Marco/Bundescrawler
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

change formatting for individual output

Browse source
This commit is contained in:
Marco Lents 2025-11-14 12:38:11 +01:00
parent 7358b47384
commit 1abf37d175
1 changed files with 11 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

15
crawler.py
View file

@ -4,6 +4,7 @@ import json
from bs4 import BeautifulSoup from bs4 import BeautifulSoup
from time import sleep from time import sleep
from os.path import commonprefix from os.path import commonprefix
from os import makedirs
from git import Repo from git import Repo
import argparse import argparse
from datetime import datetime from datetime import datetime
@ -84,7 +85,8 @@ Biographie: {self.cv}
## Sonstige Gremien: {funcs_to_str(self.additional_functions)} ## Sonstige Gremien: {funcs_to_str(self.additional_functions)}
# Mandat {self.mandate[0], self.mandate[1]} # Mandat
{self.mandate[0], self.mandate[1]}
# Veröffentlichungspflichtige Angaben # Veröffentlichungspflichtige Angaben
{self.disclosures.replace(DISCLOSURE_DISLAIMER, "")} {self.disclosures.replace(DISCLOSURE_DISLAIMER, "")}
@ -145,9 +147,14 @@ def main():
def save_individuals(bios, out): def save_individuals(bios, out):
for rep in bios: for rep in bios:
first_letter = rep.name[0][0].upper() first_letter = rep.name[0][0].upper()
name_str = f"{rep.name[1]} {rep.name[0]}".replace(" ", "_") name_str = f"{rep.name[0]} {rep.name[1]}".replace(" ", "_")
with open(f"{out}/Abgeordnete/{first_letter}/{name_str}.md") as rep_file: dir = f"{out}/Abgeordnete/{first_letter}"
json.dump(str(rep), rep_file, indent=2, ensure_ascii=False) try:
makedirs(dir)
except FileExistsError:
pass
with open(f"{dir}/{name_str}.md", "w") as rep_file:
rep_file.write(str(rep))
def save_raw(bios, out): def save_raw(bios, out):