Simplify app.py and fix path issues

app.py

@@ -3,67 +3,28 @@ import torch
 import os
 import sys
 import numpy as np
-import subprocess
 from pathlib import Path
 
-# Setup progress tracking
-progress_status = "Starting setup..."
-
-def setup_environment():
-    """Set up the environment and download necessary files"""
-    global progress_status
-    
-    # Clone the MDM repository during the first run
-    if not Path("motion-diffusion-model").exists():
-        progress_status = "Cloning Motion Diffusion Model repository..."
-        # Clone the repository
-        subprocess.run(["git", "clone", "https://github.com/GuyTevet/motion-diffusion-model.git"])
-        
-        # Install spacy language model (required by MDM)
-        progress_status = "Installing Spacy language model..."
-        subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
-    
-    # Download other necessary files if they don't exist
-    if not Path("motion-diffusion-model/data/smpl").exists():
-        progress_status = "Downloading SMPL files..."
-        subprocess.run(["bash", "motion-diffusion-model/prepare/download_smpl_files.sh"])
-    
-    if not Path("motion-diffusion-model/data/glove").exists():
-        progress_status = "Downloading Glove embeddings..."
-        subprocess.run(["bash", "motion-diffusion-model/prepare/download_glove.sh"])
-    
-    if not Path("motion-diffusion-model/save/t2m_evaluators").exists():
-        progress_status = "Downloading Text-to-Motion evaluators..."
-        subprocess.run(["bash", "motion-diffusion-model/prepare/download_t2m_evaluators.sh"])
-    
-    # Add the repository to the Python path
+# Add the motion-diffusion-model repository to the Python path if it exists
+if Path("motion-diffusion-model").exists():
     sys.path.append('./motion-diffusion-model')
-    
-    progress_status = "Setup complete! Ready to generate animations."
-    return "Environment setup complete"
-
-def get_setup_status():
-    """Return the current setup status"""
-    global progress_status
-    return progress_status
-
-# Initialize setup in background
-try:
-    setup_environment()
-except Exception as e:
-    progress_status = f"Setup error: {str(e)}"
 
 def text_to_motion(text_prompt, motion_length=3.0, seed=0):
     """Generate motion from text prompt using MDM"""
-    global progress_status
-    
     try:
-        # Import MDM functions only after the repository is cloned and path is set
-        sys.path.append('./motion-diffusion-model')
-        from sample.generate import generate
-        from visualization.visualize import visualize
+        # First, check if we need to clone the repository
+        if not Path("motion-diffusion-model").exists():
+            import subprocess
+            print("Cloning Motion Diffusion Model repository...")
+            subprocess.run(["git", "clone", "https://github.com/GuyTevet/motion-diffusion-model.git"])
+            # Add the repository to the Python path after cloning
+            sys.path.append('./motion-diffusion-model')
+            print("Installed Spacy language model...")
+            subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
         
-        progress_status = "Generating motion..."
+        # Now import the necessary functions
+        from motion_diffusion_model.sample.generate import generate
+        from motion_diffusion_model.visualization.visualize import visualize
         
         # Set the seed for reproducibility
         if seed is not None:
@@ -73,7 +34,7 @@ def text_to_motion(text_prompt, motion_length=3.0, seed=0):
         model_path = "checkpoints/mld_humanml.pt"
         
         # Generate the motion
-        progress_status = "Running MDM generation..."
+        print("Running MDM generation...")
         motion_data = generate(
             model_path=model_path,
             text_prompt=text_prompt,
@@ -82,57 +43,28 @@ def text_to_motion(text_prompt, motion_length=3.0, seed=0):
         )
         
         # Render the animation
-        progress_status = "Rendering animation..."
+        print("Rendering animation...")
         output_path = "output.mp4"
         visualize(motion_data, output_path)
         
-        progress_status = "Animation generated successfully!"
+        print("Animation generated successfully!")
         return output_path
     except Exception as e:
-        progress_status = f"Error generating motion: {str(e)}"
+        print(f"Error generating motion: {str(e)}")
         return None
 
 # Create the Gradio interface
-with gr.Blocks() as demo:
-    gr.Markdown("# Motion Diffusion Model (MDM)")
-    gr.Markdown("Generate human motions from text descriptions using MDM")
-    
-    with gr.Row():
-        with gr.Column():
-            text_input = gr.Textbox(
-                label="Text Prompt", 
-                placeholder="A person walks forward, then turns left",
-                lines=3
-            )
-            length_slider = gr.Slider(
-                minimum=1.0, 
-                maximum=9.8, 
-                value=3.0, 
-                label="Motion Length (seconds)"
-            )
-            seed_number = gr.Number(
-                label="Random Seed", 
-                value=0
-            )
-            generate_btn = gr.Button("Generate Motion")
-            status_text = gr.Textbox(
-                label="Status", 
-                value=progress_status,
-                interactive=False
-            )
-            
-        with gr.Column():
-            video_output = gr.Video(label="Generated Motion")
-    
-    # Update the status periodically
-    status_text.change(get_setup_status, inputs=None, outputs=status_text, every=1)
-    
-    # Connect the button to the generation function
-    generate_btn.click(
-        text_to_motion, 
-        inputs=[text_input, length_slider, seed_number],
-        outputs=video_output
-    )
+demo = gr.Interface(
+    fn=text_to_motion,
+    inputs=[
+        gr.Textbox(label="Text Prompt", placeholder="A person walks forward, then turns left", lines=3),
+        gr.Slider(minimum=1.0, maximum=9.8, value=3.0, label="Motion Length (seconds)"),
+        gr.Number(label="Random Seed", value=0)
+    ],
+    outputs=gr.Video(label="Generated Motion"),
+    title="Motion Diffusion Model (MDM)",
+    description="Generate human motions from text descriptions using MDM"
+)
 
 # Launch the app
 if __name__ == "__main__":