# ============================================================================
# PDF 批量处理器配置文件
# ============================================================================

# 处理器定义
processors:
  # -------------------------------------------------------------------------
  # PaddleOCR-VL 处理器
  # -------------------------------------------------------------------------
  paddleocr_vl_single_process:
    script: "/Users/zhch158/workspace/repository.git/PaddleX/zhch/paddleocr_vl_single_process.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--pipeline=/Users/zhch158/workspace/repository.git/PaddleX/zhch/my_config/PaddleOCR-VL-Client-RT-DETR-H_layout_17cls.yaml"
      - "--device=cpu"
      # - "--no-adapter"
    output_subdir: "paddleocr_vl_results"
    log_subdir: "logs/paddleocr_vl"  # 🎯 新增：日志子目录
    venv: "source /Users/zhch158/workspace/repository.git/PaddleX/paddle_env/bin/activate"
    description: "PaddleOCR-VL 处理器 - 视觉语言模型OCR"

  # -------------------------------------------------------------------------
  # PP-StructureV3 本地处理器
  # -------------------------------------------------------------------------
  ppstructurev3_single_process:
    script: "/home/ubuntu/zhch/PaddleX/zhch/ppstructurev3_single_process.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--pipeline=/home/ubuntu/zhch/PaddleX/zhch/my_config/PP-StructureV3.yaml"
      - "--device=cpu"
    output_subdir: "ppstructurev3_results"
    log_subdir: "logs/ppstructurev3"
    venv: "conda activate paddle"
    description: "PP-StructureV3 处理器 - 本地处理"

  ppstructurev3_gpu:
    script: "/home/ubuntu/zhch/PaddleX/zhch/ppstructurev3_single_process.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--pipeline=/home/ubuntu/zhch/PaddleX/zhch/my_config/PP-StructureV3.yaml"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--pipeline=/home/ubuntu/zhch/PaddleX/zhch/my_config/PP-StructureV3.yaml"
      - "--device=gpu"
    output_subdir: "ppstructurev3_gpu_results"
    log_subdir: "logs/ppstructurev3_gpu"
    venv: "conda activate paddle"
    description: "PP-StructureV3 处理器 - GPU加速"

  # -------------------------------------------------------------------------
  # PP-StructureV3 CPU 处理器
  # 明确使用 CPU 处理
  # -------------------------------------------------------------------------
  ppstructurev3_cpu:
    script: "/Users/zhch158/workspace/repository.git/PaddleX/zhch/ppstructurev3_single_process.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--pipeline=/Users/zhch158/workspace/repository.git/PaddleX/zhch/my_config/PP-StructureV3-zhch.yaml"
      - "--device=cpu"
    output_subdir: "ppstructurev3_cpu_results"
    log_subdir: "logs/ppstructurev3_cpu"
    venv: "source /Users/zhch158/workspace/repository.git/PaddleX/paddle_env/bin/activate"
    description: "PP-StructureV3 处理器 - CPU处理"

  # -------------------------------------------------------------------------
  # PP-StructureV3 API 客户端 (默认)
  # 通过 HTTP API 调用远程服务
  # -------------------------------------------------------------------------
  ppstructurev3_single_client:
    script: "/Users/zhch158/workspace/repository.git/PaddleX/zhch/ppstructurev3_single_client.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--api_url=http://10.192.72.11:8111/layout-parsing"
      - "--timeout=300"
    output_subdir: "ppstructurev3_client_results"
    log_subdir: "logs/ppstructurev3_client"
    venv: "source /Users/zhch158/workspace/repository.git/PaddleX/paddle_env/bin/activate"
    description: "PP-StructureV3 HTTP API 客户端 - 远程服务"

  # -------------------------------------------------------------------------
  # MinerU vLLM 处理器
  # 基于 MinerU 的多线程批量处理（支持 PDF 和图片）
  # -------------------------------------------------------------------------
  mineru_vllm:
    script: "/Users/zhch158/workspace/repository.git/MinerU/zhch/mineru2_vllm_multthreads.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--server_url=http://10.192.72.11:8121"
      - "--timeout=300"
      - "--batch_size=1"
    output_subdir: "mineru_vllm_results"
    log_subdir: "logs/mineru_vllm"
    venv: "conda activate mineru2"
    description: "MinerU vLLM 处理器 - 支持PDF和图片"

  # -------------------------------------------------------------------------
  # DotsOCR vLLM 处理器
  # 基于 DotsOCR 的批量处理（支持 PDF 和图片）
  # -------------------------------------------------------------------------
  dotsocr_vllm:
    script: "/Users/zhch158/workspace/repository.git/dots.ocr/zhch/dotsocr_vllm_multthreads.py"
    input_arg: "--input_file"
    output_arg: "--output_dir"
    extra_args:
      - "--ip=10.192.72.11"
      - "--port=8101"
      - "--model_name=DotsOCR"
      - "--prompt_mode=prompt_layout_all_en"
      - "--batch_size=1"
      - "--max_workers=1"
      - "--dpi=200"
    output_subdir: "dotsocr_vllm_results"
    log_subdir: "logs/dotsocr_vllm"
    venv: "conda activate py312"
    description: "DotsOCR vLLM 处理器 - 支持PDF和图片"

# ============================================================================
# 全局配置
# ============================================================================
global:
  # PDF 文件基础目录
  base_dir: "/Users/zhch158/workspace/data/流水分析"
  
  # 默认输出子目录名称（如果处理器未指定）
  output_subdir: "results"
  
  # 🎯 新增：全局日志配置
  log_dir: "logs"  # 全局日志目录（相对于 base_dir）
  log_retention_days: 30  # 日志保留天数
  log_level: "INFO"  # 日志级别: DEBUG, INFO, WARNING, ERROR