ecm_sync_system/scripts/legacy/run_full_sync_pipeline_simple.py

"""
run_full_sync_pipeline_simple.py

极简测试脚本：演示使用工厂函数创建 Pipeline
- 只需配置参数
- 调用工厂函数
- 执行 pipeline.run()
"""

import asyncio
import sys
from pathlib import Path

# 项目路径
PROJECT_ROOT = Path(__file__).resolve().parent.parent
if str(PROJECT_ROOT) not in sys.path:
    sys.path.insert(0, str(PROJECT_ROOT))

# 触发 Domain 注册
import sync_system_new.domain  # noqa: F401, E402

from sync_system_new.pipeline import create_jsonl_to_api_pipeline  # noqa: E402


# ========== 配置参数 ==========
CONFIG = {
    # 路径
    "local_jsonl_dir": str(PROJECT_ROOT / "filtered_datasource" / "datasource" / "ecm-2025-12-02"),
    "persistence_db": str(PROJECT_ROOT / "_runtime" / "test_simple.db"),
    
    # API
    "api_base_url": "http://localhost:8000/api/third/v2",
    "api_uid": "6b18c681-92d5-4c7f-802c-df413b6504c8",
    "api_secret": "1a28Az5dBD1Gc3URpmHwiR0XfzW7UtScQnqJQzgpzdg",
    
    # 节点类型
    "node_types": [
        "company",
        "supplier",
        # "user",
        "project",
        "project_detail_data",
        "preparation",
        "production",
        "lar",
        "lar_change",
        "units",
        "contract",
        "contract_change",
        "construction_task",
        "construction_task_detail",
        "material",
        "material_detail",
        "contract_settlement",
        "contract_settlement_detail",
        "personnel",
        "personnel_detail",
    ],
    
    # 项目过滤和预绑定
    "target_project_ids": ["f3e02e84-e81c-4aa6-88d4-f5aa108c8227"],
    "project_id_bindings": {
        "f3e02e84-e81c-4aa6-88d4-f5aa108c8227": "f3e02e84-e81c-4aa6-88d4-f5aa108c8227",
    },
    
    # Handler 自定义配置
    "handler_configs": {
        # supplier: 从持久化加载，不调用 API（后续运行时启用）
        "supplier": {"load_mode": "persisted_only"},
    },
    
    # Skip/Force Sync 控制
    # skip_sync_types: 跳过同步（只执行 bind，不执行 create/update）
    # 后续运行时可配置: "skip_sync_types": ["supplier"]
    "skip_sync_types": ["supplier"],
    # 行为
    "wipe_persistence_on_start": False,  # 改为 False，否则持久化数据会被清空
    "enable_persistence": True,
    
    # API 调试
    "api_debug": True,
    "poll_max_retries": 1,
    "poll_interval": 2.0,
}


async def main():
    """主函数 - 返回 True 表示成功，False 表示被中断"""
    from datetime import datetime
    import sys
    from pathlib import Path
    
    pipeline = None
    interrupted = False
    log_file = None
    original_stdout = None
    original_stderr = None
    
    # 日志设置
    log_dir = Path(PROJECT_ROOT) / "test_results"
    log_dir.mkdir(parents=True, exist_ok=True)
    log_path = log_dir / f"simple_pipeline_{datetime.now().strftime('%Y%m%d_%H%M%S')}.log"
    
    try:
        # 重定向输出到日志文件
        log_file = open(log_path, "w", encoding="utf-8")
        
        class Tee:
            def __init__(self, *streams):
                self._streams = streams
            def write(self, data):
                for stream in self._streams:
                    stream.write(data)
                    stream.flush()
                return len(data)
            def flush(self):
                for stream in self._streams:
                    stream.flush()
        
        original_stdout = sys.stdout
        original_stderr = sys.stderr
        sys.stdout = Tee(original_stdout, log_file)
        sys.stderr = Tee(original_stderr, log_file)
        
        print("\n" + "=" * 60)
        print("🚀 Creating Pipeline...")
        print(f"📝 Log file: {log_path}")
        print("=" * 60)
        
        # 一行创建完整 Pipeline
        pipeline = await create_jsonl_to_api_pipeline(**CONFIG)
        
        print(f"✅ Pipeline created!")
        print(f"   - Node types: {len(pipeline.node_types)}")
        print(f"   - Strategies: {len(pipeline.strategies)}")
        
        print("\n" + "=" * 60)
        print("🏃 Running Pipeline...")
        print("=" * 60)
        
        # 执行同步（print_summary 已整合到 stage4_post_check 中）
        stats = await pipeline.run()
        
        print("\n✅ Done!")
    
    except KeyboardInterrupt:
        print("\n\n⚠️ KeyboardInterrupt received, shutting down gracefully...")
        interrupted = True
    except Exception as e:
        print(f"\n\n❌ Error: {type(e).__name__}: {e}")
        import traceback
        traceback.print_exc()
    finally:
        if pipeline is not None:
            print("🧹 Cleaning up resources...")
            await pipeline.close()
            print("✅ Resources cleaned up")
        
        # 恢复原始输出
        if original_stdout is not None:
            sys.stdout = original_stdout
        if original_stderr is not None:
            sys.stderr = original_stderr
        if log_file is not None:
            log_file.close()
            if not interrupted:
                print(f"\n📝 Log saved to: {log_path}")
    
    return not interrupted  # 返回 False 表示被中断


if __name__ == "__main__":
    try:
        success = asyncio.run(main())
        sys.exit(0 if success else 130)  # 130 是标准的 Ctrl+C 退出码
    except KeyboardInterrupt:
        # asyncio.run() 可能会重新抛出 KeyboardInterrupt
        print("\n⚠️ Program interrupted")
        sys.exit(130)