.gitignore

@@ -1,33 +0,0 @@
-# Python
-__pycache__/
-*.py[cod]
-*$py.class
-*.so
-*.egg-info/
-dist/
-build/
-*.egg
-
-# Virtual environments
-venv/
-.venv/
-env/
-
-# IDE
-.idea/
-.vscode/
-*.swp
-*.swo
-*~
-
-# Logs
-nohup.out
-*.log
-
-# OS
-.DS_Store
-Thumbs.db
-
-# Temporary files
-*.tmp
-*.bak

README.md

@@ -1,192 +1,2 @@
 # demucs-service
 
-Vocal/accompaniment separation web service using [Demucs](https://github.com/adefossez/demucs) (htdemucs model).
-
-## Overview
-
-This service provides an async API for separating audio files into vocals and accompaniment tracks using Meta's Demucs neural network model. It follows the ahserver + longtasks + Redis pattern.
-
-## Architecture
-
-- **ahserver**: Async HTTP server framework
-- **longtasks**: Background task processing via Redis queues
-- **Redis**: Task queue for separation jobs
-- **Demucs 4.0.1**: AI-powered source separation model (htdemucs)
-
-## 模型下载（离线部署）
-
-Demucs 使用 htdemucs 模型（PyTorch hub 格式），首次运行时自动下载，也可手动预下载。
-
-### 方法1: PyTorch Hub 自动下载（默认）
-
-服务首次运行时会通过 `torch.hub.load("facebookresearch/demucs", "htdemucs")` 自动下载模型。
-
-**下载位置**: `~/.cache/torch/hub/checkpoints/`  
-**下载大小**: ~80MB
-
-### 方法2: 手动预下载
-
-如果部署环境无法访问外网，可以先在有网络的机器上下载，再拷贝：
-
-```bash
-# 在有网络的机器上运行 Python
-python3 << "PYTHON"
-import torch
-model = torch.hub.load("facebookresearch/demucs", "htdemucs", pretrained=True)
-print("Model downloaded to:", torch.hub.get_dir())
-PYTHON
-
-# 找到模型文件
-ls ~/.cache/torch/hub/checkpoints/
-# 应该看到 htdemucs-*.pt 或类似文件
-
-# 拷贝到部署服务器
-scp ~/.cache/torch/hub/checkpoints/htdemucs*.pt ymq@opencomputing.net:~/.cache/torch/hub/checkpoints/
-```
-
-### 方法3: 直接下载模型文件
-
-```bash
-# 创建缓存目录
-mkdir -p ~/.cache/torch/hub/checkpoints/
-
-# 下载 htdemucs 模型
-wget -O ~/.cache/torch/hub/checkpoints/htdemucs_v4.pt \
-  https://dl.fbaipublicfiles.com/demucs/v4.0/htdemucs.pth
-```
-
-**下载大小**: ~80MB  
-**下载时间**: 约5-15秒
-
-### 验证下载
-
-```bash
-# 启动服务后检查日志
-tail -f /data/ymq/demucs-service/nohup.out | grep -i "model"
-# 应该看到 "Model loaded" 而不是 "Downloading"
-```
-
-### 模型来源
-
-- **GitHub**: https://github.com/facebookresearch/demucs
-- **PyTorch Hub**: facebookresearch/demucs
-- **License**: MIT
-- **Paper**: Hybrid Spectrogram and Waveform Source Separation
-
-
-## API
-
-### Submit Separation Task
-
-Send a JSON payload to the longtask endpoint:
-
-```json
-{
-    "task_type": "separate",
-    "audio_path": "/path/to/audio.wav",
-    "output_dir": "/tmp/demucs_custom_output"  // optional
-}
-```
-
-**Parameters:**
-- `audio_path` (required): Absolute path to the input audio file
-- `output_dir` (optional): Output directory. Default: `/tmp/demucs_{task_id}`
-
-**Response:**
-```json
-{
-    "vocals_path": "/tmp/demucs_123/htdemucs/audio/vocals.wav",
-    "no_vocals_path": "/tmp/demucs_123/htdemucs/audio/no_vocals.wav",
-    "duration": 12.34,
-    "output_dir": "/tmp/demucs_123",
-    "model": "htdemucs"
-}
-```
-
-### Health Check
-
-```
-GET /app/health.dspy
-```
-
-Returns:
-```json
-{"status":"ok","service":"demucs-service","model":"htdemucs"}
-```
-
-## Configuration
-
-Config file: `conf/config.json`
-
-```json
-{
-    "port": 9083,
-    "queue": "demucs",
-    "filesroot": "/tmp/demucs-outputs",
-    "host": "0.0.0.0",
-    "debug": false
-}
-```
-
-## Environment Variables
-
-| Variable | Default | Description |
-|----------|---------|-------------|
-| `DEMUCS_GPU_ID` | `5` | GPU device ID for CUDA |
-| `CUDA_VISIBLE_DEVICES` | `5` | CUDA device visibility |
-| `PYTHONPATH` | `/data/ymq/demucs-service` | Python module path |
-
-## Deployment
-
-### Prerequisites
-
-- Python venv at `/data/ymq/demucs_venv` with demucs 4.0.1 and torchcodec
-- Redis server running on `127.0.0.1:6379`
-- GPU with CUDA support
-
-### Start
-
-```bash
-bash start.sh
-```
-
-### Stop
-
-```bash
-bash stop.sh
-```
-
-### Logs
-
-```bash
-tail -f nohup.out
-```
-
-## Directory Structure
-
-```
-demucs-service/
-├── ah.py                  # Main entry point
-├── workers/
-│   ├── __init__.py
-│   └── separate.py        # Separation worker
-├── conf/
-│   └── config.json        # Service configuration
-├── app/
-│   └── health.dspy        # Health check endpoint
-├── start.sh               # Start script
-├── stop.sh                # Stop script
-└── README.md              # This file
-```
-
-## Output Format
-
-Demucs outputs to: `{output_dir}/htdemucs/{basename}/`
-- `vocals.wav` - Isolated vocal track
-- `no_vocals.wav` - Accompaniment (everything except vocals)
-
-## Troubleshooting
-
-- **GPU OOM**: The htdemucs model requires significant VRAM. Ensure the assigned GPU has enough memory.
-- **Process timeout**: Long audio files may exceed the stuck_seconds timeout (default: 600s). Increase if needed.
-- **Missing output files**: Check nohup.out for demucs stderr output to diagnose issues.

