tobias/gwoe-antragspruefer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

#22 NI: Deduplizierung (Server liefert manche Treffer doppelt)

Browse Source
This commit is contained in:
Dotty Dotter 2026-04-10 17:40:46 +02:00
parent edcb4e9c76
commit 521d940611
1 changed files with 6 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
app/parlamente.py
View File

@ -464,6 +464,7 @@ class PortalaAdapter(ParlamentAdapter):
- WEV08[0].1 → PDF-URL (alternativ) - WEV08[0].1 → PDF-URL (alternativ)
""" """
results: list[Drucksache] = [] results: list[Drucksache] = []
seen: set[str] = set()
for m in self._RE_JSON_COMMENT.finditer(html): for m in self._RE_JSON_COMMENT.finditer(html):
try: try:
@ -517,6 +518,11 @@ class PortalaAdapter(ParlamentAdapter):
typ=typ or "Antrag", typ=typ or "Antrag",
) )
# Deduplizierung (NI liefert manche Treffer doppelt)
if drucksache in seen:
continue
seen.add(drucksache)
# Client-seitig Antrag-Filter (wie bei allen Adaptern) # Client-seitig Antrag-Filter (wie bei allen Adaptern)
if "antrag" not in (doc.typ or "").lower(): if "antrag" not in (doc.typ or "").lower():
continue continue