Update app.py

app.py

@@ -130,14 +130,13 @@ def step_1_upload_and_process():
             reset_session()
             st.rerun()
 
-# --- Audio Processing Function (with robust stateless logic) ---
 def process_audio():
     if not st.session_state.audio_file:
         st.error("No audio file found.")
         return
     
     tmp_file_path = None
-    log_to_browser_console("--- INFO: Starting audio processing. ---")
+    log_to_browser_console("--- INFO: Starting audio processing flow. ---")
     try:
         with tempfile.NamedTemporaryFile(delete=False, suffix=Path(st.session_state.audio_file.name).suffix) as tmp_file:
             st.session_state.audio_file.seek(0)
@@ -145,55 +144,43 @@ def process_audio():
             tmp_file_path = tmp_file.name
         
         processor = AUDIO_PROCESSOR_CLASS()
-        with st.spinner("🎤 Transcribing audio..."):
-            transcription_result = processor.transcribe_audio(tmp_file_path)
-        
-        if "Error:" in transcription_result or not transcription_result:
-            st.error(f"Transcription failed: {transcription_result}")
-            return
 
-        # --- AGGRESSIVE DEBUGGING FOR TIMESTAMPS ---
-        log_to_browser_console("--- DEBUG: Attempting to extract word timestamps... ---")
         word_timestamps = []
-        try:
-            with st.spinner("🔍 Extracting timestamps..."):
-                if hasattr(processor, 'get_word_timestamps'):
-                    word_timestamps = processor.get_word_timestamps(tmp_file_path)
-                else:
-                    log_to_browser_console("--- WARNING: processor object has no 'get_word_timestamps' method. ---")
-            
-            # This is the most important log. It will tell us the truth.
-            log_to_browser_console(f"--- CRITICAL: Extracted {len(word_timestamps)} timestamps. ---")
-            
-            if not word_timestamps:
-                st.warning("CRITICAL WARNING: No word timestamps were extracted from the audio processor. The list is empty. SYLT embedding will fail.")
-                log_to_browser_console("--- ERROR: The list of word_timestamps received from audio_processor is EMPTY. ---")
-            else:
-                # Log the first 3 entries to see if the data format is correct
-                log_to_browser_console(f"--- DEBUG: Sample of extracted timestamps: {json.dumps(word_timestamps[:3])} ---")
+        # The audio_processor now returns both data and logs
+        with st.spinner("🎤 Performing AI processing (Transcription & Timestamps)..."):
+            word_timestamps, processor_logs = processor.get_word_timestamps(tmp_file_path)
+        
+        # Log everything that happened inside the processor to the browser
+        log_to_browser_console(processor_logs)
 
-        except Exception as e:
-            st.error(f"A fatal error occurred inside 'get_word_timestamps': {e}")
-            log_to_browser_console(f"--- FATAL ERROR during get_word_timestamps call: {traceback.format_exc()} ---")
-            word_timestamps = [] # Ensure it's an empty list on failure
-        # --- END OF DEBUGGING BLOCK ---
+        # Check if the final result is empty and show a warning
+        if not word_timestamps:
+            st.warning("Could not generate timestamps. Check the browser console (F12) for detailed error logs.")
+        
+        # We still need the full text for the editor, even if timestamps fail
+        # This is a simplified way; a better way would be for get_word_timestamps to also return the text
+        full_text = " ".join([d['word'] for d in word_timestamps])
+        if not full_text and processor_logs:
+             # Try to find a fallback text from logs if available
+             full_text = "Transcription failed. Please check logs and edit text manually."
 
         st.session_state.audio_file.seek(0)
         audio_bytes = st.session_state.audio_file.read()
 
         st.session_state.transcription_data = {
-            'text': transcription_result,
+            'text': full_text, # Use the text we derived
             'word_timestamps': word_timestamps,
             'audio_bytes': audio_bytes,
             'original_suffix': Path(st.session_state.audio_file.name).suffix
         }
-        st.session_state.edited_text = transcription_result
+        st.session_state.edited_text = full_text
         st.session_state.step = 2
-        st.success("🎉 Audio processing complete!")
+        st.success("🎉 AI processing complete! Please review the results.")
         
     except Exception as e:
-        st.error(f"An unexpected error occurred during the main processing flow!")
+        st.error("An unexpected error occurred in the main processing flow!")
         st.exception(e)
+        log_to_browser_console(f"--- FATAL ERROR in app.py's process_audio: {traceback.format_exc()} ---")
     finally:
         if tmp_file_path and os.path.exists(tmp_file_path):
             os.unlink(tmp_file_path)
@@ -202,6 +189,7 @@ def process_audio():
     st.rerun()
 
 
+
 # --- Step 2: Review and Customize ---
 def step_2_review_and_customize():
     st.header("Step 2: Review & Customize")