Update app.py

app.py

@@ -3,90 +3,141 @@ import whisper
 import os
 import yt_dlp
 
-# Load Whisper model
-model = whisper.load_model("base")
-
-# Load summarization model
-summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+# Load smaller Whisper model to save memory
+print("Loading Whisper model...")
+model = whisper.load_model("tiny")
 
 def download_youtube_audio(url):
-    yt = YouTube(url)
-    audio = yt.streams.filter(only_audio=True).first()
-    out_file = audio.download(filename="temp_audio")
-    base, ext = os.path.splitext(out_file)
-    audio_file = base + '.mp3'
-    os.rename(out_file, audio_file)
-    return audio_file
+    """Download audio from YouTube using yt-dlp"""
+    try:
+        ydl_opts = {
+            'format': 'bestaudio/best',
+            'postprocessors': [{
+                'key': 'FFmpegExtractAudio',
+                'preferredcodec': 'mp3',
+                'preferredquality': '192',
+            }],
+            'outtmpl': 'temp_audio.%(ext)s',
+            'quiet': True,
+        }
+        
+        with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+            ydl.download([url])
+        
+        return "temp_audio.mp3"
+    except Exception as e:
+        raise Exception(f"Failed to download YouTube video: {str(e)}")
 
-def transcribe_and_summarize(youtube_url=None, video_file=None):
+def simple_summarize(text, max_sentences=5):
+    """Simple extractive summary"""
+    sentences = text.replace('!', '.').replace('?', '.').split('.')
+    sentences = [s.strip() for s in sentences if len(s.strip()) > 20]
+    
+    if len(sentences) <= max_sentences:
+        return '. '.join(sentences) + '.'
+    
+    summary_sentences = [sentences[0]]
+    step = len(sentences) // (max_sentences - 1)
+    
+    for i in range(1, max_sentences):
+        idx = min(i * step, len(sentences) - 1)
+        summary_sentences.append(sentences[idx])
+    
+    return '. '.join(summary_sentences) + '.'
+
+def transcribe_video(youtube_url, video_file, progress=gr.Progress()):
+    """Main transcription function"""
+    audio_path = None
+    
     try:
-        # Get audio file
-        if youtube_url:
+        if youtube_url and youtube_url.strip():
+            progress(0.2, desc="Downloading YouTube audio...")
             audio_path = download_youtube_audio(youtube_url)
-        elif video_file:
+        elif video_file is not None:
+            progress(0.2, desc="Processing uploaded video...")
             audio_path = video_file
         else:
-            return "Please provide a YouTube URL or upload a video file."
+            return "❌ Error", "Please provide a YouTube URL or upload a video file.", "", None
         
-        # Transcribe
-        result = model.transcribe(audio_path)
+        progress(0.5, desc="Transcribing audio...")
+        result = model.transcribe(audio_path, fp16=False)
         transcription = result["text"]
         
-        # Summarize (split into chunks if too long)
-        max_chunk = 1024
-        text_chunks = [transcription[i:i+max_chunk] for i in range(0, len(transcription), max_chunk)]
-        summaries = []
+        if not transcription or len(transcription) < 10:
+            return "❌ Error", "Transcription failed or audio had no speech.", "", None
         
-        for chunk in text_chunks[:3]:  # Limit to first 3 chunks
-            summary = summarizer(chunk, max_length=130, min_length=30, do_sample=False)
-            summaries.append(summary[0]['summary_text'])
+        progress(0.8, desc="Generating summary...")
+        summary = simple_summarize(transcription, max_sentences=5)
         
-        summary = " ".join(summaries)
+        progress(0.9, desc="Creating download file...")
+        output_text = f"""VIDEO TRANSCRIPTION & SUMMARY
+{"="*60}
+
+FULL TRANSCRIPTION:
+{transcription}
+
+{"="*60}
+
+SUMMARY:
+{summary}
+
+{"="*60}
+Generated by Video Transcription App
+"""
         
-        # Create downloadable text file
-        output_text = f"TRANSCRIPTION:\n{'='*50}\n\n{transcription}\n\n\nSUMMARY:\n{'='*50}\n\n{summary}"
+        output_file = "transcript_output.txt"
+        with open(output_file, "w", encoding="utf-8") as f:
+            f.write(output_text)
         
-        # Clean up
-        if youtube_url and os.path.exists(audio_path):
-            os.remove(audio_path)
+        if youtube_url and audio_path and os.path.exists(audio_path):
+            try:
+                os.remove(audio_path)
+            except:
+                pass
         
-        return transcription, summary, output_text
+        progress(1.0, desc="Done!")
+        return "✅ Success", transcription, summary, output_file
         
     except Exception as e:
-        return f"Error: {str(e)}", "", ""
+        error_msg = f"Error: {str(e)}"
+        print(error_msg)
+        return "❌ Error", error_msg, "", None
 
-# Create Gradio interface
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🎥 Video Transcription & Summary Generator")
-    gr.Markdown("Upload a video or paste a YouTube link to get AI-powered transcription and summary")
-    
-    with gr.Tab("YouTube Link"):
-        youtube_input = gr.Textbox(label="YouTube URL", placeholder="https://www.youtube.com/watch?v=...")
-        youtube_btn = gr.Button("Process YouTube Video", variant="primary")
-    
-    with gr.Tab("Upload Video"):
-        video_input = gr.Video(label="Upload Video File")
-        upload_btn = gr.Button("Process Uploaded Video", variant="primary")
+with gr.Blocks(theme=gr.themes.Soft(), title="Video Transcription") as demo:
+    gr.Markdown("""
+    # 🎥 Video Transcription & Summary Generator
+    Upload a video file or paste a YouTube link to get AI-powered transcription and summary.
+    """)
     
     with gr.Row():
         with gr.Column():
-            transcription_output = gr.Textbox(label="Full Transcription", lines=10)
+            gr.Markdown("### Input")
+            youtube_input = gr.Textbox(
+                label="YouTube URL (Optional)",
+                placeholder="https://www.youtube.com/watch?v=...",
+                lines=1
+            )
+            gr.Markdown("**OR**")
+            video_input = gr.Video(label="Upload Video File (Optional)")
+            process_btn = gr.Button("🚀 Process Video", variant="primary", size="lg")
+        
         with gr.Column():
-            summary_output = gr.Textbox(label="AI Summary", lines=10)
-    
-    download_output = gr.File(label="Download Transcript")
-    
-    # Event handlers
-    youtube_btn.click(
-        fn=lambda url: transcribe_and_summarize(youtube_url=url),
-        inputs=youtube_input,
-        outputs=[transcription_output, summary_output, download_output]
-    )
+            gr.Markdown("### Output")
+            status_output = gr.Textbox(label="Status", interactive=False)
+            
+            with gr.Accordion("Full Transcription", open=True):
+                transcription_output = gr.Textbox(label="", lines=10, show_copy_button=True)
+            
+            with gr.Accordion("Summary", open=True):
+                summary_output = gr.Textbox(label="", lines=5, show_copy_button=True)
+            
+            download_output = gr.File(label="📥 Download Complete Transcript")
     
-    upload_btn.click(
-        fn=lambda video: transcribe_and_summarize(video_file=video),
-        inputs=video_input,
-        outputs=[transcription_output, summary_output, download_output]
+    process_btn.click(
+        fn=transcribe_video,
+        inputs=[youtube_input, video_input],
+        outputs=[status_output, transcription_output, summary_output, download_output]
     )
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()