ah.py

@@ -1,34 +0,0 @@
-import os
-from ahserver.webapp import webapp
-from ahserver.serverenv import ServerEnv
-from ahserver.configuredServer import add_startup
-from longtasks.longtasks import LongTasks, schedule_once
-from appPublic.log import debug
-
-class DemucsTasks(LongTasks):
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.gpu_id = int(os.environ.get('DEMUCS_GPU_ID', '5'))
-    async def process_task(self, payload, workid=None):
-        import json
-        if isinstance(payload, str): payload = json.loads(payload)
-        task_type = payload.get('task_type', '')
-        if task_type == 'separate':
-            from workers.separate import run_separate
-            return await run_separate(self, payload)
-        raise ValueError(f'Unknown task_type: {task_type}')
-
-async def on_app_built(app):
-    env = ServerEnv()
-    lt = env.longtasks
-    if lt:
-        schedule_once(0.1, lt.run)
-        debug(f'Demucs longtasks worker started, GPU: {lt.gpu_id}')
-
-def init():
-    env = ServerEnv()
-    env.longtasks = DemucsTasks('redis://127.0.0.1:6379', 'demucs', worker_cnt=1, stuck_seconds=600, max_age_hours=24)
-    add_startup(on_app_built)
-
-if __name__ == '__main__':
-    webapp(init)

app/api/status/index.dspy

@@ -1,31 +0,0 @@
-# -*- coding:utf-8 -*-
-# GET /api/status - Demucs服务状态
-
-import subprocess
-import json
-
-result = {
-    'service': 'demucs-vocal-separation',
-    'model': 'htdemucs',
-    'gpu_id': 5,
-    'gpus': []
-}
-
-try:
-    out = subprocess.check_output(
-        ['nvidia-smi', '--query-gpu=index,utilization.gpu,memory.used,memory.total',
-         '--format=csv,noheader,nounits'],
-        timeout=5
-    ).decode().strip()
-    for line in out.split('\n'):
-        parts = [p.strip() for p in line.split(',')]
-        result['gpus'].append({
-            'id': int(parts[0]),
-            'util': int(parts[1]),
-            'mem_used': int(parts[2]),
-            'mem_total': int(parts[3])
-        })
-except Exception:
-    pass
-
-return json.dumps(result)

