LQAdminPlatform/scripts/miner_u.py

import os
import sys
import logging
from urllib.parse import urlparse

# 配置日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger("MinerU")

# 导入配置和管理器
sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', 'src'))
from app.base.mineru_connection import get_mineru_manager
from app.base.async_mysql_connection import get_db_connection

def main_cli(doc_id):
    """
    MinerU 命令行入口，供后台进程调用
    """
    manager = get_mineru_manager()
    conn = get_db_connection()
    if not conn:
        logger.error("数据库连接失败")
        return
        
    try:
        with conn.cursor() as cursor:
            # 2. 获取文档基本信息 (主表即为唯一资产中心)
            cursor.execute("SELECT title, file_url FROM t_samp_document_main WHERE id = %s", (doc_id,))
            row = cursor.fetchone()
            if not row:
                logger.warning(f"文档不存在: {doc_id}")
                return
            
            title = row['title']
            file_url = row['file_url']

            if not file_url:
                logger.error(f"未找到文件链接: {doc_id}")
                manager.update_db_status(doc_id, status=3, error="未找到文件链接(file_url)")
                return
            
            # 3. 调用管理器执行转换
            logger.info(f"开始处理文档 [{doc_id}]: {title}")
            manager.process_document(doc_id, title, file_url)
            
    except Exception as e:
        logger.exception(f"处理文档 [{doc_id}] 时发生未捕获异常: {e}")
        manager.update_db_status(doc_id, status=3, error=str(e))
    finally:
        conn.close()

if __name__ == "__main__":
    if len(sys.argv) > 1:
        # 兼容旧格式: python miner_u.py <table_type> <doc_id>
        # 兼容新格式: python miner_u.py <doc_id>
        target_id = sys.argv[-1]
        main_cli(target_id)
    else:
        print("Usage: python miner_u.py <doc_id>")