agazo commited on
Commit
e83c110
·
verified ·
1 Parent(s): ecf342c

Update audio_to_text_tool.py

Browse files
Files changed (1) hide show
  1. audio_to_text_tool.py +8 -6
audio_to_text_tool.py CHANGED
@@ -64,19 +64,17 @@ def audio_to_text_from_youtube(youtube_url: str) -> str:
64
  """
65
  try:
66
  with tempfile.TemporaryDirectory() as tmpdir:
67
- audio_output_path = os.path.join(tmpdir, "audio.mp3")
 
68
 
69
  BASE_DIR = os.path.dirname(os.path.abspath(__file__))
70
  cookies_path = os.path.join(BASE_DIR, 'files', 'cookies.txt')
71
 
72
- plain_txt_file = get_text_file_contents(cookies_path)
73
 
74
- print('cookies: \n' + plain_txt_file)
75
-
76
- # Download best audio using yt-dlp
77
  ydl_opts = {
78
  "format": "bestaudio/best",
79
- "outtmpl": audio_output_path,
80
  "quiet": True,
81
  "cookiefile": cookies_path,
82
  "postprocessors": [{
@@ -89,6 +87,10 @@ def audio_to_text_from_youtube(youtube_url: str) -> str:
89
  with yt_dlp.YoutubeDL(ydl_opts) as ydl:
90
  ydl.download([youtube_url])
91
 
 
 
 
 
92
  # Transcribe with OpenAI Whisper
93
  with open(audio_output_path, "rb") as audio_file:
94
  transcript = client.audio.transcriptions.create(
 
64
  """
65
  try:
66
  with tempfile.TemporaryDirectory() as tmpdir:
67
+ audio_stem = os.path.join(tmpdir, "audio")
68
+ audio_output_path = audio_stem + ".mp3"
69
 
70
  BASE_DIR = os.path.dirname(os.path.abspath(__file__))
71
  cookies_path = os.path.join(BASE_DIR, 'files', 'cookies.txt')
72
 
73
+ print('cookies: \n' + get_text_file_contents(cookies_path))
74
 
 
 
 
75
  ydl_opts = {
76
  "format": "bestaudio/best",
77
+ "outtmpl": audio_stem,
78
  "quiet": True,
79
  "cookiefile": cookies_path,
80
  "postprocessors": [{
 
87
  with yt_dlp.YoutubeDL(ydl_opts) as ydl:
88
  ydl.download([youtube_url])
89
 
90
+ # Ensure file exists
91
+ if not os.path.exists(audio_output_path):
92
+ raise FileNotFoundError(f"Audio file not created: {audio_output_path}")
93
+
94
  # Transcribe with OpenAI Whisper
95
  with open(audio_output_path, "rb") as audio_file:
96
  transcript = client.audio.transcriptions.create(