ai-contract-platform/scripts/check_law_updates.py

#!/usr/bin/env python3
"""
法律库更新检查脚本

功能：
1. 检查 LawRefBook/Laws 仓库是否有更新
2. 检测新增法律和修订内容
3. 过滤过期法律
4. 生成更新报告

用法：
    python scripts/check_law_updates.py
"""

import os
import sys
import json
import subprocess
from datetime import datetime
from pathlib import Path

# 配置
REPO_URL = "https://github.com/LawRefBook/Laws.git"
REPO_DIR = "temp-lawrefbook-updates"
REPORT_FILE = "project-docs/phase14-legal-fulltext-library/update-check-report.json"
DATABASE_URL = os.environ.get("DATABASE_URL", "")

def check_repo_updates():
    """检查仓库是否有新提交"""
    print("Checking LawRefBook/Laws repository for updates...")

    # 克隆或更新仓库
    if os.path.exists(REPO_DIR):
        print(f"  Updating existing clone: {REPO_DIR}")
        subprocess.run(["git", "fetch"], cwd=REPO_DIR, capture_output=True)
        subprocess.run(["git", "reset", "--hard", "origin/main"], cwd=REPO_DIR, capture_output=True)
    else:
        print(f"  Cloning repository: {REPO_URL}")
        subprocess.run(["git", "clone", REPO_URL, REPO_DIR], capture_output=True)

    # 获取最新提交信息
    result = subprocess.run(
        ["git", "log", "-1", "--format=%H|%ci|%s"],
        cwd=REPO_DIR,
        capture_output=True,
        text=True
    )

    if result.returncode == 0:
        commit_hash, commit_date, commit_msg = result.stdout.strip().split("|")
        print(f"  Latest commit: {commit_hash[:8]} ({commit_date})")
        print(f"  Message: {commit_msg}")
        return commit_hash, commit_date, commit_msg

    return None, None, None

def check_database_laws():
    """检查数据库中已有的法律"""
    if not DATABASE_URL:
        print("  Warning: DATABASE_URL not set, skipping DB check")
        return []

    import psycopg
    conn = psycopg.connect(DATABASE_URL)
    cur = conn.cursor()

    cur.execute("SELECT DISTINCT source_id FROM legal_articles")
    existing = [row[0] for row in cur.fetchall()]

    conn.close()
    return existing

def scan_new_laws(existing_sources):
    """扫描新法律"""
    print("Scanning for new laws...")

    def check_validity(file_path):
        try:
            with open(file_path, 'r', encoding='utf-8') as f:
                content = f.read()
            markers = ['已废止', '已失效', '已修改', '已修订', '被废止', '被修改', '废止', '失效']
            return not any(m in content for m in markers)
        except:
            return True

    new_laws = []
    expired_laws = []

    for root, dirs, files in os.walk(REPO_DIR):
        for file in files:
            if not file.endswith('.md') or file.startswith('_'):
                continue

            file_path = os.path.join(root, file)
            rel_path = os.path.relpath(file_path, REPO_DIR)

            # 跳过DLC目录（可选）
            if rel_path.startswith('DLC/'):
                continue

            # 生成source_id
            import re
            with open(file_path, 'r', encoding='utf-8') as f:
                content = f.read()
            title_match = re.search(r'^#\s+(.+)$', content, re.MULTILINE)
            title = title_match.group(1).strip() if title_match else file
            source_id = re.sub(r'[^\w]', '-', title[:50]).strip('-')

            # 检查是否已存在
            if source_id in existing_sources:
                continue

            # 检查时效性
            if check_validity(file_path):
                new_laws.append({
                    'source_id': source_id,
                    'title': title,
                    'path': rel_path
                })
            else:
                expired_laws.append({
                    'source_id': source_id,
                    'title': title,
                    'path': rel_path
                })

    return new_laws, expired_laws

def generate_report(commit_hash, commit_date, new_laws, expired_laws):
    """生成更新报告"""
    report = {
        'check_time': datetime.now().isoformat(),
        'repo_commit': commit_hash,
        'repo_date': commit_date,
        'new_laws_count': len(new_laws),
        'expired_laws_count': len(expired_laws),
        'new_laws': new_laws[:20],  # 只保存前20个作为示例
        'expired_laws': expired_laws[:20],
        'action_required': len(new_laws) > 0 or len(expired_laws) > 0
    }

    with open(REPORT_FILE, 'w', encoding='utf-8') as f:
        json.dump(report, f, ensure_ascii=False, indent=2)

    print(f"Report saved: {REPORT_FILE}")
    return report

def main():
    print("=" * 60)
    print("Law Library Update Checker")
    print("=" * 60)
    print()

    # 1. 检查仓库更新
    commit_hash, commit_date, commit_msg = check_repo_updates()

    # 2. 检查数据库已有法律
    existing_sources = check_database_laws()
    print(f"  Existing laws in database: {len(existing_sources)}")

    # 3. 扫描新法律
    new_laws, expired_laws = scan_new_laws(existing_sources)

    print()
    print("=" * 60)
    print("Update Check Results")
    print("=" * 60)
    print(f"New laws found: {len(new_laws)}")
    print(f"Expired laws detected: {len(expired_laws)}")

    # 4. 生成报告
    report = generate_report(commit_hash, commit_date, new_laws, expired_laws)

    # 5. 建议
    if report['action_required']:
        print()
        print("Action required:")
        if new_laws:
            print(f"  - Import {len(new_laws)} new laws")
        if expired_laws:
            print(f"  - Remove {len(expired_laws)} expired laws")
    else:
        print()
        print("No updates needed. Law library is up-to-date.")

    return 0

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