# -*- coding: utf-8 -*-
"""진안 엑셀행 57~끝 본문 렌더 스크린샷(이미지 검출용). 페이지=#sub_contents/게시판=첫글. _nshots2/"""
import glob, os, re, json, urllib.request
from collections import Counter
import openpyxl
from playwright.sync_api import sync_playwright
DIR = os.path.dirname(__file__)
SHOT = os.path.join(DIR, '_nshots2')
os.makedirs(SHOT, exist_ok=True)
HDR = {'User-Agent': 'Mozilla/5.0'}

def get(u):
    try:
        return urllib.request.urlopen(urllib.request.Request(u.replace(' ', '%20'), headers=HDR), timeout=20).read().decode('utf-8', 'replace')
    except Exception:
        return ''

t = glob.glob(os.path.join(DIR, '*.xlsx'))
t = [p for p in t if 'backup' not in p and not os.path.basename(p).startswith(('~$', '_'))][0]
ws = openpyxl.load_workbook(t).active
targets = []
for r in range(57, ws.max_row + 1):              # 엑셀행 57부터
    L = ws.cell(r, 12).value
    if L in ('페이지', '게시판'):
        targets.append({'r': r, 'L': L, 'M': ws.cell(r, 13).value, 'N': str(ws.cell(r, 14).value or ''),
                        'lab': ws.cell(r, 7).value or ws.cell(r, 6).value or ws.cell(r, 5).value or '', 'K': str(ws.cell(r, 11).value or '')})

def first_post(k):
    h = get(k).replace('&', '&')
    menu = re.search(r'menuCd=([A-Z0-9_]+)', k)
    menu = menu.group(1) if menu else ''
    vb = [b for b in re.findall(r'view\.jinan\?[^"\']*?boardId=([A-Z0-9_]+)[^"\']*?dataSid=', h) if b != 'BBS_0000157']
    if not vb:
        return None
    bid = Counter(vb).most_common(1)[0][0]
    lh = get(f'http://www.jinan.go.kr/board/list.jinan?boardId={bid}&menuCd={menu}&paging=ok&startPage=1').replace('&', '&')
    vs = [v for v in re.findall(r'view\.jinan\?([^"\']+)', lh) if f'boardId={bid}' in v and 'dataSid=' in v]
    return ('http://www.jinan.go.kr/board/view.jinan?' + vs[0].replace(' ', '%20')) if vs else None

with sync_playwright() as pw:
    br = pw.chromium.launch()
    pg = br.new_page(viewport={'width': 1100, 'height': 1500})
    for i, tg in enumerate(targets):
        out = os.path.join(SHOT, f"r{tg['r']}.png")
        if os.path.exists(out):
            continue
        url = tg['K']
        if tg['L'] == '게시판' and tg['M'] not in (0, None):
            pu = first_post(tg['K'])
            if pu:
                url = pu
        try:
            pg.goto(url, timeout=25000, wait_until='domcontentloaded')
            pg.wait_for_timeout(1100)
            el = None
            for sel in ['#sub_contents', '#contents']:
                el = pg.query_selector(sel)
                if el:
                    break
            (el or pg).screenshot(path=out)
        except Exception as e:
            print('ERR', tg['r'], str(e)[:50], flush=True)
        if (i + 1) % 30 == 0:
            print(f'... {i+1}/{len(targets)}', flush=True)
    br.close()
json.dump(targets, open(os.path.join(DIR, '_ntargets2.json'), 'w', encoding='utf-8'), ensure_ascii=False, indent=0)
print('DONE', len(targets))