df16c98366
공공기관2 작업 중. _temp 몽타주(재생성가능)는 제외. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
35 lines
1.5 KiB
Python
35 lines
1.5 KiB
Python
# -*- coding: utf-8 -*-
|
|
# 청양 LNB(좌측메뉴) 라이브 추출: 군정소식 섹션 페이지에서 depth2/3 메뉴트리 + 각 링크 href
|
|
import io,sys,json
|
|
from playwright.sync_api import sync_playwright
|
|
sys.stdout=io.TextIOWrapper(sys.stdout.buffer,encoding='utf-8')
|
|
# 군정소식 대분류 진입(새소식)
|
|
U='https://www.cheongyang.go.kr/kor/sub04_01_01.do'
|
|
JS=r'''()=>{
|
|
function txt(a){return (a.innerText||a.textContent||'').trim().replace(/\s+/g,' ');}
|
|
const out=[];
|
|
// depth2 = 중분류, depth3 = 소분류
|
|
document.querySelectorAll('.lnb_depth2__list, ul.depth2_ul').forEach(d2=>{
|
|
d2.querySelectorAll(':scope > li').forEach(li2=>{
|
|
const a2=li2.querySelector(':scope > a, :scope > .lnb_depth2__btn, :scope > button');
|
|
const name2=a2?txt(a2):'';
|
|
const kids=[];
|
|
li2.querySelectorAll('.lnb_depth3__list li > a, ul.depth3_ul li > a').forEach(a3=>{
|
|
kids.push({n:txt(a3), href:a3.getAttribute('href')});
|
|
});
|
|
out.push({d2:name2, href:(a2?a2.getAttribute('href'):null), kids});
|
|
});
|
|
});
|
|
return out;
|
|
}'''
|
|
with sync_playwright() as p:
|
|
b=p.chromium.launch(); pg=b.new_page(viewport={'width':1400,'height':2400}); pg.set_default_timeout(25000)
|
|
pg.goto(U,wait_until='networkidle'); pg.wait_for_timeout(1500)
|
|
print('진입 final:',pg.url)
|
|
data=pg.evaluate(JS)
|
|
for g in data:
|
|
print('\n[중] %s (%s)'%(g['d2'],g['href']))
|
|
for k in g['kids']:
|
|
print(' - %s | %s'%(k['n'],k['href']))
|
|
b.close()
|