Update app.py

app.py

@@ -12,6 +12,7 @@ from IPython.display import Audio, display
 import requests
 nltk.download('punkt_tab')
 from io import BytesIO
+import ffmpeg
 
 model = whisper.load_model("base")
 
@@ -31,7 +32,12 @@ def extract_yt_audio(it, video_url, video_file):
     elif it == 'URL' and ("https://www" in video_url or "https://" in video_url or "www." in video_url):
       response = requests.get(video_url)
       video_data = BytesIO(response.content)
-      sample = AudioSegment.from_file(video_data, format="mp4")
+      output_stream = io.BytesIO()
+      stream = ffmpeg.input('pipe:0')
+      stream = ffmpeg.output(stream, 'pipe:1', format='wav')
+      ffmpeg.run(stream, input=input_stream.read(), stdout=output_stream)
+      output_stream.seek(0)
+      sample = AudioSegment.from_file(output_stream, format="mp4")
     elif it == 'URL':
       sample = AudioSegment.from_file(video_url)  
     else: