#!/usr/bin/env python3
import os
"""
涪陵新城-政策文件 爬虫
站点: www.fl.gov.cn
栏目: 政策文件
"""
import requests, re, sqlite3, time, sys, json, warnings
warnings.filterwarnings('ignore')
from datetime import datetime, timedelta

SEARCH_DB = os.getenv("SEARCH_DB", "/root/search.db")
SITE_NAME = "涪陵新城-政策文件"
DOMAIN = "www.fl.gov.cn"
BASE_PATH = "/jz/xcq/zwgk_46684/fdzdgknr_46686/lzyj_46687/zcwj"

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"}


def fetch_list_page(page):
    """获取列表页"""
    if page == 1:
        url = f"http://{DOMAIN}{BASE_PATH}/index.html"
    else:
        url = f"http://{DOMAIN}{BASE_PATH}/index_{page-1}.html"
    
    for retry in range(3):
        try:
            r = requests.get(url, headers=HEADERS, timeout=30)
            if r.status_code == 200:
                r.encoding = "utf-8"
                html = r.text
                
                # Find the list container
                items = []
                ib = html.find('infos-box')
                if ib < 0:
                    return [], 0
                ul = html.find('<ul', ib)
                if ul < 0:
                    return [], 0
                ul_end = html.find('</ul>', ul)
                ul_content = html[ul:ul_end]
                
                for li in re.findall(r'<li[^>]*>.*?</li>', ul_content, re.DOTALL):
                    m = re.search(r'href="([^"]+)"[^>]*>(.*?)</a>', li)
                    if not m:
                        continue
                    dm = re.search(r'<span[^>]*>([^<]*)</span>', li)
                    date = dm.group(1).strip() if dm else ""
                    url_str = m.group(1)
                    if not url_str.startswith("http"):
                        if url_str.startswith("./"):
                            url_str = f"http://{DOMAIN}{BASE_PATH}/{url_str[2:]}"
                        elif url_str.startswith("//"):
                            url_str = f"http:{url_str}"
                        else:
                            url_str = f"http://{DOMAIN}{url_str}"
                    title = re.sub(r'<[^>]+>', '', m.group(2)).strip()
                    items.append({"url": url_str, "title": title, "date": date})
                
                # Total pages from JS
                pm = re.search(r'createPage\((\d+)', html)
                total_pages = int(pm.group(1)) if pm else 3
                return items, total_pages
        except Exception as e:
            if retry < 2:
                time.sleep(2)
    return [], 0


def extract_content(html):
    """从 trs_editor_view 容器提取"""
    for pat in ['trs_editor_view', 'TRS_UEDITOR', 'TRS_Paper', 'trs_paper']:
        idx = html.find(pat)
        if idx >= 0:
            ds = html.rfind("<div", 0, idx)
            if ds < 0:
                continue
            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 s[7:j]
                        c = re.sub(r'<(script|style)[^>]*>.*?</\1>', '', c, flags=re.DOTALL|re.I)
                        text = re.sub(r'<[^>]+>', '', c).strip()
                        if len(text) > 50:
                            return c.strip()
    return ""


def extract_meta(html):
    title = ""
    tm = re.search(r'<meta[^>]*name="ArticleTitle"[^>]*content="([^"]*)"', html)
    if tm:
        title = tm.group(1)
    if not title:
        ttm = re.search(r'<title>(.*?)<', html)
        if ttm:
            title = ttm.group(1).strip()
    
    date = ""
    pm = re.search(r'<meta[^>]*name="PubDate"[^>]*content="([^"]*)"', html)
    if pm:
        d = pm.group(1).strip()
        cm = re.match(r'(\d{4})-(\d{1,2})-(\d{1,2})', d)
        if cm:
            date = f"{cm.group(1)}-{cm.group(2).zfill(2)}-{cm.group(3).zfill(2)}"
    if not date:
        dm = re.search(r'(\d{4}-\d{2}-\d{2})', html)
        if dm:
            date = dm.group(1)
    return title, date


