import asyncio
import logging
from sqlalchemy import text
from sqlalchemy.ext.asyncio import create_async_engine, AsyncSession
from sqlalchemy.future import select

# 导入所有模型
from app.models import Base, Company, Report, ExtractedContent, AnalysisResult, TaskLog
from app.config import settings

# 配置日志
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# 旧数据库 (SQLite)
SQLITE_URL = "sqlite+aiosqlite:///./qingbao.db"

# 新数据库 (PostgreSQL) - 从配置读取
PG_URL = settings.DATABASE_URL
# 如果上面读取不正确，请手动取消下面的注释并修改
# PG_URL = "postgresql+asyncpg://user_xREpkJ:password_DZz8DQ@331002.xyz:2022/qingbao"

async def migrate():
    logger.info("开始数据迁移...")
    
    # 创建引擎
    sqlite_engine = create_async_engine(SQLITE_URL)
    pg_engine = create_async_engine(PG_URL)

    async with sqlite_engine.connect() as sqlite_conn, pg_engine.connect() as pg_conn:
        # 0. 确认连接
        logger.info("正在连接数据库...")
        
        # 定义需要迁移的表（按顺序）
        # Company, Report, ExtractedContent, AnalysisResult, TaskLog
        # 使用 SQLAlchemy Core 的 insert() 可以在不同方言间自动转换类型
        from sqlalchemy import insert
        
        # 1. Companies
        logger.info("正在迁移 Companies...")
        result = await sqlite_conn.execute(select(Company))
        companies = result.all() # ORM 模型对象或者 Row 对象
        
        if companies:
            # 将 Row 对象转换为字典
            # 注意：从 SQLite 读出来的 datetime 如果是字符串，PG 可能需要 parsing
            # 但如果我们用 ORM 模型定义了 DateTime，SQLAlchemy 应该已经处理了
            
            # 使用字典插入
            values = []
            for row in companies:
                row_dict = dict(row._mapping)
                # 确保布尔值正确
                if 'is_active' in row_dict:
                    row_dict['is_active'] = bool(row_dict['is_active'])
                values.append(row_dict)
                
            stmt = insert(Company).values(values)
            # PG 不支持 INSERT OR IGNORE，用 ON CONFLICT DO NOTHING
            from sqlalchemy.dialects.postgresql import insert as pg_insert
            stmt = pg_insert(Company).values(values).on_conflict_do_nothing(index_elements=['id'])
            
            await pg_conn.execute(stmt)
            logger.info(f"已迁移 {len(values)} 家公司")
            await pg_conn.execute(text("SELECT setval(pg_get_serial_sequence('companies', 'id'), (SELECT MAX(id) FROM companies))"))

        # 2. Reports
        logger.info("正在迁移 Reports...")
        result = await sqlite_conn.execute(select(Report))
        reports = result.all()
        
        if reports:
            values = []
            for row in reports:
                row_dict = dict(row._mapping)
                # 转换布尔值
                for bool_field in ['is_downloaded', 'is_extracted', 'is_analyzed']:
                    if bool_field in row_dict:
                        row_dict[bool_field] = bool(row_dict[bool_field])
                values.append(row_dict)
                
            stmt = pg_insert(Report).values(values).on_conflict_do_nothing(index_elements=['id'])
            await pg_conn.execute(stmt)
            logger.info(f"已迁移 {len(values)} 份报告")
            await pg_conn.execute(text("SELECT setval(pg_get_serial_sequence('reports', 'id'), (SELECT MAX(id) FROM reports))"))

        # 3. ExtractedContent
        logger.info("正在迁移 ExtractedContent...")
        result = await sqlite_conn.execute(select(ExtractedContent))
        contents = result.all()
        
        if contents:
            values = []
            for row in contents:
                row_dict = dict(row._mapping)
                values.append(row_dict)
            
            # 分批
            batch_size = 50
            for i in range(0, len(values), batch_size):
                batch = values[i:i+batch_size]
                stmt = pg_insert(ExtractedContent).values(batch).on_conflict_do_nothing(index_elements=['id'])
                await pg_conn.execute(stmt)
                logger.info(f"  - 已插入 {min(i+batch_size, len(values))}/{len(values)}")
            
            await pg_conn.execute(text("SELECT setval(pg_get_serial_sequence('extracted_contents', 'id'), (SELECT MAX(id) FROM extracted_contents))"))

        # 4. AnalysisResult
        logger.info("正在迁移 AnalysisResult...")
        result = await sqlite_conn.execute(select(AnalysisResult))
        results = result.all()
        
        if results:
            values = []
            for row in results:
                row_dict = dict(row._mapping)
                if 'is_final' in row_dict:
                    row_dict['is_final'] = bool(row_dict['is_final'])
                values.append(row_dict)
                
            stmt = pg_insert(AnalysisResult).values(values).on_conflict_do_nothing(index_elements=['id'])
            await pg_conn.execute(stmt)
            logger.info(f"已迁移 {len(values)} 条分析结果")
            await pg_conn.execute(text("SELECT setval(pg_get_serial_sequence('analysis_results', 'id'), (SELECT MAX(id) FROM analysis_results))"))

        # 5. TaskLog
        logger.info("正在迁移 TaskLog...")
        result = await sqlite_conn.execute(text("SELECT * FROM task_logs")) # TaskLog 可能没有对应 ORM 或者有变化，用 text 读
        # 注意：这里如果用 text 读，列名需要匹配 TaskLog 模型
        # 为了安全，我们还是尽量用 ORM
        # 如果 TaskLog 定义了
        try:
            result = await sqlite_conn.execute(select(TaskLog))
            logs = result.all()
            if logs:
                values = [dict(row._mapping) for row in logs]
                stmt = pg_insert(TaskLog).values(values).on_conflict_do_nothing(index_elements=['id'])
                await pg_conn.execute(stmt)
                logger.info(f"已迁移 {len(values)} 条日志")
        except Exception as e:
            logger.warning(f"TaskLog 迁移跳过（可能表结构不一致）: {e}")
            
        await pg_conn.execute(text("SELECT setval(pg_get_serial_sequence('task_logs', 'id'), (SELECT MAX(id) FROM task_logs))"))

        await pg_conn.commit()
    
    await sqlite_engine.dispose()
    await pg_engine.dispose()
    logger.info("数据迁移完成！")

if __name__ == "__main__":
    asyncio.run(migrate())