app.py

import gradio as gr
import os
import tempfile
from huggingface_hub import snapshot_download
from diffsynth import ModelManager, WanVideoPusaPipeline, save_video
import spaces

# Constants
WAN_SUBFOLDER = "Wan2.1-T2V-14B"
MODEL_REPO_ID = "RaphaelLiu/PusaV1"
MODEL_ZOO_DIR = "./model_zoo"
WAN_MODEL_PATH = os.path.join(MODEL_ZOO_DIR, WAN_SUBFOLDER)
LORA_PATH = os.path.join(MODEL_ZOO_DIR, "PusaV1", "pusa_v1.pt")

# Ensure model is downloaded
def ensure_model_downloaded():
    if not os.path.exists(WAN_MODEL_PATH):
        print("Downloading Wan2.1-T2V-14B from HuggingFace Hub...")
        snapshot_download(
            repo_id=MODEL_REPO_ID,
            local_dir=MODEL_ZOO_DIR,
            repo_type="model",
            allow_patterns=[f"{WAN_SUBFOLDER}/**"],
            local_dir_use_symlinks=False,
        )
        print("Model downloaded.")

# Video generation logic
@spaces.GPU
def generate_video(prompt: str):
    ensure_model_downloaded()

    # Load model
    manager = ModelManager(pretrained_model_dir=WAN_MODEL_PATH)
    model = manager.load_model()

    # Set up pipeline
    pipeline = WanVideoPusaPipeline(model=model)
    pipeline.set_lora_adapters(LORA_PATH)

    # Generate video
    result = pipeline(prompt)

    # Save video
    tmp_dir = tempfile.mkdtemp()
    output_path = os.path.join(tmp_dir, "video.mp4")
    save_video(result.frames, output_path, fps=8)

    return output_path

# Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("## 🎥 Wan2.1-T2V-14B with Pusa LoRA | Text-to-Video Generator")

    prompt_input = gr.Textbox(
        lines=4,
        label="Prompt",
        placeholder="Describe your video (e.g. A coral reef full of colorful fish...)"
    )

    generate_btn = gr.Button("Generate Video")
    video_output = gr.Video(label="Output")

    generate_btn.click(fn=generate_video, inputs=prompt_input, outputs=video_output)

demo.launch()