#!/usr/bin/env python3
"""茂名高新区 - 公示公告 爬虫
SiteServer/NFCMS，20条/页，16页共315条
列表页 kkpager 分页，详情页 <div class="content" id="content"> 正文
"""

import os
import sys
import re
import requests
from datetime import datetime, timedelta

DB_PATH = os.getenv("SEARCH_DB", "/root/search.db")
SITE_NAME = "mmht.gov.cn-公示公告"
BASE_URL = "http://www.mmht.gov.cn"
LIST_PATH = "/xwzx/gsgg"
TOTAL_PAGES = 16

HEADERS = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
}

CUTOFF_DATE = (datetime.now() - timedelta(days=3*365)).strftime("%Y-%m-%d")
print(f"[info] 日期过滤: >= {CUTOFF_DATE}")


def fetch_list_page(page):
    """获取列表页HTML"""
    if page == 1:
        url = f"{BASE_URL}{LIST_PATH}/index.html"
    else:
        url = f"{BASE_URL}{LIST_PATH}/index_{page}.html"
    try:
        r = requests.get(url, headers=HEADERS, timeout=15)
        r.encoding = "utf-8"
        return r.text
    except Exception as e:
        print(f"[error] 第{page}页请求失败: {e}")
        return ""


def parse_list_items(html):
    """解析列表页条目"""
    items = []
    # <li>...<a href="URL">TITLE</a><span class="date">DATE</span></li>
    pattern = r'<li>\s*<span class="point"></span>\s*<a href="([^"]+)">([^<]+)</a>\s*<span class="date">([^<]+)</span>\s*</li>'
    for m in re.finditer(pattern, html):
        url = m.group(1).strip()
        if not url.startswith("http"):
            url = BASE_URL + url
        items.append({
            "title": m.group(2).strip(),
            "date": m.group(3).strip(),
            "url": url,
        })
    return items


def fetch_detail_content(html):
    """从详情页提取 #content 正文"""
    m = re.search(r'<div class="content"[^>]*id="content">([\s\S]*?)</div>\s*</div>\s*<div class="friend-link">', html)
    if m:
        return m.group(1).strip()
    # 备用: 取 article-content 完整内容
    m = re.search(r'<div class="article-content">([\s\S]*?)</div>\s*<div class="friend-link">', html)
    if m:
        content = m.group(1)
        content = re.sub(r'<p class="title">.*?</p>', '', content)
        content = re.sub(r'<div class="desc">.*?</div>', '', content)
        return content.strip()
    return ""


def fetch_detail_page(url):
    """获取详情页HTML"""
    try:
        r = requests.get(url, headers=HEADERS, timeout=15)
        r.encoding = "utf-8"
        return r.text
    except Exception as e:
        print(f"  [warn] 详情页请求失败: {url[-60:]}, {e}")
        return ""


def main():
    import sqlite3

    conn = sqlite3.connect(DB_PATH)
    c = conn.cursor()
    c.execute("PRAGMA journal_mode=WAL")

    new_count = 0
    dup_count = 0
    old_count = 0
    err_count = 0

    for page in range(1, TOTAL_PAGES + 1):
        html = fetch_list_page(page)
        if not html:
            continue

        items = parse_list_items(html)
        if not items:
            print(f"[warn] 第{page}页未解析到条目")
            continue

        page_new = 0
        for item in items:
            date = item["date"]
            if date < CUTOFF_DATE:
                old_count += 1
                continue

            title = item["title"]
            url = item["url"]

            c.execute("SELECT COUNT(*) FROM gov_raw WHERE page_url=?", (url,))
            if c.fetchone()[0] > 0:
                dup_count += 1
                continue

            detail_html = fetch_detail_page(url)
            if not detail_html:
                err_count += 1
                continue

            content = fetch_detail_content(detail_html)
            if not content:
                content = f"[无正文] {title}"

            try:
                c.execute("""
                    INSERT OR IGNORE INTO gov_raw (page_url, title, content, site_name, publish_date, summary, category)
                    VALUES (?, ?, ?, ?, ?, ?, ?)
                """, (url, title, content, SITE_NAME, date, title, "公示公告"))
                if c.rowcount > 0:
                    page_new += 1
            except Exception as e:
                print(f"  [error] 入库失败: {e}")
                err_count += 1

        conn.commit()
        new_count += page_new
        print(f"[page {page}/{TOTAL_PAGES}] +{page_new} new, cumulative: {new_count} new")

    conn.close()
    print(f"\n[完成] 茂名高新区: 新增 {new_count} 条, 重复 {dup_count}, 超3年跳过 {old_count}, 错误 {err_count}")


if __name__ == "__main__":
    main()