df16c98366
공공기관2 작업 중. _temp 몽타주(재생성가능)는 제외. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
50 lines
2.7 KiB
Python
50 lines
2.7 KiB
Python
# -*- coding: utf-8 -*-
|
|
"""ESG 마이크로사이트 추출 + 메인 트리에 병합 → _komipo_tree.json 갱신"""
|
|
import json, re
|
|
from playwright.sync_api import sync_playwright
|
|
UA='Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/120 Safari/537.36'
|
|
BASE='https://www.komipo.co.kr'
|
|
def parse(href):
|
|
href=(href or '').strip()
|
|
m=re.search(r"goPage\('([A-Z]+)'\s*,\s*'([^']+)'", href)
|
|
if m:
|
|
typ=m.group(1); path=m.group(2); url=path if path.startswith('http') else BASE+path; return typ,url
|
|
if href.startswith('http'): return ('EXT' if 'komipo.co.kr' not in href else 'CONTENT'),href
|
|
if href.startswith('/'): return 'CONTENT',BASE+href
|
|
return '',''
|
|
with sync_playwright() as p:
|
|
b=p.chromium.launch(); pg=b.new_page(user_agent=UA,viewport={'width':1600,'height':2400})
|
|
pg.goto('https://www.komipo.co.kr/esg/main/main.do',timeout=40000,wait_until='domcontentloaded'); pg.wait_for_timeout(3000)
|
|
rows=pg.evaluate(r"""()=>{
|
|
const nav=document.querySelector('nav'); const out=[];
|
|
for(const li1 of nav.querySelectorAll(':scope>ul>li')){
|
|
const D=(li1.querySelector(':scope>a')?.textContent||'').trim().replace(/\s+/g,' ');
|
|
li1.querySelectorAll('li.more').forEach(m=>{const E=(m.querySelector(':scope>p')?.textContent||'').trim().replace(/\s+/g,' ');m.querySelectorAll('ul.depth3>li>a').forEach(a=>out.push({E2:D,F2:E,G2:(a.textContent||'').trim().replace(/\s+/g,' '),href:a.getAttribute('href')||''}));});
|
|
}
|
|
return out;
|
|
}""")
|
|
b.close()
|
|
esg=[]
|
|
for r in rows:
|
|
typ,url=parse(r['href']); mn=''
|
|
m=re.search(r'mnCd=([A-Z0-9]+)',url);
|
|
if m: mn=m.group(1)
|
|
# D=ESG경영, E=esg대분류, F=esg중분류, G=esg소분류
|
|
esg.append({'D':'ESG경영','E':r['E2'],'F':r['F2'],'G':r['G2'],'type':typ,'url':url,'mnCd':mn})
|
|
tree=json.load(open('_komipo_tree.json',encoding='utf-8'))
|
|
tree=[t for t in tree if t['D']!='ESG경영'] # 빈 ESG 제거
|
|
# G 키 없는 기존 항목에 G='' 추가
|
|
for t in tree: t.setdefault('G','')
|
|
# ESG경영 위치: 원래 5번째 대분류였음 → 순서 유지 위해 기업홍보 앞에 삽입
|
|
order=['소통채널','회사소개','기업지원','고객지원','ESG경영','기업홍보']
|
|
merged=[]
|
|
for D in order:
|
|
if D=='ESG경영': merged+=esg
|
|
else: merged+=[t for t in tree if t['D']==D]
|
|
json.dump(merged,open('_komipo_tree.json','w',encoding='utf-8'),ensure_ascii=False,indent=1)
|
|
from collections import Counter
|
|
print('ESG leaf:',len(esg),'| 최종 트리:',len(merged))
|
|
print('대분류순서:',list(dict.fromkeys(t['D'] for t in merged)))
|
|
print('type:',dict(Counter(t['type'] for t in merged)))
|
|
print('4단(G있음):',sum(1 for t in merged if t.get('G')))
|