fix Norwegian variants

More cases of 'no' being interpreted as fasle by yaml.

diff --git a/settings/icu-rules/variants-no.yaml b/settings/icu-rules/variants-no.yaml
index de75e3c389e003cd837f38c2b36cbeb9b1c9cf7b..dd17408ff21030a3b376bd921ab87073f17a02dd 100644 (file)
--- a/settings/icu-rules/variants-no.yaml
+++ b/settings/icu-rules/variants-no.yaml
@@ -1,11 +1,10 @@
 # Source: https://wiki.openstreetmap.org/wiki/Name_finder:Abbreviations#Norsk_-_Norwegian
-- lang: no
+- lang: "no"
   words:
     # convert between Nynorsk and Bookmal here
-    -  vei, veg => v,vn,vei,veg
-    -  veien, vegen -> v,vn,veien,vegen
-    -  gate -> g,gt
+    -  ~vei, ~veg -> v,vei,veg
+    -  ~veien, ~vegen -> vn,veien,vegen
     # convert between the two female forms
-    -  gaten, gata => g,gt,gaten,gata
+    -  gate, gaten, gata -> g,gt
     -  plass, plassen -> pl
     -  sving, svingen -> sv

diff --git a/settings/icu_tokenizer.yaml b/settings/icu_tokenizer.yaml
index a6545bdd1ca7ff9244496aa24ad9a314cb1dfe67..9d7778aa554edd4d2bd9243e24ab10baf4225fce 100644 (file)
--- a/settings/icu_tokenizer.yaml
+++ b/settings/icu_tokenizer.yaml
@@ -46,7 +46,7 @@ sanitizers:
     - step: strip-brace-terms
     - step: tag-analyzer-by-language
       filter-kind: [".*name.*"]
-      whitelist: [bg,ca,cs,da,de,el,en,es,et,eu,fi,fr,gl,hu,it,ja,mg,ms,nl,no,pl,pt,ro,ru,sk,sl,sv,tr,uk,vi]
+      whitelist: [bg,ca,cs,da,de,el,en,es,et,eu,fi,fr,gl,hu,it,ja,mg,ms,nl,"no",pl,pt,ro,ru,sk,sl,sv,tr,uk,vi]
       use-defaults: all
       mode: append
     - step: tag-japanese
@@ -158,7 +158,7 @@ token-analysis:
       mode: variant-only
       variants:
           - !include icu-rules/variants-nl.yaml
-    - id: no
+    - id: "no"
       analyzer: generic
       mode: variant-only
       variants: