#!/usr/bin/env python3
"""
crawl_tx.py — 桐乡市人民政府·建设项目环境影响评价信息公示
============================================================
CMS: JCMS (大汉版通)，信息公开 API 驱动，分类过滤 code=I01-A01

列表：API /api-gateway/jpaas-publish-server/front/page/build/unit
      带 paramJson 搜索参数，15条/页，共 3页 / 37条（已知）
详情：<div class="column-zjnr"> 正文 | <meta name="ArticleTitle"> /
      <meta name="PubDate"> / <meta name="ContentSource">

用法:
    python3 crawl_tx.py               # 全量
    python3 crawl_tx.py 1             # 增量（仅最新页）
"""

import os
import re
import sys
import json
import requests
from bs4 import BeautifulSoup
from urllib.parse import urljoin
from datetime import datetime, timedelta
from concurrent.futures import ThreadPoolExecutor, as_completed

DB_PATH = os.environ.get("SEARCH_DB", os.environ.get("GOV_DB_PATH", "/root/search.db"))

BASE_URL = "https://www.tx.gov.cn"
SITE_NAME = "桐乡市人民政府"
COLUMN_NAME = "建设项目环境影响评价信息公示"
COLUMN_CODE = "I01-A01"
GROUP_NAME = "重点领域信息"

API_URL = f"{BASE_URL}/api-gateway/jpaas-publish-server/front/page/build/unit"
API_BASE_PARAMS = {
    "parseType": "bulidstatic",
    "webId": "3159",
    "tplSetId": "PumzAFWTgBHhWx3260d5t",
    "pageType": "column",
    "tagId": "组配分类list",
    "pageId": "1229856913",
}

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",
    "Accept": "application/json",
    "Referer": f"https://www.tx.gov.cn/col/col1229856913/index.html?number={COLUMN_CODE}",
}

MAX_WORKERS = 20
PAGE_SIZE = 15


def log(msg):
    print(f"[tx] {msg}")


def fetch_articles_page(page_no):
    """Fetch one page of article list from the API. Returns list of (url, title, date_str)."""
    search_json = json.dumps({
        "xxgkId": COLUMN_CODE,
        "xxgkType": "",
        "className": GROUP_NAME,
    }, ensure_ascii=False)
    
    param_json = json.dumps({
        "pageNo": page_no,
        "pageSize": PAGE_SIZE,
        "search": search_json,
    }, ensure_ascii=False)

    params = dict(API_BASE_PARAMS)
    params["paramJson"] = param_json

    try:
        resp = requests.get(API_URL, headers=HEADERS, params=params, timeout=30)
        data = resp.json()
        html = data.get("data", {}).get("html", "")
    except Exception as e:
        log(f"  ERROR API page {page_no}: {e}")
        return [], 0

    items = []
    total = 0

    # Find total count
    m = re.search(r'count="(\d+)"', html)
    if m:
        total = int(m.group(1))

    # Extract article links
    for m in re.finditer(
        r'<a\s+class="fl"\s+href="([^"]+)"\s+target="_blank"\s+title="([^"]+)"[^>]*>',
        html
    ):
        href = m.group(1)
        title = m.group(2).strip()

        # Find date
        full_entry = html[m.start():m.end() + 500]
        date_m = re.search(r'<span class="fr">(\d{4}-\d{2}-\d{2})</span>', full_entry)
        date_str = date_m.group(1) if date_m else ""

        full_url = href if href.startswith("http") else urljoin(BASE_URL, href)
        items.append((full_url, title, date_str))

    return items, total


def fetch_detail(url):
    """Return (content_html, date_str, title) or (None, None, None)."""
    try:
        resp = requests.get(url, headers=HEADERS, timeout=30)
        resp.encoding = "utf-8"
        html = resp.text
    except Exception as e:
        log(f"  ERROR {url}: {e}")
        return None, None, None

    soup = BeautifulSoup(html, "html.parser")

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

    # Date from meta
    date_str = ""
    meta_date = soup.find("meta", attrs={"name": "PubDate"})
    if meta_date and meta_date.get("content"):
        m = re.match(r"(\d{4}-\d{2}-\d{2})", meta_date["content"].strip())
        if m:
            date_str = m.group(1)

    # Content: <div class="column-zjnr">
    content_div = soup.find("div", class_=lambda c: c and "column-zjnr" in (c if isinstance(c, str) else " ".join(c)))
    content_html = str(content_div) if content_div else ""

