ICU: better letter identification in normalization

[nominatim.git] / settings / icu_tokenizer.yaml

diff --git a/settings/icu_tokenizer.yaml b/settings/icu_tokenizer.yaml
index 50bb72d2eed9c03b9a9d76823b684a9413c88277..cd9c0d6dd56974888c9e12fdf834b51a5b55b22e 100644 (file)
--- a/settings/icu_tokenizer.yaml
+++ b/settings/icu_tokenizer.yaml
@@ -8,8 +8,8 @@ normalization:
     - "ª > a"
     - "º > o"
     - "[[:Punctuation:][:Symbol:]\u02bc]  > ' '"
-    - "ß > 'ss'" # German szet is unimbigiously equal to double ss
-    - "[^[:Letter:] [:Number:] [:Space:]] >"
+    - "ß > 'ss'" # German szet is unambiguously equal to double ss
+    - "[^[:alnum:] [:Canonical_Combining_Class=Virama:] [:Space:]] >"
     - "[:Lm:] >"
     - ":: [[:Number:]] Latin ()"
     - ":: [[:Number:]] Ascii ();"
@@ -41,6 +41,8 @@ sanitizers:
       mode: append
 token-analysis:
     - analyzer: generic
+    - id: "@housenumber"
+      analyzer: housenumbers
     - id: bg
       analyzer: generic
       mode: variant-only