Merge remote-tracking branch 'upstream/master'

[nominatim.git] / settings / icu_tokenizer.yaml

diff --git a/settings/icu_tokenizer.yaml b/settings/icu_tokenizer.yaml
index a3c62e67b8d2a500baee37223f56b151bce14324..50bb72d2eed9c03b9a9d76823b684a9413c88277 100644 (file)
--- a/settings/icu_tokenizer.yaml
+++ b/settings/icu_tokenizer.yaml
@@ -25,6 +25,13 @@ transliteration:
     - "[^a-z0-9[:Space:]] >"
     - ":: NFC ()"
 sanitizers:
     - "[^a-z0-9[:Space:]] >"
     - ":: NFC ()"
 sanitizers:

+    - step: clean-housenumbers
+      filter-kind:
+        - housenumber
+        - conscriptionnumber
+        - streetnumber
+      convert-to-name:
+        - (\A|.*,)[^\d,]{3,}(,.*|\Z)

     - step: split-name-list
     - step: strip-brace-terms
     - step: tag-analyzer-by-language
     - step: split-name-list
     - step: strip-brace-terms
     - step: tag-analyzer-by-language


@@ -59,6 +66,13 @@ token-analysis:
       mode: variant-only
       variants:
           - !include icu-rules/variants-de.yaml
       mode: variant-only
       variants:
           - !include icu-rules/variants-de.yaml

+      mutations:
+          - pattern: ä
+            replacements: ["ä", "ae"]
+          - pattern: ö
+            replacements: ["ö", "oe"]
+          - pattern: ü
+            replacements: ["ü", "ue"]

     - id: el
       analyzer: generic
       mode: variant-only
     - id: el
       analyzer: generic
       mode: variant-only