def insert_article(url, title, date, content):
    db = sqlite3.connect(SEARCH_DB, timeout=60)
    db.execute("PRAGMA journal_mode=WAL")
    try:
        db.execute(
            "INSERT OR IGNORE INTO gov_raw(site_name,source_url,page_url,title,publish_date,content,summary,status,category) VALUES(?,?,?,?,?,?,?,?,?)",
            (SITE_NAME, url, url, title, date, content, title[:200], "active", "政策文件"))
        affected = db.total_changes
        db.commit()
        if affected:
            db.execute(
                "INSERT INTO gov_search(rowid,title,site_name,summary) SELECT r.id,r.title,r.site_name,r.summary FROM gov_raw r WHERE r.page_url=? AND r.id NOT IN (SELECT rowid FROM gov_search)",
                (url,))
            db.commit()
        db.close()
        return affected
    except Exception as e:
        db.close()
        print(f"  DB error: {e}")
        return 0


def run(max_pages=10):
    print(f"\n{'='*50}")
    print(f"🚀 {SITE_NAME}")
    print(f"{'='*50}")

    items, total_pages = fetch_list_page(1)
    if not items:
        print("❌ 列表页无数据")
        return
    
    total_pages = min(max_pages, total_pages)
    print(f"  共{total_pages}页")
    
    all_items = list(items)
    for pg in range(2, total_pages + 1):
        its, _ = fetch_list_page(pg)
        if not its:
            break
        all_items.extend(its)
        dates = [it["date"] for it in its if it["date"]]
        if dates:
            print(f"  第{pg}页: {len(its)} 条 ({min(dates)} ~ {max(dates)})")
        else:
            print(f"  第{pg}页: {len(its)} 条")
        time.sleep(0.2)

    print(f"\n📊 共 {len(all_items)} 条")
    new = skip = no_body = err = 0
    for i, item in enumerate(all_items, 1):
        if item["date"] and item["date"] < CUTOFF:
            skip += 1
            continue

        for retry in range(3):
            try:
                r = requests.get(item["url"], headers=HEADERS, timeout=30)
                if r.status_code == 200:
                    r.encoding = "utf-8"
                    html = r.text
                    break
            except:
                if retry < 2:
                    time.sleep(2)
                html = None

        if not html:
            err += 1
            continue

        pt, pd = extract_meta(html)
        real_title = pt if pt else item["title"]
        real_date = pd if pd else item["date"]

        if real_date and real_date < CUTOFF:
            skip += 1
            continue

        content = extract_content(html)
        text_len = len(re.sub(r'<[^>]+>', '', content).strip()) if content else 0
        if text_len < 10:
            no_body += 1
            continue

        if insert_article(item["url"], real_title, real_date, content):
            new += 1
        else:
            skip += 1

        if i % 5 == 0:
            print(f"  ...{i}/{len(all_items)}")
        time.sleep(0.2)

    # FTS sync
    print(f"\n📊 FTS同步...")
    db = sqlite3.connect(SEARCH_DB, timeout=60)
    db.execute("DELETE FROM gov_search WHERE site_name=?", (SITE_NAME,))
    db.execute("INSERT INTO gov_search(rowid,title,site_name,summary) SELECT rowid,title,site_name,summary FROM gov_raw WHERE site_name=?", (SITE_NAME,))
    db.commit()
    c = db.execute("SELECT COUNT(*) FROM gov_search WHERE site_name=?", (SITE_NAME,))
    cnt = c.fetchone()[0]
    db.close()
    print(f"  FTS: {cnt} 条")
    print(f"\n✅ {SITE_NAME}: 新增{new}, 跳过{skip}, 空正文{no_body}, 错误{err}, FTS共{cnt}条")


if __name__ == "__main__":
    mp = int(sys.argv[1]) if len(sys.argv) > 1 else 10
    run(mp)