#!/usr/bin/env python3
"""
QwenPaw + MetaGPT + ECC + AnySearch 全套流水线

使用方式：
  1. 确保 metaGPT 和 ECC 已安装配置
  2. 运行: python3 pipeline_demo.py "你的需求"
"""

import subprocess
import sys
import json


def search_reference(query: str):
    """Step 1: AnySearch 搜索参考资料"""
    print(f"\n{'='*60}")
    print(f"🔍 [1/4] AnySearch 搜索: {query}")
    print(f"{'='*60}")
    # 通过 curl 调 AnySearch MCP
    cmd = f'''curl -s 'https://api.anysearch.com/mcp' -X POST \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer as_sk_f92eddbbd0360ccd793ac7aa5a391f77' \
  -d '{{"jsonrpc":"2.0","method":"tools/call","id":1,"params":{{"name":"search","arguments":{{"query":"{query}","max_results":3}}}}}}' '''
    result = subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=15)
    print(f"✅ 搜索完成")
    return result.stdout


def generate_code(requirement: str):
    """Step 2: MetaGPT 生成代码"""
    print(f"\n{'='*60}")
    print(f"🏢 [2/4] MetaGPT 生成项目: {requirement}")
    print(f"{'='*60}")
    result = subprocess.run(
        f'metagpt "{requirement}" --n-round 1',
        shell=True, capture_output=True, text=True, timeout=120
    )
    print(f"✅ MetaGPT 生成完成")
    return result


def review_code(project_path: str):
    """Step 3: ECC 代码审查"""
    print(f"\n{'='*60}")
    print(f"🔧 [3/4] ECC 代码审查")
    print(f"{'='*60}")
    # ECC 审查通过 MCP 工具调用（在 QwenPaw 会话中自动可用）
    print(f"✅ ECC 准备就绪")


def deploy():
    """Step 4: ECC 部署"""
    print(f"\n{'='*60}")
    print(f"🚀 [4/4] ECC 部署上线")
    print(f"{'='*60}")
    print(f"✅ 部署准备就绪")


if __name__ == "__main__":
    requirement = " ".join(sys.argv[1:]) or "创建一个Python命令行计算器"
    
    print(f"🔥 QwenPaw + MetaGPT + ECC + AnySearch 全套流水线")
    print(f"📝 需求: {requirement}")
    
    # 1. 搜资料
    search_reference(f"{requirement} 技术方案")
    
    # 2. 生成代码
    result = generate_code(requirement)
    
    # 3. 审查
    review_code("./workspace")
    
    # 4. 部署
    deploy()
    
    print(f"\n{'='*60}")
    print(f"🎉 流水线执行完成!")
    print(f"{'='*60}")