Clerk2.5/pdf_converter_v2/test_config.py

#!/usr/bin/env python3
# Copyright (c) Opendatalab. All rights reserved.

"""
配置文件测试脚本
验证配置文件是否正确加载
"""

import sys
from pathlib import Path

# 添加父目录到路径
sys.path.insert(0, str(Path(__file__).parent.parent))

from pdf_converter_v2.config import (
    DEVICE_KIND,
    DEFAULT_MODEL_NAME,
    DEFAULT_GPU_MEMORY_UTILIZATION,
    DEFAULT_DPI,
    DEFAULT_MAX_PAGES,
    DEFAULT_API_URL,
    DEFAULT_BACKEND,
    DEFAULT_PARSE_METHOD,
    DEFAULT_START_PAGE_ID,
    DEFAULT_END_PAGE_ID,
    DEFAULT_LANGUAGE,
    DEFAULT_RESPONSE_FORMAT_ZIP,
    DEFAULT_RETURN_MIDDLE_JSON,
    DEFAULT_RETURN_MODEL_OUTPUT,
    DEFAULT_RETURN_MD,
    DEFAULT_RETURN_IMAGES,
    DEFAULT_RETURN_CONTENT_LIST,
    DEFAULT_SERVER_URL,
    PADDLEOCR_CMD,
    PADDLE_OCR_DEVICE,
    PADDLE_OCR_DEVICES,
    PADDLE_DOC_PARSER_CMD,
    VL_REC_BACKEND,
    VL_REC_SERVER_URL,
    LOG_DIR,
    LOG_LEVEL,
    LOG_TO_FILE,
    LOG_TO_CONSOLE,
)


def print_config():
    """打印所有配置项"""
    print("=" * 80)
    print("PDF Converter v2 - 配置信息")
    print("=" * 80)
    print()
    
    print("【设备环境配置】")
    print(f"  设备类型: {DEVICE_KIND}")
    print()
    
    print("【模型配置】")
    print(f"  模型名称: {DEFAULT_MODEL_NAME}")
    print(f"  GPU 内存利用率: {DEFAULT_GPU_MEMORY_UTILIZATION}")
    print(f"  DPI: {DEFAULT_DPI}")
    print(f"  最大页数: {DEFAULT_MAX_PAGES}")
    print()
    
    print("【API 配置】")
    print(f"  API URL: {DEFAULT_API_URL}")
    print(f"  后端: {DEFAULT_BACKEND}")
    print(f"  解析方法: {DEFAULT_PARSE_METHOD}")
    print(f"  起始页ID: {DEFAULT_START_PAGE_ID}")
    print(f"  结束页ID: {DEFAULT_END_PAGE_ID}")
    print(f"  语言: {DEFAULT_LANGUAGE}")
    print(f"  服务器URL: {DEFAULT_SERVER_URL}")
    print()
    
    print("【PaddleOCR 配置】")
    print(f"  PaddleOCR 命令: {PADDLEOCR_CMD}")
    print(f"  推理设备: {PADDLE_OCR_DEVICE}")
    print(f"  多卡设备: {PADDLE_OCR_DEVICES}")
    print(f"  文档解析命令: {PADDLE_DOC_PARSER_CMD}")
    print(f"  VL 识别后端: {VL_REC_BACKEND}")
    print(f"  VL 识别服务器: {VL_REC_SERVER_URL}")
    print()
    
    print("【返回格式配置】")
    print(f"  返回ZIP格式: {DEFAULT_RESPONSE_FORMAT_ZIP}")
    print(f"  返回中间JSON: {DEFAULT_RETURN_MIDDLE_JSON}")
    print(f"  返回模型输出: {DEFAULT_RETURN_MODEL_OUTPUT}")
    print(f"  返回Markdown: {DEFAULT_RETURN_MD}")
    print(f"  返回图片: {DEFAULT_RETURN_IMAGES}")
    print(f"  返回内容列表: {DEFAULT_RETURN_CONTENT_LIST}")
    print()
    
    print("【日志配置】")
    print(f"  日志目录: {LOG_DIR}")
    print(f"  日志级别: {LOG_LEVEL}")
    print(f"  记录到文件: {LOG_TO_FILE}")
    print(f"  输出到控制台: {LOG_TO_CONSOLE}")
    print()
    
    print("=" * 80)
    print("✅ 配置加载成功！")
    print("=" * 80)


def main():
    """主函数"""
    try:
        print_config()
        return 0
    except Exception as e:
        print(f"❌ 配置加载失败: {e}", file=sys.stderr)
        import traceback
        traceback.print_exc()
        return 1


if __name__ == "__main__":
    sys.exit(main())