#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
吉安井冈山经开区-公告公示 爬虫
站点: jkq.jian.gov.cn
栏目: /xxgk-list-sotghejc.html (公告公示, 环评公示)
列表: POST /api-ajax_list-{page}.html (ajax_type[] 显式键名数组, 20条/页, total=251)
详情: div.xxgk_content 正文 (标题/日期从API, 附件 urljoin 绝对化)
"""
import re, os, sys, time, json
import requests, warnings
warnings.filterwarnings('ignore')
from datetime import datetime, timedelta

_MAX_PG = None
for _a in sys.argv[1:]:
    if _a.startswith("--pages="):
        try:
            _MAX_PG = int(_a.split("=", 1)[1])
        except Exception:
            pass
    elif _a.isdigit():
        _MAX_PG = int(_a)
if _MAX_PG is not None:
    print(f'[AutoPg] max_pages={_MAX_PG}')

SEARCH_DB = os.getenv("SEARCH_DB", "/root/search.db")
SITE_NAME = "吉安井冈山经开区-公告公示"
BASE_URL = "http://jkq.jian.gov.cn"
CUTOFF = (datetime.now() - timedelta(days=3*365)).strftime("%Y-%m-%d")
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": f"{BASE_URL}/xxgk-list-sotghejc.html",
    "Content-Type": "application/x-www-form-urlencoded",
    "X-Requested-With": "XMLHttpRequest",
}

API_LIST = f"{BASE_URL}/api-ajax_list-"
CATID = "142071"  # 公告公示栏目ID (show_lists_142071)


def fetch_list_page(page):
    data = [
        ('ajax_type[]', '70_xxgk'), ('ajax_type[]', CATID), ('ajax_type[]', '70'),
        ('ajax_type[]', 'xxgk'), ('ajax_type[]', 'Y-m-d'), ('ajax_type[]', '50'),
        ('ajax_type[]', '20'), ('ajax_type[]', 'is_top DESC'), ('ajax_type[]', 'displayorder DESC'),
        ('ajax_type[]', 'inputtime DESC'), ('ajax_type[]', ''), ('is_ds', '1'),
    ]
    try:
        r = requests.post(f"{API_LIST}{page}.html", headers=HEADERS, data=data, timeout=20, verify=False)
        if r.status_code != 200:
            print(f"  [WARN] {page} HTTP {r.status_code}")
            return []
        j = r.json()
        items = []
        for d in j.get("data", []):
            url = d.get("url", "")
            title = (d.get("title") or "").strip()
            date = (d.get("inputtime") or "").strip()
            desc = (d.get("description") or "").strip()
            if not url or not title:
                continue
            # ⚠️ 源站 url 字段损坏(仅域名根, 缺 xxgk-show-{id}.html 段) → 用 id 拼标准 URL
            if not re.search(r'xxgk-show-\d+\.html$', url):
                sid = (d.get("id") or "").strip()
                if sid:
                    url = f"{BASE_URL}/xxgk-show-{sid}.html"
            items.append({"url": url, "title": title, "date": date, "desc": desc})
        return items
    except Exception as e:
        print(f"  [WARN] {page} err: {e}")
        return []


