Refactor podcast segment regeneration to remove pydub dependency. Directly concatenate MP3 files for improved performance and simplicity.

app.py

@@ -19,7 +19,6 @@ from pathlib import Path
 from main import NotebookMg
 from dotenv import load_dotenv
 import logging
-from pydub import AudioSegment
 
 # Set up logging
 logging.basicConfig(level=logging.DEBUG)
@@ -256,7 +255,6 @@ async def regenerate_segment(
                 logger.info(f"Deleted existing segment file: {segment_path}")
             except Exception as e:
                 logger.error(f"Error deleting existing segment file: {str(e)}")
-                # Continue anyway as we'll overwrite the file
 
         # Save the regenerated segment
         audio_bytes = b"".join(audio_data)
@@ -273,30 +271,22 @@ async def regenerate_segment(
             except Exception as e:
                 logger.error(f"Error deleting existing podcast file: {str(e)}")
 
-        # Combine all segments into a new complete podcast
-        combined_audio = AudioSegment.empty()
+        # Instead of using pydub, we'll concatenate the MP3 files directly
+        with open(podcast_path, "wb") as outfile:
+            # Get all segment files and sort them correctly
+            segment_files = sorted(
+                [f for f in OUTPUT_DIR.glob(f"{base_name}_segment_*.mp3")],
+                key=lambda x: int(x.stem.split("_")[-1]),
+            )
 
-        # Get all segment files and sort them correctly
-        segment_files = sorted(
-            [f for f in OUTPUT_DIR.glob(f"{base_name}_segment_*.mp3")],
-            key=lambda x: int(x.stem.split("_")[-1]),
-        )
-
-        logger.info(f"Found {len(segment_files)} segments to combine")
+            logger.info(f"Found {len(segment_files)} segments to combine")
 
-        # Add each segment to the combined audio
-        for segment_file in segment_files:
-            try:
-                logger.info(f"Processing segment: {segment_file}")
-                audio_segment = AudioSegment.from_file(segment_file, format="mp3")
-                pause = AudioSegment.silent(duration=300)  # 300ms pause
-                combined_audio += audio_segment + pause
-            except Exception as e:
-                logger.error(f"Error processing segment {segment_file}: {str(e)}")
-                raise
+            # Concatenate all MP3 files
+            for segment_file in segment_files:
+                with open(segment_file, "rb") as infile:
+                    outfile.write(infile.read())
+                    # No pause between segments in this simple approach
 
-        # Save the new complete podcast
-        combined_audio.export(str(podcast_path), format="mp3")
         logger.info(
             f"Successfully generated new podcast with {len(segment_files)} segments"
         )