feat(norway): classifying more addresses for norway (#171)

* Activating the DirectionalClassifier and CompoundStreetClassifier for norway,
adding additional street_types and directionals for norway.

* Removing street types lien and land as it crashes with other countries addresses.

* Moving the custom tokens in separate file and adding unit tests.

---------

Co-authored-by: mansoor.sajjad <[email protected]>

classifier/CompoundStreetClassifier.js

@@ -10,13 +10,13 @@ class CompoundStreetClassifier extends WordClassifier {
     // load street suffixes
     this.suffixes = {}
 
-    libpostal.load(this.suffixes, ['de', 'nl', 'sv'], 'concatenated_suffixes_separable.txt', {
+    libpostal.load(this.suffixes, ['de', 'nl', 'sv', 'nb'], 'concatenated_suffixes_separable.txt', {
       // remove any suffixes which contain less than 3 characters (excluding a period)
       // this removes suffixes such as 'r.' which can be ambiguous
       minlength: 3
     })
 
-    libpostal.load(this.suffixes, ['de', 'nl'], 'concatenated_suffixes_inseparable.txt', {
+    libpostal.load(this.suffixes, ['de', 'nl', 'nb'], 'concatenated_suffixes_inseparable.txt', {
       // remove any suffixes which contain less than 3 characters (excluding a period)
       // this removes suffixes such as 'r.' which can be ambiguous
       minlength: 3

classifier/DirectionalClassifier.js

@@ -9,7 +9,7 @@ const libpostal = require('../resources/libpostal/libpostal')
 
 // optionally control which languages are included
 // note: reducing the languages will have a considerable performance benefit
-const languages = ['en', 'es', 'de', 'fr', 'nl']
+const languages = ['en', 'es', 'de', 'fr', 'nl', 'nb']
 
 class DirectionalClassifier extends WordClassifier {
   setup () {

resources/pelias/dictionaries/libpostal/nb/concatenated_suffixes_inseparable.txt

@@ -0,0 +1,40 @@
+plass|pl.|p.
+haugen
+skogen|skog
+gården|gård
+åsen
+haugen
+skogen|skog
+gården|gård
+åsen
+gård|gården
+holm|holmen
+haugen|Haugane
+krysset
+skogen|skog
+grend|grenda
+flat|flaten
+våg|vågen
+strand|stranda|stranden
+vik|vika
+platå
+stad
+terrasse
+hage|hagen
+øya
+lund|lunden
+holm|holmen
+berg|berget
+landa|lande|landet
+dal|dalen
+sund
+bakke|bakken
+høgda|høgde
+borg|borgen
+gard|garden|garda
+mark|marka
+voll|vollen
+sjø|sjøen
+lie
+sti|stien
+tangen

resources/pelias/dictionaries/libpostal/nb/directionals.txt

@@ -0,0 +1,8 @@
+øvre|ovre
+øvste|ovste
+nedre
+midtre
+ytre
+indre
+sør|sor|s
+sørre|sorre

resources/pelias/dictionaries/libpostal/nb/street_types.txt

@@ -0,0 +1,36 @@
+plass
+haugen
+skogen|skog
+gården|gård
+åsen
+gård|gården
+holm|holmen
+haugen|Haugane
+krysset
+skogen|skog
+grend|grenda
+flat|flaten
+våg|vågen
+strand|stranda|stranden
+vik|vika
+platå
+stad
+terrasse
+hage|hagen|hager
+øya
+lund|lunden
+holm|holmen
+berg|berget
+landa|lande|landet
+dal|dalen
+sund
+bakke|bakken
+høgda|høgde
+borg|borgen
+gard|garden|garda
+mark|marka
+voll|vollen
+sjø|sjøen
+lie
+sti|stien
+tangen

test/address.nor.test.js

@@ -0,0 +1,47 @@
+const testcase = (test, common) => {
+  let assert = common.assert(test)
+
+  assert('Skredderplassen 20', [
+    { street: 'Skredderplassen' }, { housenumber: '20' }
+  ])
+
+  assert('Henrik Walters plass 3', [
+    { street: 'Henrik Walters plass' }, { housenumber: '3' }
+  ])
+
+  assert('Vestre Haugen 74', [
+    { street: 'Vestre Haugen' }, { housenumber: '74' }
+  ])
+
+  assert('Maria Dehlis vei 15', [
+    { street: 'Maria Dehlis vei' }, { housenumber: '15' }
+  ])
+
+  assert('Gamle Drammensvei 163', [
+    { street: 'Gamle Drammensvei' }, { housenumber: '163' }
+  ])
+
+  assert('Grense Jakobselv veien 311', [
+    { street: 'Grense Jakobselv veien' }, { housenumber: '311' }
+  ])
+
+  assert('Epleskogen 39A', [
+    { street: 'Epleskogen' }, { housenumber: '39A' }
+  ])
+
+  assert('Øvste Skogen 121', [
+    { street: 'Øvste Skogen' }, { housenumber: '121' }
+  ])
+
+  assert('Tindvegen nedre 44B', [
+    { street: 'Tindvegen nedre' }, { housenumber: '44B' }
+  ])
+}
+
+module.exports.all = (tape, common) => {
+  function test (name, testFunction) {
+    return tape(`address AUS: ${name}`, testFunction)
+  }
+
+  testcase(test, common)
+}