From 521d94061101f4924068023ee5dd52e0655dce47 Mon Sep 17 00:00:00 2001
From: Dotty Dotter <dotty@Mac.wideopen.space>
Date: Fri, 10 Apr 2026 17:40:46 +0200
Subject: [PATCH] #22 NI: Deduplizierung (Server liefert manche Treffer
 doppelt)

---
 app/parlamente.py | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/app/parlamente.py b/app/parlamente.py
index 4c21646..573a143 100644
--- a/app/parlamente.py
+++ b/app/parlamente.py
@@ -464,6 +464,7 @@ class PortalaAdapter(ParlamentAdapter):
         - WEV08[0].1    → PDF-URL (alternativ)
         """
         results: list[Drucksache] = []
+        seen: set[str] = set()
 
         for m in self._RE_JSON_COMMENT.finditer(html):
             try:
@@ -517,6 +518,11 @@ class PortalaAdapter(ParlamentAdapter):
                 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)
             if "antrag" not in (doc.typ or "").lower():
                 continue