Update app.py

app.py

@@ -2,6 +2,8 @@ from transformers import pipeline
 import gradio as gr
 from pytube import YouTube
 from datasets import Dataset, Audio
+import os
+from moviepy.editor import AudioFileClip
 
 pipe1 = pipeline(model="khalidey/ID2223_Lab2_Whisper_SV")  # change to "your-username/the-name-you-picked"
 pipe2 = pipeline('text-generation', model='birgermoell/swedish-gpt')
@@ -19,11 +21,21 @@ def youtube_link(url):
     path = streams.first().download()
     return path
 
+def convert_to_wav(path):
+    
+    sound = AudioFileClip(path)
+
+    segment = sound.subclip(0, int(sound.duration))
+    segment.write_audiofile(f"segment.wav")
+
+    return f"segment.wav"
+
 def youtube_transcribe(url):
     
     path = youtube_link(url)
+    path_wav = convert_to_wav(path)
     
-    audio_dataset = Dataset.from_dict({"audio": list(path)}).cast_column("audio", Audio(sampling_rate=16000))
+    audio_dataset = Dataset.from_dict({"audio": list(path_wav)}).cast_column("audio", Audio(sampling_rate=16000))
     text = pipe1(audio_dataset["audio"])["text"]
 
     return text