try to chunk into pieces

app.py

@@ -5,6 +5,8 @@ from video_downloader import download_video
 from moviepy.editor import AudioFileClip
 import datetime
 import os
+from pydub import AudioSegment
+from pydub.silence import split_on_silence
 
 pipe = pipeline("automatic-speech-recognition", model="Artanis1551/whisper_romanian3")
 
@@ -17,10 +19,18 @@ def process_video(date):
     audio_path = f"audio_{date}.wav"
     AudioFileClip(video_path).write_audiofile(audio_path)
 
-    # Transcribe the audio
-    with open(audio_path, "rb") as audio_file:
-        audio = audio_file.read()
-    transcription = pipe(audio)["text"]
+    # Split the audio into chunks
+    audio = AudioSegment.from_wav(audio_path)
+    chunks = split_on_silence(audio, min_silence_len=500, silence_thresh=-40)
+
+    # Transcribe each chunk
+    transcription = ""
+    for i, chunk in enumerate(chunks):
+        chunk.export(f"chunk{i}.wav", format="wav")
+        with open(f"chunk{i}.wav", "rb") as audio_file:
+            audio = audio_file.read()
+        transcription += pipe(audio)["text"] + "\n "
+        os.remove(f"chunk{i}.wav")
 
     # Remove the audio file
     os.remove(audio_path)

requirements.txt

@@ -6,4 +6,5 @@ gradio
 transformers
 torch
 urllib3
-moviepy
+moviepy
+pydub