ewoooc/services/crawler_config_loader.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
爬蟲配置載入器
提供動態讀取和管理爬蟲配置的功能
"""
import os
import json
from typing import Dict, Any, Optional

CONFIG_PATH = os.path.join(os.path.dirname(__file__), 'data', 'crawler_config.json')

def load_crawler_config() -> Dict[str, Any]:
    """
    載入爬蟲配置文件

    Returns:
        dict: 完整的配置字典
    """
    try:
        with open(CONFIG_PATH, 'r', encoding='utf-8') as f:
            return json.load(f)
    except FileNotFoundError:
        # 如果配置文件不存在，返回預設配置
        return get_default_config()
    except json.JSONDecodeError as e:
        print(f"❌ 配置文件格式錯誤: {e}")
        return get_default_config()

def get_default_config() -> Dict[str, Any]:
    """返回預設配置"""
    return {
        "crawlers": {
            "momo_main": {"enabled": True, "schedule_hours": 4},
            "edm_promo": {"enabled": True, "schedule_hours": 4},
            "festival_11": {"enabled": False, "schedule_hours": 6},
            "mothers_day_2026": {"enabled": False, "schedule_hours": 4, "lpn_code": "", "activity_name": "母親節超值限時購", "page_type": "mothers_day"},
            "valentine_520_2026": {"enabled": False, "schedule_hours": 4, "lpn_code": "", "activity_name": "520情人節限定購物", "page_type": "valentine_520"},
            "labor_day_2026": {"enabled": False, "schedule_hours": 4, "lpn_code": "", "activity_name": "勞動節購物優惠", "page_type": "labor_day"}
        },
        "settings": {
            "auto_import_enabled": True,
            "auto_import_interval_minutes": 30
        }
    }

def is_crawler_enabled(crawler_key: str) -> bool:
    """
    檢查指定爬蟲是否啟用

    Args:
        crawler_key: 爬蟲識別碼 (momo_main, edm_promo, festival_11)

    Returns:
        bool: True 表示啟用，False 表示停用
    """
    config = load_crawler_config()
    crawler = config.get('crawlers', {}).get(crawler_key, {})
    return crawler.get('enabled', False)

def get_crawler_schedule(crawler_key: str) -> Optional[int]:
    """
    取得爬蟲的執行頻率（小時）

    Args:
        crawler_key: 爬蟲識別碼

    Returns:
        int: 執行頻率（小時），如果爬蟲停用則返回 None
    """
    if not is_crawler_enabled(crawler_key):
        return None

    config = load_crawler_config()
    crawler = config.get('crawlers', {}).get(crawler_key, {})
    return crawler.get('schedule_hours', 4)

def get_crawler_info(crawler_key: str) -> Optional[Dict[str, Any]]:
    """
    取得爬蟲的完整資訊

    Args:
        crawler_key: 爬蟲識別碼

    Returns:
        dict: 爬蟲資訊，如果不存在則返回 None
    """
    config = load_crawler_config()
    return config.get('crawlers', {}).get(crawler_key)

def get_enabled_crawlers() -> Dict[str, Dict[str, Any]]:
    """
    取得所有啟用的爬蟲

    Returns:
        dict: 啟用的爬蟲字典 {crawler_key: crawler_info}
    """
    config = load_crawler_config()
    crawlers = config.get('crawlers', {})
    return {key: info for key, info in crawlers.items() if info.get('enabled', False)}

def get_paused_crawlers() -> Dict[str, Dict[str, Any]]:
    """
    取得所有暫停的爬蟲

    Returns:
        dict: 暫停的爬蟲字典 {crawler_key: crawler_info}
    """
    config = load_crawler_config()
    crawlers = config.get('crawlers', {})
    return {key: info for key, info in crawlers.items() if not info.get('enabled', False)}

def update_crawler_status(crawler_key: str, enabled: bool, reason: str = "") -> bool:
    """
    更新爬蟲啟用狀態

    Args:
        crawler_key: 爬蟲識別碼
        enabled: True 啟用，False 停用
        reason: 停用原因（可選）

    Returns:
        bool: 更新成功返回 True
    """
    try:
        config = load_crawler_config()

        if crawler_key not in config.get('crawlers', {}):
            print(f"❌ 爬蟲 {crawler_key} 不存在")
            return False

        config['crawlers'][crawler_key]['enabled'] = enabled

        if not enabled and reason:
            config['crawlers'][crawler_key]['pause_reason'] = reason
            config['crawlers'][crawler_key]['status'] = 'paused'
        elif enabled:
            config['crawlers'][crawler_key]['status'] = 'active'
            config['crawlers'][crawler_key].pop('pause_reason', None)

        # 更新時間戳
        from datetime import datetime
        config['metadata']['last_updated'] = datetime.now().isoformat()

        # 寫回配置文件
        with open(CONFIG_PATH, 'w', encoding='utf-8') as f:
            json.dump(config, f, ensure_ascii=False, indent=2)

        return True
    except Exception as e:
        print(f"❌ 更新配置失敗: {e}")
        return False

def print_crawler_status():
    """印出所有爬蟲的狀態（用於檢查）"""
    config = load_crawler_config()
    crawlers = config.get('crawlers', {})

    print("\n" + "=" * 60)
    print("爬蟲配置狀態")
    print("=" * 60)

    for key, info in crawlers.items():
        status_icon = "✅" if info.get('enabled') else "⏸️"
        print(f"\n{status_icon} [{key}] {info.get('name', 'N/A')}")
        print(f"   狀態: {info.get('status', 'unknown')}")
        print(f"   啟用: {info.get('enabled', False)}")

        if not info.get('enabled'):
            print(f"   暫停原因: {info.get('pause_reason', 'N/A')}")
            print(f"   暫停日期: {info.get('paused_date', 'N/A')}")
        else:
            print(f"   執行頻率: 每 {info.get('schedule_hours', 'N/A')} 小時")

        if 'notes' in info:
            print(f"   備註: {info.get('notes')}")

    print("\n" + "=" * 60 + "\n")

if __name__ == "__main__":
    # 測試：印出當前配置
    print_crawler_status()