#!/usr/bin/env python3
"""crawl_ldnews_ggb.py - 娄底新闻网-公告吧 (湖南娄底)

站点: https://news.ldnews.cn/loudi/thedosee/
CMS: cmstop (JS 瀑布流分页)
列表: <li><div class="news-item">...<a href="http://news.ldnews.cn/thedosee/YYYYMM/ID.shtml" title="标题">标题</a>...
分页: ⚠️ 瀑布流 JS: var page=1; pagetotal=50; url+page+'.shtml' → 2.shtml~50.shtml 每页20条
      每页数据在 <span id="start-data-N"></span>...<span id="stop-data-N"> 之间
详情: ⚠️ http 301 → https (须 https 或跟随跳转)
      标题: <title> 分割 '_娄底新闻网'; 日期: meta 或文本 2026-08-04 11:42
      正文: div.right_body 平衡 div (纯正文, 尾部混入登录区需清洗)
"""
import sys, os, requests, re, argparse, json
from datetime import datetime, timedelta
from urllib.parse import urljoin

sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from crawler_lib import push_to_searchdb

CUTOFF = (datetime.now() - timedelta(days=365 * 3)).strftime('%Y-%m-%d')
SITE_NAME = '娄底新闻网-公告吧'
GROUP_NAME = '湖南'
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',
    'Referer': 'https://news.ldnews.cn/',
}
BASE = 'https://news.ldnews.cn'
LIST_URL = BASE + '/loudi/thedosee/'


def fetch(url):
    try:
        r = requests.get(url, timeout=(5, 20), headers=HEADERS, verify=False)
        r.encoding = 'utf-8'
        return r.text
    except Exception as e:
        print(f"[WARN] fetch failed: {url} - {e}", file=sys.stderr)
        return ''


def extract_items(html, page_no):
    """列表条目: thedosee/YYYYMM/ID.shtml + title 属性/文本"""
    items = []
    # 首页全页; 分页取 start-data-N 段
    seg = html
    if page_no > 1:
        s = html.find(f'<span id="start-data-{page_no}"')
        e = html.find(f'<span id="stop-data-{page_no}"')
        if s < 0 or e < 0 or e <= s:
            return []
        seg = html[s:e]
    for m in re.finditer(r'<a[^>]+href=["\']?([^"\'>\s]*thedosee/\d{6}/\d+\.shtml)[^>]*?(?:title=["\']([^"\']*)["\'])?[^>]*>(.*?)</a>', seg, re.DOTALL):
        href, title_attr, title_html = m.group(1), m.group(2) or '', m.group(3)
        title = (title_attr or re.sub(r'<[^>]+>', '', title_html)).strip()
        title = title.replace('\u200b', '').strip()
        title = re.sub(r'^\s*(?:&middot;|&nbsp;|·|\s)+', '', title)
        if len(title) < 4:
            continue
        url = href if href.startswith('http') else urljoin(BASE, href)
        url = url.replace('http://', 'https://')
        items.append({'url': url, 'title': title})
    return items


def extract_balanced_div(html, open_tag_re):
    """平衡 div 匹配 (大小写不敏感)"""
    m = re.search(open_tag_re, html, re.I)
    if not m:
        return ''
    gt = html.find('>', m.start())
    if gt == -1:
        return ''
    i = gt + 1
    depth = 1
    for mm in re.finditer(r'<div[\s>]|</div>', html[i:], re.I):
        if mm.group(0).startswith('<div') or mm.group(0).startswith('<DIV'):
            depth += 1
        else:
            depth -= 1
            if depth == 0:
                return html[i:i + mm.start()]
    return html[i:]


