infer-web.py

import os
import shutil
import subprocess
from mega import Mega
import gradio as gr

def download_from_url(url, model):
    try:
        model = model.replace('.pth', '').replace('.index', '').replace('.zip', '')
        url = url.replace('/blob/main/', '/resolve/main/').strip()

        for directory in ["downloads", "unzips", "zip"]:
            os.makedirs(directory, exist_ok=True)

        if url.endswith('.pth'):
            subprocess.run(["wget", url, "-O", f'assets/weights/{model}.pth'])
        elif url.endswith('.index'):
            os.makedirs(f'logs/{model}', exist_ok=True)
            subprocess.run(["wget", url, "-O", f'logs/{model}/added_{model}.index'])
        elif url.endswith('.zip'):
            subprocess.run(["wget", url, "-O", f'downloads/{model}.zip'])
        else:
            if "drive.google.com" in url:
                url = url.split('/')[0]
                subprocess.run(["gdown", url, "--fuzzy", "-O", f'downloads/{model}'])
            elif "mega.nz" in url:
                Mega().download_url(url, 'downloads')
            else:
                subprocess.run(["wget", url, "-O", f'downloads/{model}'])

        downloaded_file = next((f for f in os.listdir("downloads")), None)
        if downloaded_file:
            if downloaded_file.endswith(".zip"):
                shutil.unpack_archive(f'downloads/{downloaded_file}', "unzips", 'zip')
                for root, _, files in os.walk('unzips'):
                    for file in files:
                        file_path = os.path.join(root, file)
                        if file.endswith(".index"):
                            os.makedirs(f'logs/{model}', exist_ok=True)
                            shutil.copy2(file_path, f'logs/{model}')
                        elif file.endswith(".pth") and "G_" not in file and "D_" not in file:
                            shutil.copy(file_path, f'assets/weights/{model}.pth')
            elif downloaded_file.endswith(".pth"):
                shutil.copy(f'downloads/{downloaded_file}', f'assets/weights/{model}.pth')
            elif downloaded_file.endswith(".index"):
                os.makedirs(f'logs/{model}', exist_ok=True)
                shutil.copy(f'downloads/{downloaded_file}', f'logs/{model}/added_{model}.index')
            else:
                return "Failed to download file"
        return f"Successfully downloaded {model} voice models"
    except Exception as e:
        return f"Error: {str(e)}"
    finally:
        shutil.rmtree("downloads", ignore_errors=True)
        shutil.rmtree("unzips", ignore_errors=True)
        shutil.rmtree("zip", ignore_errors=True)

def listen_to_model(model_path, index_path, pitch, input_path, f0_method, save_as, index_rate, volume_normalization, consonant_protection):
    if not os.path.exists(model_path):
        return "Model path not found"
    if not os.path.exists(index_path):
        return f"{index_path} was not found"
    if not os.path.exists(input_path):
        return f"{input_path} was not found"
    
    os.environ['index_root'] = os.path.dirname(index_path)
    index_path = os.path.basename(index_path)
    model_name = os.path.basename(model_path)
    os.environ['weight_root'] = os.path.dirname(model_path)

    try:
        command = [
            "python", "tools/infer_cli.py",
            "--f0up_key", str(pitch),
            "--input_path", input_path,
            "--index_path", index_path,
            "--f0method", f0_method,
            "--opt_path", save_as,
            "--model_name", model_name,
            "--index_rate", str(index_rate),
            "--device", "cuda:0",
            "--is_half", "True",
            "--filter_radius", "3",
            "--resample_sr", "0",
            "--rms_mix_rate", str(volume_normalization),
            "--protect", str(consonant_protection)
        ]
        subprocess.run(command, check=True)
        return save_as
    except subprocess.CalledProcessError as e:
        return f"Error: {str(e)}"

with gr.Blocks() as demo:
    gr.Markdown("# RVC V2 Web UI")

    with gr.Tabs():
        with gr.TabItem("Download Model"):
            gr.Markdown("### Download RVC Model")
            url_input = gr.Textbox(label="Model URL", placeholder="Enter the model URL here")
            model_input = gr.Textbox(label="Model Name", placeholder="Enter the model name here")
            download_button = gr.Button("Download")
            download_output = gr.Textbox(label="Download Status")
            download_button.click(download_from_url, inputs=[url_input, model_input], outputs=download_output)

        with gr.TabItem("Listen to Model"):
            gr.Markdown("### Listen to Your Model")
            model_path_input = gr.Textbox(label="Model Path", value="/content/RVC/assets/weights/Sonic.pth")
            index_path_input = gr.Textbox(label="Index Path", value="/content/RVC/logs/Sonic/added_IVF905_Flat_nprobe_1_Sonic_v2.index")
            input_path_input = gr.Textbox(label="Input Audio Path", value="/content/RVC/audios/astronauts.mp3")
            save_as_input = gr.Textbox(label="Save Output As", value="/content/RVC/audios/cli_output.wav")
            f0_method_input = gr.Radio(label="F0 Method", choices=["rmvpe", "pm", "harvest"], value="rmvpe")

            with gr.Row():
                pitch_input = gr.Slider(label="Pitch", minimum=-12, maximum=12, step=1, value=0)
                index_rate_input = gr.Slider(label="Index Rate", minimum=0, maximum=1, step=0.01, value=0.5)
                volume_normalization_input = gr.Slider(label="Volume Normalization", minimum=0, maximum=1, step=0.01, value=0)
                consonant_protection_input = gr.Slider(label="Consonant Protection", minimum=0, maximum=1, step=0.01, value=0.5)

            listen_button = gr.Button("Generate and Listen")
            audio_output = gr.Audio(label="Output Audio")
            listen_button.click(
                listen_to_model, 
                inputs=[
                    model_path_input, 
                    index_path_input, 
                    pitch_input, 
                    input_path_input, 
                    f0_method_input, 
                    save_as_input, 
                    index_rate_input, 
                    volume_normalization_input, 
                    consonant_protection_input
                ], 
                outputs=audio_output
            )

demo.launch()