ocr_platform/table_line_generator/backend/api/config.py

"""
配置 API 路由
提供数据源配置的读取和管理
"""

from fastapi import APIRouter, HTTPException
from pathlib import Path
from typing import List, Dict, Any
import yaml
from loguru import logger

from models.schemas import HealthResponse


router = APIRouter(prefix="/api/config", tags=["config"])


def _get_config_dir() -> Path:
    """获取配置目录路径"""
    return Path(__file__).parent.parent / "config" / "data_sources"


def _resolve_template(template: str, variables: Dict[str, str]) -> str:
    """解析模板变量"""
    resolved = template
    for key, value in variables.items():
        resolved = resolved.replace(f"{{{{{key}}}}}", value)
        resolved = resolved.replace(f"{{{{ {key} }}}}", value)
    return resolved


def _parse_data_source(config: Dict[str, Any]) -> Dict[str, Any]:
    """解析数据源配置，展开模板变量"""
    if 'data_source' not in config:
        return config
    
    ds = config['data_source']
    
    # 准备变量字典
    variables = {
        'name': ds.get('name', ''),
        'base_dir': ds.get('base_dir', '')
    }
    
    # 解析路径模板
    if 'json_dir' in ds:
        ds['json_dir'] = _resolve_template(ds['json_dir'], variables)
    if 'image_dir' in ds:
        ds['image_dir'] = _resolve_template(ds['image_dir'], variables)
    if 'output' in ds and 'directory' in ds['output']:
        ds['output']['directory'] = _resolve_template(ds['output']['directory'], variables)
    
    return ds


@router.get("/data-sources")
async def list_data_sources() -> Dict[str, List[Dict[str, Any]]]:
    """
    列出所有预定义数据源
    
    Returns:
        包含 sources 列表的字典，每个数据源包含 filename 字段（不含 .yaml 扩展名）
    """
    config_dir = _get_config_dir()
    sources = []
    
    try:
        if not config_dir.exists():
            logger.warning(f"配置目录不存在: {config_dir}")
            return {"sources": []}
        
        for yaml_file in sorted(config_dir.glob("*.yaml")):
            if yaml_file.name == "README.md":
                continue
                
            try:
                with open(yaml_file, 'r', encoding='utf-8') as f:
                    config = yaml.safe_load(f)
                
                if config and 'data_source' in config:
                    parsed = _parse_data_source(config)
                    # 添加文件名作为唯一标识符（不含 .yaml 扩展名）
                    filename = yaml_file.stem  # 获取不含扩展名的文件名
                    parsed['filename'] = filename
                    sources.append(parsed)
                    logger.debug(f"加载数据源配置: {yaml_file.name} (filename: {filename})")
            except Exception as e:
                logger.error(f"加载配置文件失败 {yaml_file}: {e}")
                continue
        
        logger.info(f"成功加载 {len(sources)} 个数据源配置")
        return {"sources": sources}
        
    except Exception as e:
        logger.exception(f"列出数据源失败: {e}")
        raise HTTPException(status_code=500, detail=f"列出数据源失败: {e}")


@router.get("/data-sources/{filename}")
async def get_data_source(filename: str) -> Dict[str, Any]:
    """
    获取指定数据源配置
    
    Args:
        filename: 数据源配置文件名（不含 .yaml 扩展名）
        
    Returns:
        数据源配置
    """
    config_dir = _get_config_dir()
    
    try:
        # 直接使用文件名查找配置文件
        yaml_file = config_dir / f"{filename}.yaml"
        
        if not yaml_file.exists():
            raise HTTPException(status_code=404, detail=f"数据源配置文件不存在: {filename}.yaml")
        
        with open(yaml_file, 'r', encoding='utf-8') as f:
            config = yaml.safe_load(f)
        
        if config and 'data_source' in config:
            parsed = _parse_data_source(config)
            # 添加文件名作为唯一标识符
            parsed['filename'] = filename
            return parsed
        else:
            raise HTTPException(status_code=404, detail=f"配置文件格式错误: {filename}.yaml")
        
    except HTTPException:
        raise
    except Exception as e:
        logger.exception(f"获取数据源失败: {e}")
        raise HTTPException(status_code=500, detail=f"获取数据源失败: {e}")


@router.get("/health", response_model=HealthResponse)
async def config_health_check():
    """配置服务健康检查"""
    return HealthResponse(
        status="ok",
        service="config"
    )