[languages] If a non-Latin script in a string would prohibit the found language, return ambiguous. Adding some test cases for sanity checking the labeling

scripts/geodata/language_id/disambiguation.py

@@ -124,7 +124,7 @@ street_types_gazetteer = DictionaryPhraseFilter('street_types.txt',
                                                 'stopwords.txt',)
 
 char_scripts = get_chars_by_script()
-script_languages = get_script_languages()
+script_languages = {script: set(langs) for script, langs in get_script_languages().iteritems()}
 
 UNKNOWN_SCRIPT = 'Unknown'
 COMMON_SCRIPT = 'Common'
@@ -148,15 +148,23 @@ def get_string_script(s):
             last_script = script
     return (last_script, script_len, is_ascii)
 
-
+LATIN_SCRIPT = 'Latin'
 UNKNOWN_LANGUAGE = 'unk'
 AMBIGUOUS_LANGUAGE = 'xxx'
 
 
 def disambiguate_language(text, languages):
-    num_defaults = sum((1 for lang, default in languages if default))
     valid_languages = OrderedDict(languages)
-    tokens = tokenize(safe_decode(text).replace(u'-', u' ').lower())
+    script_langs = {}
+    read_len = 0
+    while read_len < len(text):
+        script, script_len, is_ascii = get_string_script(text[read_len:])
+        if script != LATIN_SCRIPT:
+            script_langs[script] = set([l for l, d in languages if l in script_languages[script]])
+        read_len += script_len
+
+    num_defaults = sum((1 for lang, default in valid_languages.iteritems() if default))
+    tokens = [(c, t.rstrip('.')) for c, t in tokenize(safe_decode(text).replace(u'-', u' ').lower())]
 
     current_lang = None
     possible_lang = None
@@ -172,7 +180,7 @@ def disambiguate_language(text, languages):
             for lang, canonical, stopword in data:
                 canonical = int(canonical)
                 stopword = int(stopword)
-                if lang not in valid_languages or stopword:
+                if lang not in valid_languages or (stopword and len(potentials) > 1):
                     continue
                 is_default = valid_languages[lang]
 
@@ -180,7 +188,7 @@ def disambiguate_language(text, languages):
                     valid.append(lang)
                 elif is_default and num_defaults > 1 and current_lang != lang:
                     return AMBIGUOUS_LANGUAGE
-                elif not seen_languages and len(t[0][1]) > 1:
+                elif not seen_languages and len(potentials) == 1 and len(t[0][1]) > 1:
                     possible_lang = lang if possible_lang is None or possible_lang == lang else None
 
             if seen_languages and valid and not any((l in seen_languages for l in valid)):
@@ -189,16 +197,20 @@ def disambiguate_language(text, languages):
             if len(valid) == 1:
                 current_lang = valid[0]
             else:
-                valid = [l for l in valid if valid_languages.get(l)]
-                if len(valid) == 1 and current_lang is not None and valid[0] != current_lang:
+                valid_default = [l for l in valid if valid_languages.get(l)]
+                if len(valid_default) == 1 and current_lang is not None and valid_default[0] != current_lang:
                     return AMBIGUOUS_LANGUAGE
-                elif len(valid) == 1:
-                    current_lang = valid[0]
-
-            seen_languages.update(valid)
+                elif len(valid_default) == 1:
+                    current_lang = valid_default[0]
 
     if current_lang is not None:
-        return current_lang
+        if not any((current_lang not in langs for script, langs in script_langs.iteritems())):
+            return current_lang
+        else:
+            return AMBIGUOUS_LANGUAGE
     elif possible_lang is not None:
-        return possible_lang
+        if not any((possible_lang not in langs for script, langs in script_langs.iteritems())):
+            return possible_lang
+        else:
+            return AMBIGUOUS_LANGUAGE
     return UNKNOWN_LANGUAGE

