#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# 太和县 - 政务公告
# CMS: 自定义
# List: /Content/showList/149/page_{N}.html (20/page)
# Content: div.content
# Date: meta.PubDate

import requests
import re
import sqlite3
from bs4 import BeautifulSoup
from datetime import datetime

BASE_URL = "https://www.taihe.gov.cn"
LIST_URL = "https://www.taihe.gov.cn/Content/showList/149/page_{}.html"
HEADERS = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
}
DB_PATH = "/root/search.db"
MAX_PAGES = 5

def get_pages():
    """Get all article URLs from list pages"""
    articles = []
    for page_num in range(1, MAX_PAGES + 1):
        url = LIST_URL.format(page_num)
        
        try:
            r = requests.get(url, headers=HEADERS, verify=False, timeout=30)
            r.encoding = "utf-8"
            if r.status_code != 200:
                continue
            
            soup = BeautifulSoup(r.text, "html.parser")
            
            # Find the ul with most date-stamped items
            best_ul = None
            best_count = 0
            for ul in soup.find_all("ul"):
                count = 0
                for li in ul.find_all("li"):
                    span = li.find("span")
                    a = li.find("a")
                    if span and a and span.get_text(strip=True).startswith("20"):
                        count += 1
                if count > best_count:
                    best_count = count
                    best_ul = ul
            
            if not best_ul:
                continue
            
            count = 0
            for li in best_ul.find_all("li"):
                a = li.find("a")
                span = li.find("span")
                if a and span:
                    href = a.get("href", "")
                    # Use title attribute for clean title
                    title = a.get("title", "")
                    date_str = span.get_text(strip=True)[:10]
                    
                    if href and title:
                        if href.startswith("/"):
                            full_url = BASE_URL + href
                        else:
                            full_url = href
                        
                        articles.append({
                            "url": full_url,
                            "title": title,
                            "date": date_str
                        })
                        count += 1
            
            print(f"  [OK] Page {page_num}: {count} items")
            
        except Exception as e:
            print(f"  [ERR] Page {page_num}: {e}")
    
    return articles

def get_detail(url):
    """Fetch detail page and extract content and date"""
    try:
        r = requests.get(url, headers=HEADERS, verify=False, timeout=30)
        r.encoding = "utf-8"
        if r.status_code != 200 or len(r.text) < 1000:
            return None, None
        
        soup = BeautifulSoup(r.text, "html.parser")
        
        # Content: m-dtfonts (primary), fallback m-dttexts
        content = None
        for cls in ["m-dtfonts", "m-dttexts"]:
            div = soup.find("div", class_=cls)
            if div:
                inner = "".join(str(tag) for tag in div.contents).strip()
                text_only = re.sub(r'<[^>]+>', '', inner).strip()
                if len(text_only) >= 50:
                    content = inner
                    break
        
        if not content:
            return None, None
        
        # Date from PubDate meta
        pub_date = ""
        meta = soup.find("meta", attrs={"name": "PubDate"})
        if meta and meta.get("content"):
            pub_date = meta["content"].strip()[:10]
        
        return content, pub_date
    
    except Exception as e:
        print(f"  [ERR] {url[:80]}: {e}")
        return None, None

def main():
    print("=" * 60)
    print("太和县 - 政务公告 爬虫")
    print(f"启动时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
    print("=" * 60)
    
    print("\n[1/3] 抓取列表页 (前5页)...")
    articles = get_pages()
    print(f"  共找到 {len(articles)} 条记录")
    
    if not articles:
        return
    
    print("\n[2/3] 检查数据库去重...")
    conn = sqlite3.connect(DB_PATH, timeout=60)
    c = conn.cursor()
    
    c.execute("SELECT page_url FROM gov_raw")
    existing = set(row[0] for row in c.fetchall())
    
    new_articles = [a for a in articles if a["url"] not in existing]
    print(f"  新增: {len(new_articles)}/{len(articles)}")
    
    if not new_articles:
        print("  没有新记录")
        conn.close()
        return
    
    print("\n[3/3] 抓取详情页...")
    imported = 0
    skipped = 0
    
    for i, article in enumerate(new_articles, 1):
        title_short = article["title"][:45].replace("\n", " ")
        print(f"  [{i}/{len(new_articles)}] {title_short}...", end=" ", flush=True)
        
        content, pub_date = get_detail(article["url"])
        
        if not content:
            print("✗ 无正文")
            skipped += 1
            continue
        
        date = pub_date or article["date"]
        
        text_only = re.sub(r'<[^>]+>', '', content).strip()
        summary = text_only[:200]
        
        try:
            c.execute("""INSERT OR IGNORE INTO gov_raw 
                (title, content, page_url, summary, site_name, publish_date, category)
                VALUES (?, ?, ?, ?, ?, ?, ?)""",
                (article["title"], content, article["url"], summary,
                 "太和县", date, "环评公示"))
            if c.rowcount > 0:
                conn.commit()
                imported += 1
                print(f"✓ ({len(text_only)}字)")
            else:
                print("• 已存在")
        except Exception as e:
            print(f"✗ DB错误: {e}")
            conn.rollback()
    
    conn.close()
    print(f"\n{'=' * 60}")
    print(f"导入完成: {imported} 条 (跳过 {skipped} 条)")
    print(f"{'=' * 60}")

if __name__ == "__main__":
    main()