module.py

import os
import re
import random
from scipy.io.wavfile import write
from scipy.io.wavfile import read
import numpy as np
import gradio as gr
import yt_dlp
import subprocess
from pydub import AudioSegment
from audio_separator.separator import Separator
from lib.infer import infer_audio
import edge_tts
import tempfile
import anyio
from pathlib import Path
from lib.language_tts import language_dict
import shutil
import time
from argparse import ArgumentParser
from download_model import download_online_model

main_dir = Path().resolve()
print(main_dir)

os.chdir(main_dir)
models_dir = main_dir / "rvc_models"
audio_separat_dir = main_dir / "audio_input"
AUDIO_DIR = main_dir / 'audio_input'


# Function to list all folders in the models directory
def get_folders():
    if models_dir.exists() and models_dir.is_dir():
        return [folder.name for folder in models_dir.iterdir() if folder.is_dir()]
    return []


# Function to refresh and return the list of folders
def refresh_folders():
    return gr.Dropdown.update(choices=get_folders())


# Function to get the list of audio files in the specified directory
def get_audio_files():
    if not os.path.exists(AUDIO_DIR):
        os.makedirs(AUDIO_DIR)
    return [f for f in os.listdir(AUDIO_DIR) if f.lower().endswith(('.mp3', '.wav', '.flac', '.ogg', '.aac'))]


# Function to return the full path of audio files for playback
def load_audio_files():
    audio_files = get_audio_files()
    return [os.path.join(AUDIO_DIR, f) for f in audio_files]


def refresh_audio_list():
    audio_files = load_audio_files()
    return gr.Dropdown.update(choices=audio_files)


def download_audio(url):
    ydl_opts = {
        'format': 'bestaudio/best',
        'outtmpl': 'ytdl/%(title)s.%(ext)s',
        'postprocessors': [{
            'key': 'FFmpegExtractAudio',
            'preferredcodec': 'wav',
            'preferredquality': '192',
        }],
    }

    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
        info_dict = ydl.extract_info(url, download=True)
        file_path = ydl.prepare_filename(info_dict).rsplit('.', 1)[0] + '.wav'
        return file_path


async def text_to_speech_edge(text, language_code):
    voice = language_dict.get(language_code, "default_voice")
    communicate = edge_tts.Communicate(text, voice)
    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
        tmp_path = tmp_file.name
        await communicate.save(tmp_path)
    return tmp_path


if __name__ == '__main__':
    parser = ArgumentParser()
    parser.add_argument("--share", action="store_true", dest="share_enabled", default=False)
    parser.add_argument("--listen", action="store_true", default=False)
    parser.add_argument('--listen-host', type=str)
    parser.add_argument('--listen-port', type=int)
    args = parser.parse_args()

# Gradio Interface
with gr.Blocks(title="Hex RVC", theme=gr.themes.Base(primary_hue="red", secondary_hue="pink")) as app:
    gr.Markdown("# Hex RVC")
    gr.Markdown("Join [AIHub](https://discord.gg/aihub) to get the RVC model!")
    
    with gr.Tab("Inference"):
        with gr.Row():
            MODEL_NAME = gr.Dropdown(
                label="Select a Model",
                choices=get_folders(),
                interactive=True
            )
            SOUND_PATH = gr.Dropdown(
                choices=load_audio_files(),
                label="Select an audio file",
                interactive=True
            )
            upload_audio = gr.Audio(label="Upload Audio", type='filepath')

        with gr.Accordion("Hex TTS"):
            input_text = gr.Textbox(lines=5, label="Input Text")
            language = gr.Dropdown(choices=list(language_dict.keys()), label="Choose the Voice Model")
            tts_convert = gr.Button("Convert")
            tts_output = gr.Audio(label="Generated TTS Audio", type='filepath')
            
            tts_convert.click(
                fn=text_to_speech_edge, 
                inputs=[input_text, language], 
                outputs=tts_output
            )

        output_audio = gr.Audio(label="Generated Audio", type='filepath')
        with gr.Row():
            refresh_btn = gr.Button("Refresh")
            run_button = gr.Button("Convert")

        refresh_btn.click(
            lambda: (refresh_audio_list(), refresh_folders()),
            outputs=[SOUND_PATH, MODEL_NAME]
        )

    with gr.Tab("Download RVC Model"):
        url = gr.Textbox(label="Your Model URL")
        dirname = gr.Textbox(label="Your Model Name")
        download_button = gr.Button("Download Model")
        download_output = gr.Textbox(label="Download Status")

        download_button.click(
            download_online_model,
            inputs=[url, dirname],
            outputs=download_output
        )

    with gr.Tab("UVR5 (Demo)"):
        input_audio = gr.Audio(type="filepath", label="Upload Audio")
        roformer_link = gr.Textbox(label="Audio Link")
        roformer_download_button = gr.Button("Download")
        
        separate_button = gr.Button("Separate Audio")
        separation_output = gr.Textbox(label="Separation Output Path")

        roformer_download_button.click(download_audio, inputs=[roformer_link], outputs=[input_audio])

app.launch(
    share=args.share_enabled,
    server_name=None if not args.listen else (args.listen_host or '0.0.0.0'),
    server_port=args.listen_port,
)