def extract_detail(html, page_url):
    """div.xxgk_content 平衡匹配, 附件/图片 urljoin 绝对化, 保留 <p>"""
    i = html.find('class="xxgk_content"')
    if i < 0:
        return ""
    # 找该 div 起点 (可能带 style)
    ds = html.rfind('<div', 0, i)
    if ds < 0:
        return ""
    s = html[ds:]
    d = 0
    for j in range(len(s)):
        if s[j:j+4] == "<div" and (j+4 >= len(s) or s[j+4] in " >\n\r\t"):
            d += 1
        elif s[j:j+6] == "</div>":
            d -= 1
            if d == 0:
                gt = s.find(">", 0, j)
                c = s[gt+1:j] if gt > 0 else ""
                # 附件/图片绝对化
                for m in re.finditer(r'<a[^>]*href="([^"]+)"', c):
                    h = m.group(1)
                    if h.startswith("/"):
                        c = c.replace(h, BASE_URL + h)
                for m in re.finditer(r'<img[^>]*src="([^"]+)"', c):
                    h = m.group(1)
                    if h.startswith("/"):
                        c = c.replace(h, BASE_URL + h)
                # 纯 PDF 型: iframe src=*.pdf -> 附件段落 (保留 PDF)
                for m in re.finditer(r'<iframe[^>]*src="([^"]+\.pdf)"[^>]*>', c, re.I):
                    h = m.group(1)
                    if h.startswith("/"):
                        h = BASE_URL + h
                    c = c.replace(m.group(0), f'<p><a href="{h}">PDF原文</a></p>')
                c = re.sub(r'<(script|style)[^>]*>.*?</\1>', '', c, flags=re.DOTALL|re.I)
                c = re.sub(r'<p[^>]*>', '<p>', c)
                text = re.sub(r'<[^>]+>', '', c).strip()
                if text or '<a href=' in c or '<img' in c:
                    return c.strip()
    return ""


def run(max_pages=20):
    print(f"\n{'='*50}")
    print(f"🚀 {SITE_NAME}")
    print(f"{'='*50}")

    all_items = []
    for pg in range(1, min(max_pages, _MAX_PG or max_pages) + 1):
        its = fetch_list_page(pg)
        if not its:
            print(f"  第{pg}页: 空 -> 结束")
            break
        dates = [it["date"] for it in its if it["date"]]
        print(f"  第{pg}页: {len(its)} 条 ({min(dates) if dates else '?'} ~ {max(dates) if dates else '?'})")
        all_items.extend(its)
        if dates and max(dates) < CUTOFF:
            print(f"  该页已全部早于截断日{CUTOFF}，停止翻页")
            break
        time.sleep(0.3)

    print(f"\n📊 列表共 {len(all_items)} 条")
    kept = [it for it in all_items if not it["date"] or it["date"] >= CUTOFF]
    print(f"  窗口内(>= {CUTOFF}): {len(kept)} 条")

    out = []
    err = 0
    for i, it in enumerate(kept, 1):
        url = it["url"]
        if not url.startswith("http"):
            url = BASE_URL + url
        try:
            resp = requests.get(url, headers={"User-Agent": HEADERS["User-Agent"], "Referer": f"{BASE_URL}/xxgk-list-sotghejc.html"}, timeout=20, verify=False)
            resp.encoding = "utf-8"   # ⚠️ 必须在 .text 前设置, 否则 requests 猜 Latin-1 → 乱码
            detail_html = resp.text
        except Exception as e:
            print(f"  ⚠ 详情失败: {it['title'][:40]} {e}")
            err += 1
            continue
        content = extract_detail(detail_html, url)
        text_len = len(re.sub(r'<[^>]+>', '', content).strip()) if content else 0
        if text_len < 5 and '<img' not in (content or ''):
            # ⚠️ 详情提取失败 → 回退 API description 摘要作正文 (源站 url 字段损坏条目的详情页是别的文章)
            if it.get("desc"):
                _d = re.sub(r'\s+', ' ', it['desc'])
                content = f"<p>{_d}</p>"
                print(f"  ↳ 用API摘要回退: {it['title'][:40]}")
            else:
                print(f"  ⚠ 空正文: {it['title'][:40]}")
                err += 1
                continue
        out.append({
            "site_name": SITE_NAME,
            "title": it["title"],
            "pub_date": it["date"],
            "content": content,
            "source_url": url,
            "url": url,
        })
        if i % 20 == 0:
            print(f"  ...{i}/{len(kept)}")
        time.sleep(0.2)

    print(f"\n📦 待入库: {len(out)}, 失败: {err}")
    sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
    from crawler_lib import push_to_searchdb
    push_to_searchdb(out, batch_label=SITE_NAME)


if __name__ == "__main__":
    run(max_pages=20)