app/api/submit/index.dspy

@@ -1,50 +0,0 @@
-# -*- coding:utf-8 -*-
-# POST /api/submit - 提交Demucs人声分离任务
-
-import json
-import uuid
-from ahserver.serverenv import ServerEnv
-
-method = request.method
-
-if method == 'POST':
-    audio_path = params_kw.get('audio_path', '')
-    if not audio_path:
-        return json.dumps({'error': 'audio_path is required'}, ensure_ascii=False)
-
-    task_id = params_kw.get('task_id', str(uuid.uuid4()).replace("-", "")[:12])
-    model = params_kw.get('model', 'htdemucs')
-
-    payload = {
-        'task_type': 'separate',
-        'task_id': task_id,
-        'audio_path': audio_path,
-        'model': model
-    }
-
-    env = ServerEnv()
-    longtasks = env.longtasks
-    if longtasks is None:
-        return json.dumps({'error': 'service not ready'}, ensure_ascii=False)
-
-    result = await longtasks.submit_task(payload)
-    real_task_id = result.get('task_id', str(result)) if isinstance(result, dict) else str(result)
-
-    return json.dumps({
-        'task_id': real_task_id,
-        'status': 'queued',
-        'audio_path': audio_path,
-        'model': model,
-        'message': 'task submitted',
-        'check_url': f'/api/task?task_id={real_task_id}'
-    }, ensure_ascii=False)
-
-else:
-    return json.dumps({
-        'usage': 'POST with JSON body',
-        'params': {
-            'audio_path': 'string (required, server path to audio file)',
-            'model': 'string (default htdemucs, options: htdemucs/htdemucs_ft/mdx_extra_q)',
-            'task_id': 'string (optional, auto-generated)',
-        }
-    }, ensure_ascii=False)

app/api/task/index.dspy

@@ -1,17 +0,0 @@
-# -*- coding:utf-8 -*-
-# GET /api/task?task_id=xxx - 查询任务状态
-
-import json
-from ahserver.serverenv import ServerEnv
-
-task_id = params_kw.get('task_id', '')
-if not task_id:
-    return json.dumps({'error': 'task_id is required'}, ensure_ascii=False)
-
-env = ServerEnv()
-longtasks = env.longtasks
-if longtasks is None:
-    return json.dumps({'error': 'service not ready'}, ensure_ascii=False)
-
-status = await longtasks.get_status(task_id)
-return json.dumps(status)

app/health.dspy

@@ -1,3 +0,0 @@
-import json
-result = {"status": "ok", "service": "$svc"}
-print(json.dumps(result))

