app.py

import gradio as gr
import os
import subprocess
from gtts import gTTS
from pydub import AudioSegment

def generate_video(image, text):
    # Save input image
    image_path = "input.jpg"
    image.save(image_path)

    # Create 5 sec mute video from image
    cmd = [
        "ffmpeg", "-y", "-loop", "1", "-i", image_path,
        "-t", "5", "-vf", "fps=25,scale=512:512",
        "-c:v", "libx264", "-pix_fmt", "yuv420p", "input.mp4"
    ]
    subprocess.run(cmd, check=True)

    # Generate audio from text using gTTS
    tts = gTTS(text, lang="en")
    tts.save("tts.mp3")
    AudioSegment.from_mp3("tts.mp3").export("tts.wav", format="wav")

    # Run Wav2Lip inference
    cmd = [
        "python", "Wav2Lip/inference.py",
        "--checkpoint_path", "checkpoints/wav2lip.pth",
        "--face", "input.mp4",
        "--audio", "tts.wav",
        "--outfile", "output.mp4"
    ]
    subprocess.run(cmd, check=True)

    return "output.mp4"

# Build Gradio interface
iface = gr.Interface(
    fn=generate_video,
    inputs=[
        gr.Image(type="pil", label="Upload Cartoon Image"),
        gr.Textbox(lines=2, placeholder="Enter text for speech...", label="Text to Speak")
    ],
    outputs=gr.Video(label="Talking Cartoon Video"),
    title="Cartoon Talking Video Generator",
    description="Upload a cartoon image and enter text to generate a talking video using Wav2Lip."
)

if __name__ == "__main__":
    iface.launch()