app.py

'''
import os
import shutil
from moviepy.editor import VideoFileClip, AudioFileClip
from gradio_client import Client

def process_videos(input_folder, output_folder):
    # 确保输出文件夹存在
    if not os.path.exists(output_folder):
        os.makedirs(output_folder)
    
    # 初始化Gradio客户端
    client = Client("http://127.0.0.1:7861")
    
    # 遍历输入文件夹中的所有文件
    for filename in os.listdir(input_folder):
        if filename.endswith('.mp4'):
            # 构建完整的文件路径
            mp4_path = os.path.join(input_folder, filename)
            txt_path = os.path.join(input_folder, filename.replace('.mp4', '.txt'))
            
            try:
                # 使用上下文管理器确保资源释放
                with VideoFileClip(mp4_path) as video:
                    # 将MP4转换为WAV
                    temp_wav_path = os.path.join(input_folder, filename.replace('.mp4', '.wav'))
                    video.audio.write_audiofile(temp_wav_path, logger=None)  # 禁用进度条避免冲突[8](@ref)
                    
                    # 调用API处理音频
                    result = client.predict(temp_wav_path, api_name="/predict")
                    
                    if len(result) >= 2:
                        processed_wav_path = result[1]
                        
                        # 关键修复：加载为音频对象而非字符串
                        with AudioFileClip(processed_wav_path) as processed_audio:
                            # 用处理后的音频替换原始视频的音频
                            video_with_new_audio = video.set_audio(processed_audio)
                            
                            # 保存处理后的视频到输出文件夹
                            output_mp4_path = os.path.join(output_folder, filename)
                            
                            # 写入视频文件（禁用进度条）
                            video_with_new_audio.write_videofile(
                                output_mp4_path,
                                codec='libx264',
                                audio_codec='aac',
                                logger=None  # 避免进度条导致的冲突[8](@ref)
                            )
                        
                        # 复制对应的TXT文件
                        if os.path.exists(txt_path):
                            output_txt_path = os.path.join(output_folder, filename.replace('.mp4', '.txt'))
                            shutil.copy2(txt_path, output_txt_path)
                    
                    else:
                        print(f"API返回结果不足，无法处理文件: {filename}")
                
                # 删除临时文件
                if os.path.exists(temp_wav_path):
                    os.remove(temp_wav_path)
                    
            except Exception as e:
                print(f"处理文件 {filename} 时出错: {str(e)}")

# 使用示例
input_folder = "wan_gen_videos_HunyuanVideo_Foley_sound_captioned"
output_folder = "wan_gen_videos_HunyuanVideo_Foley_no_vocal_sound_captioned"
process_videos(input_folder, output_folder)
'''

import os
import gradio as gr
from scipy.io.wavfile import write


def inference(audio):
  os.makedirs("out", exist_ok=True)
  write('test.wav', audio[0], audio[1])
  os.system("python3 -m demucs.separate -n htdemucs --two-stems=vocals test.wav -o out")
  return "./out/htdemucs/test/vocals.wav","./out/htdemucs/test/no_vocals.wav"
    
title = "Demucs Music Source Separation (v4)"
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/1911.13254' target='_blank'>Music Source Separation in the Waveform Domain</a> | <a href='https://github.com/facebookresearch/demucs' target='_blank'>Github Repo</a> | <a href='https://www.thafx.com' target='_blank'>//THAFX</a></p>"

gr.Interface(
    inference, 
    gr.Audio(type="numpy", label="Input"),
    [gr.Audio(type="filepath", label="Vocals"),gr.Audio(type="filepath", label="No Vocals / Instrumental")],
    title=title,
    article=article,
    ).launch(share = True)