build.sh

@@ -1,150 +0,0 @@
-#!/bin/bash
-# 一键部署脚本模板
-# 用法: ./build.sh [deploy|update|stop|status]
-
-set -e
-
-SERVICE_NAME="demucs-service"
-GIT_REPO="git@git.opencomputing.cn:yumoqing/demucs-service.git"
-SERVICE_PORT=9083
-DEPLOY_DIR="/data/ymq/$SERVICE_NAME"
-VENV_PATH="/data/ymq/wan22-service/py3"
-GPU_ID="5"
-
-# 颜色输出
-RED='\033[0;31m'
-GREEN='\033[0;32m'
-YELLOW='\033[1;33m'
-NC='\033[0m'
-
-log_info() { echo -e "${GREEN}[INFO]${NC} $1"; }
-log_warn() { echo -e "${YELLOW}[WARN]${NC} $1"; }
-log_error() { echo -e "${RED}[ERROR]${NC} $1"; }
-
-check_deps() {
-    command -v git >/dev/null || { log_error "git not found"; exit 1; }
-    [ -f "$VENV_PATH/bin/python" ] || { log_error "Python venv not found: $VENV_PATH"; exit 1; }
-}
-
-deploy() {
-    log_info "Deploying $SERVICE_NAME..."
-    
-    # 检查依赖
-    check_deps
-    
-    # 克隆或更新代码
-    if [ -d "$DEPLOY_DIR/.git" ]; then
-        log_info "Updating existing deployment..."
-        cd "$DEPLOY_DIR"
-        git fetch origin
-        git reset --hard origin/master
-    else
-        log_info "Cloning repository..."
-        cd /data/ymq
-        git clone "$GIT_REPO" "$SERVICE_NAME"
-        cd "$DEPLOY_DIR"
-    fi
-    
-    # 创建必要目录
-    mkdir -p "$DEPLOY_DIR/app/api/status"
-    mkdir -p "$DEPLOY_DIR/app/api/submit"
-    mkdir -p "$DEPLOY_DIR/app/api/task"
-    
-    # 设置权限
-    chmod +x start.sh stop.sh 2>/dev/null || true
-    
-    # 启动服务
-    start_service
-}
-
-start_service() {
-    log_info "Starting $SERVICE_NAME on port $SERVICE_PORT..."
-    
-    # 停止旧进程
-    if [ -f stop.sh ]; then
-        bash stop.sh 2>/dev/null || true
-        sleep 2
-    fi
-    
-    # 启动新进程
-    bash start.sh
-    
-    # 等待启动
-    sleep 3
-    
-    # 验证
-    if ss -tlnp | grep -q ":$SERVICE_PORT "; then
-        log_info "✓ Service started successfully"
-        verify_api
-    else
-        log_error "✗ Service failed to start"
-        log_error "Check logs: $DEPLOY_DIR/nohup.out"
-        exit 1
-    fi
-}
-
-verify_api() {
-    log_info "Verifying API endpoints..."
-    
-    # 检查 status endpoint
-    if curl -s "http://127.0.0.1:$SERVICE_PORT/api/status" | grep -q "service"; then
-        log_info "✓ /api/status OK"
-    else
-        log_warn "✗ /api/status failed"
-    fi
-}
-
-stop_service() {
-    log_info "Stopping $SERVICE_NAME..."
-    if [ -f "$DEPLOY_DIR/stop.sh" ]; then
-        cd "$DEPLOY_DIR"
-        bash stop.sh
-        log_info "✓ Service stopped"
-    else
-        log_warn "stop.sh not found"
-    fi
-}
-
-show_status() {
-    echo "=== $SERVICE_NAME Status ==="
-    echo "Port: $SERVICE_PORT"
-    echo "Deploy Dir: $DEPLOY_DIR"
-    echo ""
-    
-    # 检查进程
-    if ss -tlnp | grep -q ":$SERVICE_PORT "; then
-        echo -e "Status: ${GREEN}RUNNING${NC}"
-        PID=$(ss -tlnp | grep ":$SERVICE_PORT " | grep -oP 'pid=\K[0-9]+')
-        echo "PID: $PID"
-    else
-        echo -e "Status: ${RED}STOPPED${NC}"
-    fi
-    echo ""
-    
-    # 检查 API
-    echo "API Endpoints:"
-    curl -s "http://127.0.0.1:$SERVICE_PORT/api/status" 2>/dev/null | python3 -m json.tool 2>/dev/null || echo "  (not responding)"
-}
-
-# 主入口
-case "${1:-deploy}" in
-    deploy|install)
-        deploy
-        ;;
-    update|upgrade)
-        deploy
-        ;;
-    stop)
-        stop_service
-        ;;
-    start)
-        start_service
-        ;;
-    status)
-        show_status
-        ;;
-    *)
-        echo "Usage: $0 {deploy|update|stop|start|status}"
-        exit 1
-        ;;
-esac

conf/config.json

@@ -1,38 +0,0 @@
-{
-    "password_key": "DemucsService2026Key",
-    "databases": {},
-    "session_redis": {
-        "host": "127.0.0.1",
-        "port": 6379,
-        "db": 1
-    },
-    "website": {
-        "paths": [
-            [
-                "$[workdir]$/app",
-                ""
-            ]
-        ],
-        "host": "0.0.0.0",
-        "port": 9083,
-        "coding": "utf-8",
-        "indexes": [
-            "index.html",
-            "index.dspy"
-        ],
-        "processors": [
-            [
-                ".dspy",
-                "dspy"
-            ]
-        ],
-        "startswiths": [
-            {
-                "leading": "/idfile",
-                "registerfunction": "idfile"
-            }
-        ]
-    },
-    "hot_reload": false,
-    "filesroot": "/tmp/demucs-outputs"
-}

start.sh

@@ -1,7 +0,0 @@
-#!/bin/bash
-cd /data/ymq/demucs-service
-export DEMUCS_GPU_ID=5
-export CUDA_VISIBLE_DEVICES=5
-export PYTHONPATH=/data/ymq/demucs-service
-nohup /data/ymq/wan22-service/py3/bin/python ah.py > nohup.out 2>&1 &
-echo "demucs-service started, PID: $!, GPU: $DEMUCS_GPU_ID"

