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

improve no disclosure handling

Browse source
This commit is contained in:
Marco Lents 2025-11-16 13:01:56 +01:00
parent 5501f07cf7
commit 3787f756f3
1 changed files with 4 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
crawler.py
View file

@ -303,7 +303,11 @@ def get_disclosures(elem):
continue continue
if child.text.strip() == "": if child.text.strip() == "":
continue continue
if child.text.strip() == "Keine veröffentlichungspflichtigen Angaben.":
continue
current_body.append(child.text.strip()) current_body.append(child.text.strip())
if current_heading == "" and current_body == []:
continue
out.append((current_heading, current_body)) out.append((current_heading, current_body))
return out return out