reinstate word column in icu word table

[nominatim.git] / settings / legacy_icu_tokenizer.yaml

diff --git a/settings/legacy_icu_tokenizer.yaml b/settings/legacy_icu_tokenizer.yaml
index 34cd8b0ba763c9d0abb784c0f1388dd9b053e53c..5fd30bd8f2ee194806c28942ced3f7ab3191d5c6 100644 (file)
--- a/settings/legacy_icu_tokenizer.yaml
+++ b/settings/legacy_icu_tokenizer.yaml
@@ -1,116 +1,56 @@
 normalization:
-    - ":: NFD ()"
-    - "[[:Nonspacing Mark:] [:Cf:]] >"
     - ":: lower ()"
+    - !include icu-rules/unicode-digits-to-decimal.yaml
+    - "'№' > 'no'"
+    - "'n°' > 'no'"
+    - "'nº' > 'no'"
+    - "ª > a"
+    - "º > o"
+    - "[[:Punctuation:][:Symbol:]]  > ' '"
     - "ß > 'ss'" # German szet is unimbigiously equal to double ss
-    - "[[:Punctuation:][:Space:]]+ > ' '"
+    - "[^[:Letter:] [:Number:] [:Space:]] >"
+    - "[:Lm:] >"
+    - ":: [[:Number:]] Latin ()"
+    - ":: [[:Number:]] Ascii ();"
+    - ":: [[:Number:]] NFD ();"
+    - "[[:Nonspacing Mark:] [:Cf:]] >;"
+    - "[:Space:]+ > ' '"
+transliteration:
+    - ":: Latin ()"
+    - !include icu-rules/extended-unicode-to-asccii.yaml
+    - ":: Ascii ()"
+    - ":: NFD ()"
+    - "[^[:Ascii:]] >"
+    - ":: lower ()"
     - ":: NFC ()"
-transliteration: icu_transliteration.rules
-compound_suffixes:
-    # Danish
-    - hal
-    - hallen
-    - hallerne
-    # German
-    - berg
-    - brücke
-    - fabrik
-    - gasse
-    - graben
-    - haus
-    - höhle
-    - hütte
-    - kapelle
-    - kogel
-    - pfad
-    - platz
-    - quelle
-    - spitze
-    - stiege
-    - strasse
-    - teich
-    - universität
-    - wald
-    - weg
-    - wiese
-    # Dutch
-    - gracht
-    - laan
-    - markt
-    - plein
-    - straat
-    - vliet
-    - weg
-    # Norwegian
-    - vei
-    - veien
-    - veg
-    - vegen
-    - gate
-    - gaten
-    - gata
-    - plass
-    - plassen
-    - sving
-    - svingen
-    # Finnish
-    - alue
-    - asema
-    - aukio
-    - kaari
-    - katu
-    - kuja
-    - kylä
-    - penger
-    - polku
-    - puistikko
-    - puisto
-    - raitti
-    - ranta
-    - rinne
-    - taival
-    - tie
-    - tori
-    - väylä
-    # Swedish
-    - väg
-    - vägen
-    - gatan
-    - gata
-    - gränd
-    - gränden
-    - stig
-    - stigen
-    - plats
-    - platsen
-abbreviations:
-    # German
-    - am => a
-    - an der => a d
-    - allgemeines krankenhaus => akh
-    - altstoffsammelzentrum => asz
-    - auf der => a d
-    - bach => b
-    - bad => b
-    - bahnhof => bhf,bf
-    - berg => bg
-    - bezirk => bez
-    - brücke => br
-    - burg => bg
-    - chaussee => ch
-    - deutsche,deutscher,deutsches => dt
-    - dorf => df
-    - doktor => dr
-    - fachhochschule => fh
-    - Freiwillige Feuerwehr => ff
-    - sankt => st
-    - strasse => str
-    - weg => wg
-    # English
-    - alley => al
-    - beach => bch
-    - street => st
-    - road => rd
-    - bridge => brdg
-
-
+variants:
+    - !include icu-rules/variants-bg.yaml
+    - !include icu-rules/variants-ca.yaml
+    - !include icu-rules/variants-cs.yaml
+    - !include icu-rules/variants-da.yaml
+    - !include icu-rules/variants-de.yaml
+    - !include icu-rules/variants-el.yaml
+    - !include icu-rules/variants-en.yaml
+    - !include icu-rules/variants-es.yaml
+    - !include icu-rules/variants-et.yaml
+    - !include icu-rules/variants-eu.yaml
+    - !include icu-rules/variants-fi.yaml
+    - !include icu-rules/variants-fr.yaml
+    - !include icu-rules/variants-gl.yaml
+    - !include icu-rules/variants-hu.yaml
+    - !include icu-rules/variants-it.yaml
+    - !include icu-rules/variants-ja.yaml
+    - !include icu-rules/variants-mg.yaml
+    - !include icu-rules/variants-ms.yaml
+    - !include icu-rules/variants-nl.yaml
+    - !include icu-rules/variants-no.yaml
+    - !include icu-rules/variants-pl.yaml
+    - !include icu-rules/variants-pt.yaml
+    - !include icu-rules/variants-ro.yaml
+    - !include icu-rules/variants-ru.yaml
+    - !include icu-rules/variants-sk.yaml
+    - !include icu-rules/variants-sl.yaml
+    - !include icu-rules/variants-sv.yaml
+    - !include icu-rules/variants-tr.yaml
+    - !include icu-rules/variants-uk.yaml
+    - !include icu-rules/variants-vi.yaml