# -*- coding: utf-8 -*-
# 증평 트리: ul.depth1_ul>li.th1>a.th1_lnk(d1) / ul.depth2_ul>li>a(d2) / ul.depth3_ul>li>a.lnk_2th(d3) [/depth4_ul]
import sys,io,json,urllib.request,ssl
from bs4 import BeautifulSoup
sys.stdout=io.TextIOWrapper(sys.stdout.buffer,encoding='utf-8')
ctx=ssl.create_default_context(); ctx.check_hostname=False; ctx.verify_mode=ssl.CERT_NONE
def get(u):
    req=urllib.request.Request(u,headers={'User-Agent':'Mozilla/5.0'})
    return urllib.request.urlopen(req,timeout=25,context=ctx).read().decode('utf-8','ignore')
h=get('https://www.jp.go.kr/kor/sitemap_11.do')
s=BeautifulSoup(h,'html.parser')
def txt(a): return a.get_text(strip=True) if a else ''
tree=[]
for li1 in s.select('ul.depth1_ul > li.th1'):
    a1=li1.select_one('a.th1_lnk'); d1=txt(a1)
    d2s=li1.select('ul.depth2_ul > li')
    if not d2s:
        tree.append([d1,'','','',a1.get('href')]); continue
    for li2 in d2s:
        a2=li2.find('a', recursive=False) or li2.find('a'); d2=txt(a2)
        d3s=li2.select('ul.depth3_ul > li')
        if not d3s:
            tree.append([d1,d2,'','',a2.get('href')]); continue
        for li3 in d3s:
            a3=li3.find('a'); d3=txt(a3)
            d4s=li3.select('ul.depth4_ul > li')
            if not d4s:
                tree.append([d1,d2,d3,'',a3.get('href')])
            else:
                for li4 in d4s:
                    a4=li4.find('a'); tree.append([d1,d2,d3,txt(a4),a4.get('href')])
json.dump(tree,open('_jp_tree.json','w',encoding='utf-8'),ensure_ascii=False)
print('증평 트리노드',len(tree))
for t in tree[:5]: print('  ',t[0],'>',t[1],'>',t[2],'>',t[3],'|',(t[4] or '')[:40])