#!/usr/bin/env python3
"""
灵石县 - 晋中市生态环境局灵石分局 通知公告 (lingshi.gov.cn)
列表: /zwgk/bmxxgkml/xzfgzbm/20hbj/fdzdgknrhjj/zcwj20hbj/tzgg20hbj
      /zwgk/bmxxgkml/xzfgzbm/20hbj/fdzdgknrhjj/zcwj20hbj/tzgg20hbj_N (N=2..42)
详情: /zwgk/.../tzgg20hbj/content_XXXXX
正文: div.conTxt
注意: DNS不解析，需 IP + Host header
"""
import re, sys, os, json, time
import urllib.request, urllib.error
from bs4 import BeautifulSoup
import sqlite3

BASE_URL = "http://218.26.234.211"
SITE_NAME = "晋中市生态环境局灵石分局-通知公告"
GROUP_NAME = "山西环评平台"
DB = os.environ.get("SEARCH_DB", "/root/search.db")
HEADERS = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
    "Host": "www.lingshi.gov.cn",
}
LIST_PATH = "/zwgk/bmxxgkml/xzfgzbm/20hbj/fdzdgknrhjj/zcwj20hbj/tzgg20hbj"

_MAX_PG = None
for i, a in enumerate(sys.argv):
    if a == "--pages" and i + 1 < len(sys.argv):
        _MAX_PG = int(sys.argv[i + 1])
        break


def html_table_to_html(table, base_url=""):
    """保留 HTML 表格结构，仅将相对链接/图片转绝对 URL"""
    import urllib.parse
    tbl = BeautifulSoup(str(table), 'html.parser')
    for a in tbl.find_all('a'):
        href = a.get('href', '')
        if href and not href.startswith(('http', 'javascript', '#')):
            a['href'] = urllib.parse.urljoin(base_url, href) if base_url else href
    for img in tbl.find_all('img'):
        src = img.get('src', '')
        if src and not src.startswith(('http', '//', 'data:')):
            img['src'] = urllib.parse.urljoin(base_url, src) if base_url else src
    return str(tbl)
def fetch(url):
    req = urllib.request.Request(url, headers=HEADERS)
    try:
        return urllib.request.urlopen(req, timeout=30).read().decode("utf-8", errors="replace")
    except Exception as e:
        print(f"[WARN] fetch failed: {url} - {e}", file=sys.stderr)
        return ""


def extract_items(html):
    """Extract items from list page"""
    items = []
    soup = BeautifulSoup(html, "html.parser")
    for li in soup.select("ul.infoList > li"):
        a = li.find("a", href=True) if li else None
        if not a:
            continue
        href = a.get("href", "")
        if "content_" not in href:
            continue

        # Title from title attribute (preferred) or text
        title_attr = a.get("title", "")
        if title_attr.startswith("标题："):
            title = title_attr[3:].strip()
        else:
            title = a.get_text(strip=True)

        if not title or len(title) < 5:
            continue

        # Make absolute URL (store the original domain URL for DB)
        if not href.startswith("http"):
            if href.startswith("/"):
                orig_url = "http://www.lingshi.gov.cn" + href
            else:
                orig_url = "http://www.lingshi.gov.cn/" + href
        else:
            orig_url = href
        # Replace domain with IP for fetching
        fetch_url = orig_url.replace("http://www.lingshi.gov.cn", BASE_URL)

        # Date from span.date
        date = ""
        span = li.find("span", class_="date")
        if span:
            date = span.get_text(strip=True)

        items.append({"title": title, "url": fetch_url, "orig_url": orig_url, "date": date})
    return items


def parse_detail(html, url):
    """Parse detail page"""
    soup = BeautifulSoup(html, "html.parser")

    title = ""
    meta = soup.find("meta", attrs={"name": "ArticleTitle"})
    if meta and meta.get("content"):
        title = meta["content"].strip()

    date_text = ""
    meta = soup.find("meta", attrs={"name": "PubDate"})
    if meta and meta.get("content"):
        m = re.search(r"(\d{4}-\d{2}-\d{2})", meta["content"])
        if m:
            date_text = m.group(1)

    content = ""
    attachments = []
    content_div = soup.select_one("div.conTxt")
    if not content_div:
        return {"title": title, "date": date_text, "content": "", "attachments": []}

    parts = []
    for el in content_div.find_all(["p", "table", "img"], recursive=True):
        if el.name == "p" and el.find_parent("table"):
            continue
        if el.name == "table" and el.find_parent("table"):
            continue

        if el.name == "img":
            src = el.get("src", "")
            if "icon_" in src.lower() or "/sysimage/" in src.lower():
                continue
            alt = el.get("alt", "") or ""
            if src.startswith("//"):
                src = "http:" + src
            elif src.startswith("/"):
                src = "http://www.lingshi.gov.cn" + src
            elif not src.startswith("http"):
                src = url.rsplit("/", 1)[0] + "/" + src
            parts.append(f"![{alt}]({src})")
            continue

