df16c98366
공공기관2 작업 중. _temp 몽타주(재생성가능)는 제외. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
40 lines
1.5 KiB
Python
40 lines
1.5 KiB
Python
# -*- coding: utf-8 -*-
|
|
# 새창열림(_blank) → 사이트 소급적용: python _blank_apply.py <sitemap_url> <host> <excel>
|
|
import sys,io,urllib.request,ssl,re
|
|
from bs4 import BeautifulSoup
|
|
import openpyxl
|
|
sys.stdout=io.TextIOWrapper(sys.stdout.buffer,encoding='utf-8')
|
|
SM=sys.argv[1]; HOST=sys.argv[2]; F=sys.argv[3]
|
|
BASE='https://'+HOST
|
|
ctx=ssl.create_default_context(); ctx.check_hostname=False; ctx.verify_mode=ssl.CERT_NONE
|
|
try: ctx.set_ciphers('DEFAULT@SECLEVEL=1')
|
|
except: pass
|
|
def absu(u):
|
|
if not u: return ''
|
|
if u.startswith('http'): return u
|
|
if u.startswith('/'): return BASE+u
|
|
return BASE+'/'+u
|
|
def norm(u): return absu(u).split('#')[0].rstrip('/')
|
|
req=urllib.request.Request(SM,headers={'User-Agent':'Mozilla/5.0'})
|
|
h=urllib.request.urlopen(req,timeout=25,context=ctx).read().decode('utf-8','ignore')
|
|
s=BeautifulSoup(h,'html.parser')
|
|
blank=set()
|
|
for a in s.find_all('a'):
|
|
tg=(a.get('target') or ''); ti=(a.get('title') or ''); tx=a.get_text(strip=True)
|
|
if tg=='_blank' or '새창' in ti or '새창' in tx:
|
|
href=a.get('href')
|
|
if href and not href.startswith('javascript') and href!='#':
|
|
blank.add(norm(href))
|
|
print('새창열림 URL',len(blank))
|
|
wb=openpyxl.load_workbook(F); ws=wb.active
|
|
chg=0
|
|
for r in range(3,ws.max_row+1):
|
|
u=ws.cell(r,11).value
|
|
if not u: continue
|
|
if norm(u) in blank and ws.cell(r,12).value!='사이트':
|
|
ws.cell(r,12).value='사이트'
|
|
for c in (13,14,15,16,17): ws.cell(r,c).value=None
|
|
chg+=1
|
|
wb.save(F)
|
|
print('사이트 정정',chg,'행')
|