transcription

app.py

@@ -21,34 +21,42 @@ from transformers import pipeline
 from moviepy.editor import VideoFileClip
 import tempfile
 import os
+from pydub import AudioSegment
+from huggingface_hub import login
+
+with open("../../token.txt", "r") as file:
+    token = file.readline().strip()
+
+
+login(token=token, add_to_git_credential=True)
 
 pipe = pipeline(model="dacavi/whisper-small-hi")
 
 def transcribe_video(video_url):
     # Download video and extract audio
-    with tempfile.NamedTemporaryFile(suffix=".mp4", delete=False) as temp_video:
-        os.system(f"youtube-dl -o {temp_video.name} {video_url}")
-        video_clip = VideoFileClip(temp_video.name)
-        audio_clip = video_clip.audio
-        temp_audio_path = tempfile.NamedTemporaryFile(suffix=".wav", delete=False).name
-        audio_clip.write_audiofile(temp_audio_path, codec="wav")
+    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as temp_audio:
+        # os.system(f"yt-dlp -o {temp_audio.name} -x --audio-format wav {video_url}")
+        os.system(f"yt-dlp -o audioSample.wav -x --audio-format wav {video_url}")
+
+        print("Downloaded audio:", temp_audio.name)
+
 
     # Transcribe audio
-    text = pipe(temp_audio_path)["text"]
+        text = pipe("audioSample.wav")["text"]
 
     # Clean up temporary files
-    os.remove(temp_video.name)
-    os.remove(temp_audio_path)
-
-    return text
+        os.remove("audioSample.wav")
 
-iface = gr.Interface(
-    fn=transcribe_video,
-    inputs="text",
-    outputs="text",
-    live=True,
-    title="Video Transcription",
-    description="Paste the URL of a video to transcribe the spoken content.",
-)
 
-iface.launch()
+    return text
+print(transcribe_video("https://www.youtube.com/watch?v=8FkLRUJj-o0"))
+# iface = gr.Interface(
+#     fn=transcribe_video,
+#     inputs="text",
+#     outputs="text",
+#     live=True,
+#     title="Video Transcription",
+#     description="Paste the URL of a video to transcribe the spoken content.",
+# )
+#
+# iface.launch()

requirements.txt

@@ -2,4 +2,7 @@ transformers
 torch
 tensorflow
 moviepy==1.0.3
-ffmpeg
+ffmpeg
+ffprobe
+yt-dlp
+pydub