平台政策变更自适应监控 — 跨境电商合规政策智能预警与快速响应

"""
平台政策变更自适应监控系统
功能：语义变更检测 + 影响分级 + 自动响应触发
"""
import numpy as np
import hashlib
import re
from dataclasses import dataclass, field
from typing import List, Dict, Optional
from datetime import datetime, timedelta
import warnings
warnings.filterwarnings('ignore')


@dataclass
class PolicyPage:
    """政策页面数据结构"""
    url: str
    platform: str  # 'amazon', 'shopee', 'tiktok'
    category: str  # 'product_safety', 'ad_policy', 'seller_rules'
    last_hash: Optional[str] = None
    last_content: Optional[str] = None
    last_checked: Optional[datetime] = None


@dataclass
class PolicyChange:
    """政策变更记录"""
    page_url: str
    platform: str
    detected_at: datetime
    old_snippet: str
    new_snippet: str
    semantic_similarity: float
    impact_level: str  # 'critical', 'high', 'low'
    affected_categories: List[str] = field(default_factory=list)
    summary: str = ""


def simple_sentence_similarity(text1: str, text2: str) -> float:
    """
    简化版语义相似度（生产环境替换为 Sentence-BERT）
    使用词袋模型 + Jaccard 相似度作为近似
    """
    def tokenize(text):
        return set(re.findall(r'\b\w+\b', text.lower()))
    
    words1 = tokenize(text1)
    words2 = tokenize(text2)
    
    if not words1 and not words2:
        return 1.0
    intersection = words1 & words2
    union = words1 | words2
    return len(intersection) / len(union)


def detect_policy_changes(old_content: str, new_content: str, 
                          similarity_threshold: float = 0.85) -> List[Dict]:
    """