Create app.py

app.py

@@ -0,0 +1,130 @@
+import gradio as gr
+import torch
+import yt_dlp
+import os
+import subprocess
+import json
+from threading import Thread
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import spaces
+import time
+import langdetect
+import uuid
+
+HF_TOKEN = os.environ.get("HF_TOKEN")
+print("Starting the program...")
+
+model_path = "Qwen/Qwen2.5-7B-Instruct"
+print(f"Loading model {model_path}...")
+tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype=torch.float16, trust_remote_code=True).cuda()
+model = model.eval()
+print("Model successfully loaded.")
+
+def generate_unique_filename(extension):
+    return f"{uuid.uuid4()}{extension}"
+
+def cleanup_files(*files):
+    for file in files:
+        if file and os.path.exists(file):
+            os.remove(file)
+            print(f"Removed file: {file}")
+
+def download_youtube_audio(url):
+    print(f"Downloading audio from YouTube: {url}")
+    output_path = generate_unique_filename(".wav")
+    ydl_opts = {
+        'format': 'bestaudio/best',
+        'postprocessors': [{
+            'key': 'FFmpegExtractAudio',
+            'preferredcodec': 'wav',
+        }],
+        'outtmpl': output_path,
+        'keepvideo': True,
+    }
+    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+        ydl.download([url])
+    
+    if os.path.exists(output_path + ".wav"):
+        os.rename(output_path + ".wav", output_path)
+    
+    return output_path
+
+@spaces.GPU(duration=90)
+def transcribe_audio(file_path):
+    print(f"Starting transcription of file: {file_path}")
+    temp_audio = None
+    if file_path.endswith(('.mp4', '.avi', '.mov', '.flv')):
+        print("Video file detected. Extracting audio using ffmpeg...")
+        temp_audio = generate_unique_filename(".wav")
+        command = ["ffmpeg", "-i", file_path, "-q:a", "0", "-map", "a", temp_audio]
+        subprocess.run(command, check=True)
+        file_path = temp_audio
+    
+    output_file = generate_unique_filename(".json")
+    command = [
+        "insanely-fast-whisper",
+        "--file-name", file_path,
+        "--device-id", "0",
+        "--model-name", "openai/whisper-large-v3",
+        "--task", "transcribe",
+        "--timestamp", "chunk",
+        "--transcript-path", output_file
+    ]
+    subprocess.run(command, check=True)
+    
+    with open(output_file, "r") as f:
+        transcription = json.load(f)
+    
+    result = transcription.get("text", " ".join([chunk["text"] for chunk in transcription.get("chunks", [])]))
+    
+    cleanup_files(output_file)
+    if temp_audio:
+        cleanup_files(temp_audio)
+    
+    return result
+
+def generate_summary_stream(transcription):
+    detected_language = langdetect.detect(transcription)
+    prompt = f"""Summarize the following video transcription in 150-300 words in {detected_language}:
+    {transcription[:300000]}..."""
+    
+    response, history = model.chat(tokenizer, prompt, history=[])
+    return response
+
+def process_youtube(url):
+    if not url:
+        return "Please enter a YouTube URL.", None
+    audio_file = download_youtube_audio(url)
+    transcription = transcribe_audio(audio_file)
+    cleanup_files(audio_file)
+    return transcription, None
+
+def process_uploaded_video(video_path):
+    transcription = transcribe_audio(video_path)
+    return transcription, None
+
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("""
+    # 🎥 Video Transcription and Smart Summary
+    Upload a video or provide a YouTube link to get a transcription and AI-generated summary.
+    """)
+    
+    with gr.Tabs():
+        with gr.TabItem("📤 Video Upload"):
+            video_input = gr.Video()
+            video_button = gr.Button("🚀 Process Video")
+        
+        with gr.TabItem("🔗 YouTube Link"):
+            url_input = gr.Textbox(placeholder="https://www.youtube.com/watch?v=...")
+            url_button = gr.Button("🚀 Process URL")
+    
+    transcription_output = gr.Textbox(label="📝 Transcription", lines=10, show_copy_button=True)
+    summary_output = gr.Textbox(label="📊 Summary", lines=10, show_copy_button=True)
+    summary_button = gr.Button("📝 Generate Summary")
+    
+    video_button.click(process_uploaded_video, inputs=[video_input], outputs=[transcription_output, summary_output])
+    url_button.click(process_youtube, inputs=[url_input], outputs=[transcription_output, summary_output])
+    summary_button.click(generate_summary_stream, inputs=[transcription_output], outputs=[summary_output])
+
+demo.launch()