scripts/geodata/tests/test_disambiguation.py

@@ -0,0 +1,108 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+from __future__ import unicode_literals
+
+import os
+import sys
+import unittest
+
+this_dir = os.path.realpath(os.path.dirname(__file__))
+sys.path.append(os.path.realpath(os.path.join(os.pardir, os.pardir)))
+
+from geodata.i18n.languages import init_languages, get_country_languages, get_regional_languages
+from geodata.language_id.disambiguation import disambiguate_language, street_types_gazetteer, UNKNOWN_LANGUAGE, AMBIGUOUS_LANGUAGE
+
+
+country_test_cases = [
+    # String, country, expected language
+
+    # US has some Spanish and French street names
+    ('Avenue P', 'us', 'en'),
+    ('Avenue du', 'us', 'fr'),
+    ('Avenida de la Plata', 'us', 'es'),
+    ('Pl', 'us', UNKNOWN_LANGUAGE),
+    ('No 2 School House', 'us', UNKNOWN_LANGUAGE),
+    ('E Thetford Rd', 'us', 'en'),
+    ('El Camino', 'us', 'es'),
+    ('Rue Louis Phillippe', 'us', 'fr'),
+    ('Calle Street', 'us', AMBIGUOUS_LANGUAGE),
+
+    # Avenue + stopword
+    ('Avenue du Bourget-du-Lac', 'je', 'fr'),
+
+    # UAE, English is non-default, has abbreviation
+    ('128 A St', 'ae', 'en'),
+    ('128 A St.', 'ae', 'en'),
+
+    # English / Arabic street address
+    ('Omar Street ﺵﺍﺮﻋ ﻊﻣﺭ', 'iq', AMBIGUOUS_LANGUAGE),
+
+
+    # Random script
+    ('Bayard Street - 擺也街', 'us', AMBIGUOUS_LANGUAGE),
+
+
+    # Brussels address
+    ('Avenue Paul Héger - Paul Hégerlaan', 'be', AMBIGUOUS_LANGUAGE),
+    ('Smaragdstraat', 'be', 'nl'),
+
+
+    # India
+    ('Kidwai nagar', 'in', 'hi'),
+    ('Mavoor Rd.', 'in', 'en'),
+
+    # Sri Lanka
+    ('Sri Sadathissa Mawatha', 'lk', 'si'),
+
+    # Russian
+    ('Фрунзе улица', 'kg', 'ru'),
+]
+
+regional_test_cases = [
+    # Spain
+    ('Carrer de la Morella', 'es', 'qs_a1r', 'Cataluña/Catalunya', 'ca'),
+    ('Calle de la Morella', 'es', 'qs_a1r', 'Cataluña/Catalunya', 'es'),
+    ('autobidea', 'es', 'qs_a1r', 'Comunidad Foral de Navarra', 'eu'),
+    ('Calle', 'es', 'qs_a1r', 'Comunidad Foral de Navarra', 'es'),
+    ('Txurruka', 'es', 'qs_a1r', 'País Vasco/Euskadi', UNKNOWN_LANGUAGE),
+
+    # Belgium
+    ('Lutticherstrasse', 'be', 'qs_a1', 'Liège', 'de'),
+    ('Chaussée de Charleroi', 'be', 'qs_a1', 'Namur', 'fr'),
+
+
+    # France / Occitan
+    ('Carriera de Brasinvert', 'fr', 'qs_a1r', 'Rhône-Alpes', 'oc'),
+
+]
+
+
+class TestNormalization(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        init_languages()
+        street_types_gazetteer.configure()
+
+    def test_countries(self):
+        for s, country, expected in country_test_cases:
+            languages = get_country_languages(country)
+            self.assertTrue(bool(languages))
+
+            lang = disambiguate_language(s, languages.items())
+            self.assertEqual(lang, expected, '{} != {} for {}, langs={}'.format(lang, expected, s, languages.items()))
+
+    def test_regional(self):
+        for s, country, k, v, expected in regional_test_cases:
+            languages = get_country_languages(country)
+            self.assertTrue(bool(languages))
+            regional = get_regional_languages(country, k, v)
+            self.assertTrue(bool(regional))
+            regional.update(languages)
+
+            lang = disambiguate_language(s, regional.items())
+
+            self.assertEqual(lang, expected, '{} != {} for {}, langs={}'.format(lang, expected, s, regional.items()))
+
+if __name__ == '__main__':
+    unittest.main()