# -*- coding: utf-8 -*- """익산 91~끝: 부모컨테이너 페이지의 .snb 자식탭을 lc+1열에 전개(재정운용상황공개식). 가드: 행 menuUid가 snb탭 목록에 없음(=부모컨테이너) & 탭>=2 & lc+1 빈칸. 부모행 lc+1=첫탭·나머지 탭 행삽입·lc병합·process_row(L/M/N). 1~90 보존. 사용: python -X utf8 _익산_snb탭.py [--write] """ import sys, os, re, shutil, importlib.util from urllib.parse import urljoin import openpyxl from openpyxl.worksheet.hyperlink import Hyperlink HERE=os.path.dirname(os.path.abspath(__file__)) s=importlib.util.spec_from_file_location('m',os.path.join(HERE,'_jeonbuk_phase234_all.py')) M=importlib.util.module_from_spec(s); s.loader.exec_module(M) XLSX=r'D:\01.프로젝트\DB수집\작업파일\광역_사이트맵\전북특별자치도\9.익산시\전북특별자치도_익산시.xlsx' BASE='https://www.iksan.go.kr' START_ROW=91 SESS=M.make_session() def clean(t): return re.sub(r'현재단계|현재페이지|하위메뉴.*|닫기|열기','',t or '').strip() def muid(u): mm=re.search(r'menuUid=([0-9a-f]+)',u or ''); return mm.group(1) if mm else None def leafcol(ws,r): for c in range(10,3,-1): if ws.cell(r,c).value not in (None,''): return c return None def snb_tabs(url): soup=M.fetch(SESS,url) if soup is None: return None snb=soup.select_one('ul.snb') or soup.select_one('.snb') if not snb: return None out=[] for li in snb.find_all('li',recursive=False): a=li.find('a',href=True) if a: t=clean(a.get_text()); if t: out.append((t, urljoin(BASE,a['href']))) return out def remap_insert(ws, ins, n): ranges=[(x.min_col,x.min_row,x.max_col,x.max_row) for x in list(ws.merged_cells.ranges)] for x in list(ws.merged_cells.ranges): ws.unmerge_cells(str(x)) ws.insert_rows(ins,n) for mc,mr,xc,xr in ranges: if mr>=ins: mr+=n; xr+=n elif xr>=ins: xr+=n ws.merge_cells(start_row=mr,start_column=mc,end_row=xr,end_column=xc) def main(): write='--write' in sys.argv wb=openpyxl.load_workbook(XLSX); ws=wb.active C=ws.cell(3,3).value plan=[] for r in range(START_ROW, ws.max_row+1): K=ws.cell(r,11).value if not(isinstance(K,str) and 'menuUid=' in K): continue lc=leafcol(ws,r) if not lc or lc>=10: continue if ws.cell(r,lc+1).value not in (None,''): continue # 이미 전개됨 tabs=snb_tabs(K) if not tabs or len(tabs)<2: continue tabids=[muid(u) for _,u in tabs] if muid(K) in tabids: continue # 행이 탭 중 하나(리프/형제) → 부모아님 plan.append((r,lc,ws.cell(r,lc).value,tabs)) print('전개대상 %d개'%len(plan)) for r,lc,lab,tabs in plan: print(' r%d lc%d %s → %d탭: %s'%(r,lc,lab,len(tabs),[t for t,_ in tabs])) if not write: return shutil.copy(XLSX, XLSX.replace('.xlsx','_backup_snb탭전.xlsx')) tot=0 for r,lc,lab,tabs in sorted(plan,key=lambda t:-t[0]): # bottom-up ws.cell(r,lc+1).value=tabs[0][0] out0=M.process_row(SESS,tabs[0][1],M.BODY_SEL) ws.cell(r,11).value=tabs[0][1] ws.cell(r,12).value=out0['L'] or '페이지' ws.cell(r,13).value=out0['M'] if out0['M']!='' else 1 ws.cell(r,14).value=out0['N']; ws.cell(r,15).value=out0['O']; ws.cell(r,16).value=out0['P']; ws.cell(r,17).value=out0['Q'] rest=tabs[1:]; n=len(rest) if n: remap_insert(ws,r+1,n) for i,(clab,curl) in enumerate(rest): rr=r+1+i out=M.process_row(SESS,curl,M.BODY_SEL) ws.cell(rr,3).value=C ws.cell(rr,lc+1).value=clab ws.cell(rr,11).value=curl ws.cell(rr,12).value=out['L'] or '페이지' ws.cell(rr,13).value=out['M'] if out['M']!='' else 1 ws.cell(rr,14).value=out['N']; ws.cell(rr,15).value=out['O']; ws.cell(rr,16).value=out['P']; ws.cell(rr,17).value=out['Q'] ws.row_dimensions[rr].height=15 ws.merge_cells(start_row=r,start_column=lc,end_row=r+n,end_column=lc) tot+=n print(' r%d %s +%d (첫탭=%s)'%(r,lab,n,tabs[0][0])) for r in range(3,ws.max_row+1): v=ws.cell(r,11).value; cell=ws.cell(r,11) if isinstance(v,str) and v.startswith('http'): cell.hyperlink=Hyperlink(ref=cell.coordinate,target=v) else: cell.hyperlink=None wb.save(XLSX) print('적용: +%d행 · 하이퍼링크재구성 · 백업'%tot) if __name__=='__main__': main()