    # Fallback: bt_content
    if not content_html:
        content_div = soup.find("div", class_=lambda c: c and "bt_content" in (c if isinstance(c, str) else " ".join(c)))
        content_html = str(content_div) if content_div else ""

    if not content_html:
        log(f"  WARNING: no content for {url}")
        return None, date_str, title

    return content_html, date_str, title


def fetch_detail_batch(urls):
    results = {}
    with ThreadPoolExecutor(max_workers=MAX_WORKERS) as ex:
        fut_map = {ex.submit(fetch_detail, url): url for url in urls}
        for fut in as_completed(fut_map):
            url = fut_map[fut]
            try:
                results[url] = fut.result()
            except Exception as e:
                results[url] = (None, None, None)
    return results


def save_to_db(items):
    """items: list of (url, title, date_str, content_html)"""
    import sqlite3
    if not items:
        return 0
    conn = sqlite3.connect(DB_PATH, timeout=60)
    c = conn.cursor()
    inserted = 0
    for url, title, date_str, content_html in items:
        try:
            c.execute(
                """INSERT OR REPLACE INTO gov_raw (site_name, page_url, title, publish_date, content, category, script_name) VALUES (?, ?, ?, ?, ?, ?, 'crawl_tx.py')""",
                (SITE_NAME, url, title, date_str, content_html, COLUMN_NAME),
            )
            inserted += 1
        except Exception as e:
            log(f"  DB error for {url}: {e}")
    conn.commit()
    conn.close()
    return inserted


def main():
    incremental = len(sys.argv) > 1 and sys.argv[1] == "1"
    three_years_ago = (datetime.now() - timedelta(days=365 * 3)).strftime("%Y-%m-%d")
    log(f"3-year cutoff: {three_years_ago}")

    # Phase 1: Fetch all list pages
    all_articles = []
    total_expected = 0

    page = 1
    while True:
        items, total = fetch_articles_page(page)
        if page == 1:
            total_expected = total
            log(f"Total expected: {total_expected}")

        if not items:
            if page == 1:
                log("No articles found")
                return
            break

        # Filter by date (3 years)
        filtered = [(u, t, d) for u, t, d in items if d and d >= three_years_ago]
        all_articles.extend(filtered)

        if incremental:
            log(f"Incremental: page 1, {len(filtered)} articles")
            break

        if filtered and filtered != items and len(items) == len(filtered):
            pass  # All items in range, continue
        elif filtered != items:
            log(f"Page {page}: some items before cutoff, stopping")
            break

        if page * PAGE_SIZE >= total_expected:
            break

        page += 1

    log(f"Total articles: {len(all_articles)}")

    # Phase 2: Fetch details in parallel
    results = []
    batch_size = MAX_WORKERS * 3
    for batch_start in range(0, len(all_articles), batch_size):
        batch = all_articles[batch_start:batch_start + batch_size]
        batch_urls = [u for u, _, _ in batch]
        log(f"  Details [{batch_start+1}-{batch_start+len(batch)}/{len(all_articles)}]...")
        detail_map = fetch_detail_batch(batch_urls)
        for url, title, date_str in batch:
            content_html, detail_date, detail_title = detail_map.get(url, (None, None, None))
            final_title = detail_title or title
            final_date = detail_date or date_str
            results.append((url, final_title, final_date, content_html or ""))

    # Save
    saved = save_to_db(results)

    log(f"\n{'=' * 50}")
    log(f"Done. Total: {len(results)} articles (saved: {saved})")
    if results:
        dates = sorted([r[2] for r in results if r[2]])
        log(f"Date range: {dates[0]} ~ {dates[-1]}")

    # FTS rebuild
    try:
        import sqlite3
        conn = sqlite3.connect(DB_PATH, timeout=60)
        conn.execute("""
            INSERT OR REPLACE INTO gov_search(rowid, title, site_name, summary)
            SELECT rowid, title, site_name,
                   CASE WHEN length(content) > 200 THEN substr(content, 1, 200) ELSE content END
            FROM gov_raw WHERE category = ?
        """, (COLUMN_NAME,))
        conn.commit()
        conn.close()
        log("FTS index rebuilt")
    except Exception as e:
        log(f"FTS rebuild note: {e}")


if __name__ == "__main__":
    main()