        if el.name == 'table':
            tbl_html = html_table_to_html(el, url)
            if tbl_html:
                attachments.append(tbl_html)
        text = el.get_text(" ", strip=True)
        if text:
            for a in el.find_all("a", href=True):
                ahref = a["href"]
                if re.search(r'\.(docx?|pdf|xlsx?|rar|zip)$', ahref, re.I):
                    atitle = a.get_text(strip=True) or a.get("download", "")
                    if ahref.startswith("/"):
                        ahref = "http://www.lingshi.gov.cn" + ahref
                    elif not ahref.startswith("http"):
                        ahref = url.rsplit("/", 1)[0] + "/" + ahref
                    attachments.append({"title": atitle or "附件", "url": ahref})
            parts.append(text)

    content = "\n\n".join(p for p in parts if p.strip())
    content = content.replace("\u00a0", " ").replace("&nbsp;", " ")

    seen = set()
    unique_att = []
    for att in attachments:
        if att["url"] not in seen:
            seen.add(att["url"])
            unique_att.append(att)

    return {"title": title, "date": date_text, "content": content, "attachments": unique_att}


def main():
    conn = sqlite3.connect(DB, timeout=60)
    c = conn.cursor()

    # Clean old data
    c.execute("DELETE FROM gov_search WHERE site_name=?", (SITE_NAME,))
    c.execute("DELETE FROM gov_raw WHERE site_name=?", (SITE_NAME,))
    conn.commit()
    print("[清理] 已清理该站旧数据", flush=True)

    all_items = []

    # Page 1
    html = fetch(f"{BASE_URL}{LIST_PATH}")
    if not html:
        print("[ERROR] Cannot fetch list page", file=sys.stderr)
        sys.exit(1)
    items = extract_items(html)
    print(f"[INFO] Page 1: {len(items)} items", flush=True)
    all_items.extend(items)

    # Determine total pages
    m = re.search(r'共(\d+)条信息/共(\d+)页', html)
    total_pages = int(m.group(2)) if m else 42
    total_records = int(m.group(1)) if m else 835
    print(f"[INFO] Total: {total_records}条 / {total_pages}页", flush=True)

    if _MAX_PG and total_pages > _MAX_PG:
        total_pages = _MAX_PG
    print(f"[INFO] 实际抓取页数: {total_pages}", flush=True)

    # Pages 2+: tzgg20hbj_N
    for pg in range(2, total_pages + 1):
        url = f"{BASE_URL}{LIST_PATH}_{pg}"
        html = fetch(url)
        if not html:
            break
        items = extract_items(html)
        if not items:
            break
        print(f"[INFO] Page {pg}: {len(items)} items", flush=True)
        all_items.extend(items)

    print(f"[INFO] Total items from list: {len(all_items)}", flush=True)

    new_count = 0
    skip_count = 0
    empty_count = 0
    total_att = 0

    for idx, item in enumerate(all_items):
        c.execute("SELECT id FROM gov_raw WHERE page_url=?", (item["orig_url"],))
        if c.fetchone():
            skip_count += 1
            continue

        if idx % 5 == 0:
            print(f"[INFO] [{idx}/{len(all_items)}] {item['title'][:40]}...", flush=True, end=" ", file=sys.stderr)

        html = fetch(item["url"])
        if not html:
            print("FETCH FAILED", file=sys.stderr)
            continue

        detail = parse_detail(html, item["url"])
        title = detail["title"] or item["title"]
        date_text = detail["date"] or item["date"]
        content = detail["content"]
        attachments_list = detail["attachments"]

        if not content.strip():
            empty_count += 1
            if idx % 5 == 0:
                print("EMPTY", file=sys.stderr)
            continue

        plain = re.sub(r'\s+', ' ', content).strip()
        summary = plain[:200] if plain else title[:200]
        att_json = json.dumps(attachments_list, ensure_ascii=False) if attachments_list else "[]"

        date_rank = 0
        if date_text:
            m2 = re.search(r"(\d{4})-(\d{2})-(\d{2})", date_text)
            if m2:
                date_rank = int(m2.group(1) + m2.group(2) + m2.group(3))

        try:
            c.execute(
                """INSERT OR IGNORE INTO gov_raw 
                (title, site_name, group_name, page_url, publish_date, content, summary, attachments, date_rank)
                VALUES (?,?,?,?,?,?,?,?,?)""",
                (title, SITE_NAME, GROUP_NAME, item["orig_url"], date_text, content, summary, att_json, date_rank)
            )
            if c.rowcount > 0:
                new_count += 1
                total_att += len(attachments_list)
        except Exception as e:
            if idx % 5 == 0:
                print(f"DB ERROR: {e}", file=sys.stderr)

        time.sleep(0.3)

    conn.commit()
    conn.close()

    print(f"\n[RESULT] {SITE_NAME}: {new_count} new, {skip_count} skip, {empty_count} empty, {total_att} attachments", flush=True)


if __name__ == "__main__":
    main()