Update app.py

app.py

@@ -1,62 +1,74 @@
 import gradio as gr
-from gradio_client import Client, handle_file
-import shutil
+import subprocess
+import os
 from pathlib import Path
 
-def generate_video(image, audio):
-    """Use existing HF Space API"""
+def setup():
+    """One-time setup"""
+    if Path("setup_done.txt").exists():
+        return
+    
+    print("Installing dependencies...")
+    os.system("pip install -q gfpgan realesrgan basicsr")
+    
+    print("Downloading model...")
+    os.system("git clone https://github.com/vinthony/video-retalking.git vrt")
+    
+    # Download checkpoints
+    os.chdir("vrt")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/30_net_gen.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/BFM.zip -P checkpoints/")
+    os.system("unzip -q checkpoints/BFM.zip -d checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/DNet.pt -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/ENet.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/expression.mat -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/face3d_pretrain_epoch_20.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/GFPGANv1.3.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/GPEN-BFR-512.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/LNet.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/ParseNet-latest.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/RetinaFace-R50.pth -P checkpoints/")
+    os.system("wget https://github.com/vinthony/video-retalking/releases/download/v0.0.1/shape_predictor_68_face_landmarks.dat -P checkpoints/")
+    os.chdir("..")
+    
+    Path("setup_done.txt").touch()
+    print("✅ Setup complete!")
+
+def generate(image, audio):
+    if not image or not audio:
+        return None, "Upload both files!"
+    
     try:
-        if not image or not audio:
-            return None, "❌ Please upload both files!"
+        setup()
         
-        print("🔄 Connecting to API...")
+        output = "result.mp4"
         
-        # Use the working MuseTalk space
-        client = Client("TMElyralab/MuseTalk")
+        # Run inference
+        cmd = f"cd vrt && python inference.py --face ../{image} --audio ../{audio} --outfile ../{output}"
         
-        print("📤 Uploading files...")
+        result = os.system(cmd)
         
-        result = client.predict(
-            audio_path=handle_file(audio),
-            video_path=None,
-            bbox_shift=0,
-            api_name="/predict"
-        )
-        
-        # Result is a file path
-        if result and Path(result).exists():
-            # Copy to local output
-            output = "result.mp4"
-            shutil.copy(result, output)
-            return output, "✅ Video generated successfully!"
+        if Path(output).exists():
+            return output, "✅ Done!"
         else:
-            return None, "❌ API returned no result"
+            return None, "❌ Failed to generate"
             
     except Exception as e:
         return None, f"❌ Error: {str(e)}"
 
-# Gradio UI
-with gr.Blocks(theme=gr.themes.Soft()) as app:
-    gr.Markdown("# 🎬 AI Lip Sync Generator")
-    gr.Markdown("Upload a face image and audio to create lip-synced video")
-    
-    with gr.Row():
-        with gr.Column():
-            img = gr.Image(type="filepath", label="📷 Face Image")
-            aud = gr.Audio(type="filepath", label="🎵 Audio File")
-            btn = gr.Button("🚀 Generate Video", variant="primary")
-        
-        with gr.Column():
-            vid = gr.Video(label="📹 Result")
-            status = gr.Textbox(label="Status", lines=2)
-    
-    btn.click(generate_video, [img, aud], [vid, status])
-    
-    gr.Markdown("""
-    ### 💡 Notes:
-    - Uses MuseTalk API (no local installation needed)
-    - Processing time: 30-90 seconds
-    - Best with clear front-facing images
-    """)
+# Simple Gradio UI
+demo = gr.Interface(
+    fn=generate,
+    inputs=[
+        gr.Image(type="filepath", label="Face Image"),
+        gr.Audio(type="filepath", label="Audio File")
+    ],
+    outputs=[
+        gr.Video(label="Result"),
+        gr.Textbox(label="Status")
+    ],
+    title="AI Lip Sync Generator",
+    description="Upload a face image and audio to generate lip-synced video"
+)
 
-app.launch()
+demo.launch()