Update main.py

main.py

@@ -386,6 +386,43 @@ async def create_slideshow(request: Request):
 
         # Download audio file
         audio_file = await download_file(audio_url, ".mp3")
+        
+        # Pre-process audio file to ensure it's valid
+        normalized_audio = os.path.join(temp_dir, "normalized_audio.wav")
+        audio_check_cmd = (
+            f'ffmpeg -i {audio_file} -af "aformat=sample_fmts=fltp:sample_rates=44100:channel_layouts=stereo" '
+            f'-y {normalized_audio}'
+        )
+        audio_process = await asyncio.create_subprocess_shell(
+            audio_check_cmd,
+            stdout=asyncio.subprocess.PIPE,
+            stderr=asyncio.subprocess.PIPE
+        )
+        _, audio_stderr = await audio_process.communicate()
+
+        # If audio normalization fails, create a silent audio track
+        if audio_process.returncode != 0:
+            print(f"Audio normalization failed: {audio_stderr.decode()}")
+            print("Creating silent audio track instead...")
+            
+            # Calculate total video duration based on number of images and duration per image
+            total_duration = len(image_urls) * duration
+            
+            silent_cmd = (
+                f'ffmpeg -f lavfi -i anullsrc=r=44100:cl=stereo -t {total_duration} '
+                f'-y {normalized_audio}'
+            )
+            silent_process = await asyncio.create_subprocess_shell(
+                silent_cmd,
+                stdout=asyncio.subprocess.PIPE,
+                stderr=asyncio.subprocess.PIPE
+            )
+            _, silent_stderr = await silent_process.communicate()
+            
+            if silent_process.returncode != 0:
+                print(f"Silent audio creation failed: {silent_stderr.decode()}")
+                # Fall back to no audio if even silent audio fails
+                normalized_audio = None
 
         # Create a text file with the list of images and their duration
         images_list_path = os.path.join(temp_dir, "images_list.txt")
@@ -415,31 +452,57 @@ async def create_slideshow(request: Request):
             print(f"FFmpeg slideshow creation error: {stderr.decode()}")
             raise HTTPException(status_code=500, detail=f"FFmpeg slideshow creation failed: {stderr.decode()}")
 
-        # Add audio to the slideshow
-        # Increase analyzeduration and probesize for better audio detection
-        final_cmd = (
-            f'ffmpeg -analyzeduration 10000000 -probesize 10000000 -i {intermediate_video} '
-            f'-analyzeduration 10000000 -probesize 10000000 -i {audio_file} '
-            f'-c:v copy -c:a aac -strict experimental -shortest {output_path}'
-        )
-        process = await asyncio.create_subprocess_shell(
-            final_cmd,
-            stdout=asyncio.subprocess.PIPE,
-            stderr=asyncio.subprocess.PIPE
-        )
-        _, stderr = await process.communicate()
+        # Add audio to the slideshow only if normalized_audio is available
+        if normalized_audio:
+            final_cmd = (
+                f'ffmpeg -i {intermediate_video} -i {normalized_audio} '
+                f'-c:v copy -c:a aac -b:a 192k -shortest {output_path}'
+            )
+            process = await asyncio.create_subprocess_shell(
+                final_cmd,
+                stdout=asyncio.subprocess.PIPE,
+                stderr=asyncio.subprocess.PIPE
+            )
+            _, stderr = await process.communicate()
 
-        if process.returncode != 0:
-            print(f"FFmpeg audio addition error: {stderr.decode()}")
-            raise HTTPException(status_code=500, detail=f"FFmpeg audio addition failed: {stderr.decode()}")
+            if process.returncode != 0:
+                print(f"FFmpeg audio addition error: {stderr.decode()}")
+                # If adding audio fails, use just the video
+                shutil.copy(intermediate_video, output_path)
+        else:
+            # If no normalized audio available, use just the video
+            shutil.copy(intermediate_video, output_path)
 
         # Clean up temporary files
         for file in image_files:
-            os.remove(file)
-        os.remove(images_list_path)
-        os.remove(intermediate_video)
-        os.remove(audio_file)
-        os.rmdir(temp_dir)
+            if os.path.exists(file):
+                os.remove(file)
+        
+        if os.path.exists(images_list_path):
+            os.remove(images_list_path)
+        
+        if os.path.exists(intermediate_video):
+            os.remove(intermediate_video)
+            
+        if os.path.exists(audio_file):
+            os.remove(audio_file)
+            
+        if normalized_audio and os.path.exists(normalized_audio):
+            os.remove(normalized_audio)
+            
+        try:
+            os.rmdir(temp_dir)
+        except OSError:
+            # Directory might not be empty, try to clean up remaining files
+            for remaining_file in os.listdir(temp_dir):
+                try:
+                    os.remove(os.path.join(temp_dir, remaining_file))
+                except:
+                    pass
+            try:
+                os.rmdir(temp_dir)
+            except:
+                pass
 
         # Return the URL path to the output file
         return f"/output/{output_filename}"