From: Sarah Hoffmann <lonvia@denofr.de>
Date: Wed, 19 Feb 2025 09:26:06 +0000 (+0100)
Subject: strip normalisation results of normal and special spaces
X-Git-Url: https://git.openstreetmap.org./nominatim.git/commitdiff_plain/55c31769576e3143b8e747a782c114a5198a17dc

strip normalisation results of normal and special spaces
---

diff --git a/src/nominatim_api/query_preprocessing/normalize.py b/src/nominatim_api/query_preprocessing/normalize.py
index afb93ded..0bb0c8ed 100644
--- a/src/nominatim_api/query_preprocessing/normalize.py
+++ b/src/nominatim_api/query_preprocessing/normalize.py
@@ -27,5 +27,5 @@ def create(config: QueryConfig) -> QueryProcessingFunc:
 
     return lambda phrases: list(
         filter(lambda p: p.text,
-               (Phrase(p.ptype, cast(str, normalizer.transliterate(p.text)))
+               (Phrase(p.ptype, cast(str, normalizer.transliterate(p.text)).strip('-: '))
                 for p in phrases)))
diff --git a/src/nominatim_api/search/icu_tokenizer.py b/src/nominatim_api/search/icu_tokenizer.py
index 6f1dcf79..487dd171 100644
--- a/src/nominatim_api/search/icu_tokenizer.py
+++ b/src/nominatim_api/search/icu_tokenizer.py
@@ -244,7 +244,7 @@ class ICUQueryAnalyzer(AbstractQueryAnalyzer):
             standardized form search will work with. All information removed
             at this stage is inevitably lost.
         """
-        return cast(str, self.normalizer.transliterate(text))
+        return cast(str, self.normalizer.transliterate(text)).strip('-: ')
 
     def split_query(self, query: qmod.QueryStruct) -> Tuple[QueryParts, WordDict]:
         """ Transliterate the phrases and split them into tokens.