PearlTree dirty export from HTML (when it's too big to export by zip)

pearltrees_html_to_wget.py

#!/usr/bin/python3
# pearltress export HTML sauvegarde
import re
import random
from urllib.parse import unquote

RE_FOLDER = re.compile('<H3 FOLDED ADD_DATE="[^"]*">([^<]*)<\/H3>')
RE_LINK = re.compile('HREF="([^"]*)"')
RE_NAME = re.compile('filename="([^"]*)"')

with open("export.html", "r") as f:
  for line in f.readlines():
    if "H3 FOLDED" in line:
      folder = RE_FOLDER.search(line).group(1)
      print(f'mkdir "{folder}"; cd "{folder}"')
    if "</DL>" in line:
      print("cd ..")
    if "HREF" in line:
      link = RE_LINK.search(line).group(1)
      print(f'echo wget --content-disposition "{link}"')
      print(f'wget --content-disposition "{link}"')