Update app.py

app.py

@@ -39,6 +39,7 @@ import librosa
 import noisereduce as nr
 import soundfile as sf
 
+
 logger = logging.getLogger(__name__)
 
 # Configure logging
@@ -223,16 +224,21 @@ def transcribe_video_with_speakers(video_path):
         speaker = segment["speaker"]
         end = segment["end"]
         start = segment["start"]
-        if speaker not in speaker_audio:
-            speaker_audio[speaker] = []
         if end > start and (end - start) > 0.05:  # Require >50ms duration
-            speaker_audio[speaker].append((segment["start"], segment["end"]))
-
+            if speaker not in speaker_audio:
+                speaker_audio[speaker] = [(segment["start"], segment["end"])]
+            else:
+                speaker_audio[speaker].append((segment["start"], segment["end"]))
     # Collapse and truncate speaker audio
     speaker_sample_paths = {}
     audio_clip = AudioFileClip(speech_audio_path)
     for speaker, segments in speaker_audio.items():
         speaker_clips = [audio_clip.subclip(start, end) for start, end in segments]
+        # Add a check to ensure speaker_clips is not empty
+        if not speaker_clips:
+            logger.warning(f"No valid audio segments found for speaker {speaker} meeting the duration requirement. Skipping sample creation.")
+            continue # Skip the rest of the loop for this speaker
+
         combined_clip = concatenate_audioclips(speaker_clips)
         truncated_clip = combined_clip.subclip(0, min(30, combined_clip.duration))