app.py

import os
import sys

# Apply compatibility patches first
try:
    import compatibility_patches
except ImportError:
    print("Warning: compatibility_patches not found")

# Print environment information
print("==== Environment Information ====")
print(f"Python version: {sys.version}")
print(f"Working directory: {os.getcwd()}")
print(f"Directory contents: {os.listdir('.')}")

# Import key packages and print versions
packages = {
    "torch": None,
    "transformers": None,
    "diffusers": None,
    "accelerate": None,
    "einops": None,
    "omegaconf": None,
    "librosa": None 
}

for package in packages.keys():
    try:
        module = __import__(package)
        try:
            packages[package] = module.__version__
            print(f"{package}: {module.__version__}")
        except AttributeError:
            packages[package] = "Installed (version unknown)"
            print(f"{package}: Installed (version unknown)")
    except ImportError as e:
        print(f"{package}: Not installed")
        packages[package] = "Not installed"

# Create necessary directories
os.makedirs("assets", exist_ok=True)
os.makedirs("ckpts", exist_ok=True)
os.makedirs("outputs", exist_ok=True)

# Download example audio files if they don't exist
if not os.path.exists("assets/fire_crackling.wav") or not os.path.exists("assets/plastic_bag.wav"):
    try:
        from download_assets import download_gdrive_file
        
        assets = {
            "assets/fire_crackling.wav": "1vOAZcbkpo_hre2g26n--lUXdwbTQp22k",
            "assets/plastic_bag.wav": "15igeDor7a47a-oluSCfO6GeUvFVl2ttb"
        }
        
        for path, file_id in assets.items():
            if not os.path.exists(path):
                download_gdrive_file(file_id, path)
    except Exception as e:
        print(f"Error downloading example audio files: {e}")

# Import required packages
import gradio as gr

# Import our controller
from controller import SonicDiffusionController

# Initialize controller
controller = SonicDiffusionController()

# Create the Gradio interface
with gr.Blocks(title="SonicDiffusion") as demo:
    gr.Markdown("# SonicDiffusion - Audio-to-Image Generation")
    gr.Markdown("Generate images conditioned on audio inputs using Stable Diffusion")
    
    status_output = gr.Textbox(label="Status", value="System initialized. Start by checking dependencies and downloading assets.", lines=8)
    
    with gr.Tab("1. Setup"):
        with gr.Row():
            with gr.Column():
                check_deps_btn = gr.Button("Check Dependencies")
                
                def format_deps():
                    deps = controller.check_dependencies()
                    return "\n".join([f"{pkg}: {vers}" for pkg, vers in deps.items()])
                
                check_deps_btn.click(
                    fn=format_deps,
                    outputs=status_output
                )
            
            with gr.Column():
                check_assets_btn = gr.Button("Check Assets")
                
                def format_assets():
                    assets = controller.check_assets()
                    return "\n".join([f"{path}: {'✓' if exists else '✗'}" for path, exists in assets.items()])
                
                check_assets_btn.click(
                    fn=format_assets,
                    outputs=status_output
                )
                
        download_assets_btn = gr.Button("Download Required Assets", variant="primary")
        download_assets_btn.click(fn=controller.download_assets, outputs=status_output)
    
    with gr.Tab("2. Generate"):
        with gr.Row():
            with gr.Column():
                model_dropdown = gr.Dropdown(
                    label="Select Model",
                    choices=["Landscape Model", "Greatest Hits Model"],
                    value="Landscape Model"
                )
                
                load_model_btn = gr.Button("Load Selected Model", variant="primary")
                load_model_btn.click(
                    fn=controller.load_model,
                    inputs=[model_dropdown],
                    outputs=status_output
                )
                
                prompt_input = gr.Textbox(
                    label="Prompt",
                    placeholder="Enter a descriptive prompt...",
                    value="a high quality photograph of a fantasy landscape"
                )
                
                # Use the preloaded example audio files
                audio_dropdown = gr.Dropdown(
                    label="Select Example Audio",
                    choices=["assets/fire_crackling.wav", "assets/plastic_bag.wav", "Upload my own audio"],
                    value="assets/fire_crackling.wav"
                )
                
                audio_input = gr.Audio(
                    label="Or Upload Your Own Audio",
                    type="filepath",
                    sources=["upload", "microphone"],
                    visible=False
                )
                
                def toggle_audio_upload(choice):
                    return gr.update(visible=(choice == "Upload my own audio"))
                
                audio_dropdown.change(
                    fn=toggle_audio_upload,
                    inputs=[audio_dropdown],
                    outputs=[audio_input]
                )
                
                with gr.Row():
                    cfg_scale = gr.Slider(
                        label="CFG Scale",
                        minimum=1.0,
                        maximum=20.0,
                        value=7.5,
                        step=0.5
                    )
                    
                    steps = gr.Slider(
                        label="Steps",
                        minimum=20,
                        maximum=100,
                        value=50,
                        step=5
                    )
                
                def prepare_audio_path(dropdown_choice, uploaded_audio):
                    if dropdown_choice == "Upload my own audio":
                        return uploaded_audio
                    else:
                        return dropdown_choice
                
                generate_btn = gr.Button("Generate Image", variant="primary")
            
            with gr.Column():
                output_image = gr.Image(label="Generated Image", type="pil")
                output_status = gr.Textbox(label="Generation Status", value="")
                
                generate_btn.click(
                    fn=lambda p, d, u, c, s: (
                        controller.generate(p, prepare_audio_path(d, u), c, s),
                        "Generation completed"
                    ),
                    inputs=[prompt_input, audio_dropdown, audio_input, cfg_scale, steps],
                    outputs=[output_image, output_status]
                )

if __name__ == "__main__":
    # Change the server parameters
    demo.launch(server_name="0.0.0.0", share=True)