stop.sh

@@ -1,26 +0,0 @@
-#!/bin/bash
-echo "Stopping demucs-service..."
-
-# Find and kill processes running ah.py for demucs-service
-PIDS=$(ps aux | grep '[d]emucs.*ah.py' | awk '{print $2}')
-
-if [ -z "$PIDS" ]; then
-    echo "No demucs-service processes found."
-    exit 0
-fi
-
-for pid in $PIDS; do
-    echo "Killing PID: $pid"
-    kill "$pid" 2>/dev/null
-done
-
-# Wait briefly then force kill if still running
-sleep 2
-for pid in $PIDS; do
-    if kill -0 "$pid" 2>/dev/null; then
-        echo "Force killing PID: $pid"
-        kill -9 "$pid" 2>/dev/null
-    fi
-done
-
-echo "demucs-service stopped."

workers/separate.py

@@ -1,87 +0,0 @@
-import os
-import json
-import asyncio
-import time
-from appPublic.log import debug, error
-
-async def run_separate(task_obj, payload):
-    """
-    Run demucs vocal/accompaniment separation.
-
-    payload:
-        audio_path (str, required): Path to input audio file
-        output_dir (str, optional): Output directory, default /tmp/demucs_{task_id}
-    """
-    audio_path = payload.get('audio_path')
-    if not audio_path:
-        raise ValueError('audio_path is required')
-
-    if not os.path.isfile(audio_path):
-        raise FileNotFoundError(f'Audio file not found: {audio_path}')
-
-    task_id = payload.get('task_id', str(int(time.time())))
-    output_dir = payload.get('output_dir', f'/tmp/demucs_{task_id}')
-
-    gpu_id = task_obj.gpu_id
-    basename = os.path.splitext(os.path.basename(audio_path))[0]
-
-    # Expected output paths from demucs
-    result_dir = os.path.join(output_dir, 'htdemucs', basename)
-    vocals_path = os.path.join(result_dir, 'vocals.wav')
-    no_vocals_path = os.path.join(result_dir, 'no_vocals.wav')
-
-    # Build the command
-    env = os.environ.copy()
-    env['CUDA_VISIBLE_DEVICES'] = str(gpu_id)
-
-    cmd = [
-        '/data/ymq/demucs_venv/bin/python', '-m', 'demucs',
-        '--two-stems', 'vocals',
-        audio_path,
-        '-o', output_dir
-    ]
-
-    debug(f'[demucs] Running separation: audio={audio_path}, output={output_dir}, gpu={gpu_id}')
-    debug(f'[demucs] Command: {" ".join(cmd)}')
-
-    start_time = time.time()
-
-    proc = await asyncio.create_subprocess_exec(
-        *cmd,
-        stdout=asyncio.subprocess.PIPE,
-        stderr=asyncio.subprocess.PIPE,
-        env=env
-    )
-
-    stdout, stderr = await proc.communicate()
-    elapsed = time.time() - start_time
-
-    if proc.returncode != 0:
-        stderr_text = stderr.decode('utf-8', errors='replace')
-        stdout_text = stdout.decode('utf-8', errors='replace')
-        error(f'[demucs] Process failed (rc={proc.returncode})')
-        error(f'[demucs] stdout: {stdout_text[-2000:]}')
-        error(f'[demucs] stderr: {stderr_text[-2000:]}')
-        raise RuntimeError(
-            f'Demucs separation failed (rc={proc.returncode}): {stderr_text[-500:]}'
-        )
-
-    # Verify output files exist
-    if not os.path.isfile(vocals_path):
-        raise FileNotFoundError(f'Expected vocals output not found: {vocals_path}')
-    if not os.path.isfile(no_vocals_path):
-        raise FileNotFoundError(f'Expected no_vocals output not found: {no_vocals_path}')
-
-    vocals_size = os.path.getsize(vocals_path)
-    no_vocals_size = os.path.getsize(no_vocals_path)
-
-    debug(f'[demucs] Separation complete in {elapsed:.1f}s')
-    debug(f'[demucs] vocals.wav: {vocals_size} bytes, no_vocals.wav: {no_vocals_size} bytes')
-
-    return {
-        'vocals_path': vocals_path,
-        'no_vocals_path': no_vocals_path,
-        'duration': round(elapsed, 2),
-        'output_dir': output_dir,
-        'model': 'htdemucs'
-    }