def fetch_detail(url):
    html = fetch(url)
    if not html:
        return '', '', ''
    # 标题
    title = ''
    m = re.search(r'<title>\s*([^<]+?)\s*</title>', html, re.DOTALL)
    if m:
        title = re.sub(r'\s+', '', m.group(1)).strip()
        title = re.split(r'_娄底新闻网|_|[-–—]', title)[0].strip()
    title = title.replace('\u200b', '').strip()
    title = re.sub(r'^\s*(?:&middot;|&nbsp;|·|\s)+', '', title)
    # 日期: 文本 2026-08-04 11:42 (right_body 前)
    date_text = ''
    m = re.search(r'(\d{4}-\d{2}-\d{2})\s+\d{2}:\d{2}', html)
    if m:
        date_text = m.group(1)
    if not date_text:
        m = re.search(r'<meta[^>]*name="[Pp]ub[Dd]ate"[^>]*content="([^"]*)"', html)
        if m:
            md = re.search(r'(\d{4}-\d{2}-\d{2})', m.group(1))
            if md:
                date_text = md.group(1)
    # 正文: div.right_body 平衡 div
    content = ''
    c = extract_balanced_div(html, r'<div[^>]*class="[^"]*right_body[^"]*"[^>]*>')
    if c.strip():
        content = c.strip()
    # 清洗: 移除尾部登录区/评论相关
    for kw in ['用户登录', '用户名', '密码', '下次自动登录', '没有账号', '马上注册',
               '使用第三方账号登录', 'QQ账户', '新浪微博', '腾讯微博', '验证码']:
        idx = content.find(kw)
        if idx > 0 and idx > len(content) * 0.5:
            content = content[:idx]
            break
    content = re.sub(r'<script.*?</script>', '', content, flags=re.DOTALL | re.I)
    content = re.sub(r'<style.*?</style>', '', content, flags=re.DOTALL | re.I)
    # 附件/图片相对路径 → 绝对化
    content = re.sub(r'href="(?!https?://|//|#|javascript:)([^"]+)"',
                     lambda mm: f'href="{urljoin(url, mm.group(1))}"', content)
    content = re.sub(r'src="(?!https?://|//)([^"]+)"',
                     lambda mm: f'src="{urljoin(url, mm.group(1))}"', content)
    # 清首尾空 p
    content = re.sub(r'^(<p[^>]*>(\s*(<br\s*/?>)?\s*)?</p>)+', '', content)
    content = re.sub(r'(<p[^>]*>(\s*(<br\s*/?>)?\s*)?</p>)+$', '', content)
    return title, date_text, content.strip()


def run(max_pages=5):
    records = []
    seen = set()
    for pg in range(max_pages):
        url = LIST_URL if pg == 0 else BASE + f'/loudi/thedosee/{pg + 1}.shtml'
        html = fetch(url)
        if not html:
            print(f"  Page {pg + 1}: fetch fail (stop)")
            break
        items = extract_items(html, pg + 1)
        if not items:
            print(f"  Page {pg + 1}: 0 items (stop)")
            break
        print(f"  Page {pg + 1}: {len(items)} items")
        for it in items:
            if it['url'] in seen:
                continue
            seen.add(it['url'])
            title, date_text, content = fetch_detail(it['url'])
            if not title:
                title = it['title']
            if not content:
                print(f"    SKIP empty content: {it['url']}")
                continue
            pd = date_text
            if pd and pd < CUTOFF:
                continue
            # 附件提取
            atts = []
            for am in re.finditer(r'<a[^>]*href="([^"]+)"[^>]*>(.*?)</a>', content, re.DOTALL):
                h, t = am.group(1), re.sub(r'<[^>]+>', '', am.group(2)).strip()
                if re.search(r'\.(pdf|doc|docx|xls|xlsx|zip|rar|caj|wps)$', h, re.I) or 'download' in h.lower():
                    atts.append({'title': t or '附件', 'url': h})
            records.append({
                'title': title, 'url': it['url'], 'source_url': it['url'],
                'pub_date': pd, 'site_name': SITE_NAME, 'content': content,
                'summary': '', 'attachments': json.dumps(atts, ensure_ascii=False) if atts else '[]',
                'group_name': GROUP_NAME,
            })
    print(f"  Total: {len(records)} items")
    valid = [r for r in records if r['content'].strip()]
    if valid:
        push_to_searchdb(valid, "ldnews_ggb")
    return len(valid)


if __name__ == '__main__':
    ap = argparse.ArgumentParser()
    ap.add_argument('--pages', type=int, default=5)
    args = ap.parse_args()
    n = run(args.pages)
    print(f"[RESULT] {SITE_NAME}: {n} items")