestException as e:
         log_message(f"❌ Failed to upload state file to API ({url}): {str(e)}", "ERROR")
         return False
@@ -269,12 +269,12 @@ def lock_file_for_processing(wav_filename: str, state: Dict[str, Any]) -> bool:
             del state["file_states"][wav_filename]
         return False
 
-def unlock_file_as_processed(wav_filename: str, state: Dict[str, Any]) -> bool:
+def unlock_file_as_processed(wav_filename: str, state: Dict[str, Any], next_index: int) -> bool:
     """Marks a file as 'processed', updates the index, and uploads the state via API."""
     log_message(f"🔓 Attempting to unlock file: {wav_filename} (Marking as 'processed')", "INFO")
     
     state["file_states"][wav_filename] = "processed"
-
+    state["next_download_index"] = next_index
     
     if upload_state_to_api(state):
         log_message(f"✅ Successfully unlocked and marked as processed: {wav_filename} via API state upload", "INFO")
@@ -544,9 +544,7 @@ def process_audio_file(audio_path: str, reference_map: Dict[str, str], output_fi
         return False
         
     # 3. Upload transcription to API
-    upload_success = upload_transcription_to_api(final_json_path, final_json_filename)
-    
-    if upload_success:
+    if upload_transcription_to_api(final_json_path, final_json_filename):
         processing_status["transcribed_files"] += 1
         # Clean up the local transcription file after successful upload
         try:
@@ -581,7 +579,8 @@ def main_processing_loop():
         while processing_status["is_running"]:
             time.sleep(PROCESSING_DELAY)
             
-            # 1. Download state from the new API
+            # 1. Download FRESH state from the API at the start of each iteration
+            # This ensures we respect the next_download_index that other workers may have set
             current_state = download_state_from_api()
             next_file_info = get_next_file_to_process(SOURCE_REPO_ID, current_state)
             
@@ -601,10 +600,9 @@ def main_processing_loop():
             try:
                 # 2. Lock file by updating state on the API
                 # IMPORTANT: Update next_download_index when locking to prevent other workers from picking same file
-
-                
-                # The index should be incremented before the lock attempt to prevent other workers from picking it up
+                old_index = current_state["next_download_index"]
                 current_state["next_download_index"] = target_index + 1
+                log_message(f"📍 Incrementing next_download_index from {old_index} to {current_state['next_download_index']}", "INFO")
                 
                 if not lock_file_for_processing(target_file, current_state):
                     log_message(f"❌ Failed to lock file {target_file}. Skipping.", "ERROR")
@@ -638,18 +636,17 @@ def main_processing_loop():
                 
             finally:
                 # 4. Unlock/Mark as processed by updating state on the API
-                # IMPORTANT: Don't re-download the state here because it will lose the incremented index!
-                # Instead, use the current_state which already has the correct next_download_index
+                # IMPORTANT: Keep the incremented next_download_index from locking
                 
                 if success:
-                    # Mark as processed and upload state (index was already updated before lock)
-                    unlock_file_as_processed(target_file, current_state)
+                    # Mark as processed and keep the incremented index, then upload state
+                    unlock_file_as_processed(target_file, current_state, current_state["next_download_index"])
                     processing_status["processed_files"] += 1
                 else:
                     # Mark as failed but keep the incremented index so next worker can proceed
                     log_message(f"⚠️ File {target_file} failed. Marking as 'failed' and updating state.", "WARNING")
                     current_state["file_states"][target_file] = "failed"
-                    # Don't overwrite next_download_index here - keep it at the incremented value
+                    # Keep the incremented next_download_index - don't change it
                     upload_state_to_api(current_state)
                 
                 # Clean up the downloaded audio file regardless of success