Merge pull request #2486 from lonvia/fix-special-phrases

Fix parsing of operator in special phrases

nominatim/tools/special_phrases/special_phrase.py

@@ -16,4 +16,5 @@ class SpecialPhrase():
         # Hack around a bug where building=yes was imported with quotes into the wiki
         self.p_type = re.sub(r'\"|"', '', p_type.strip())
         # Needed if some operator in the wiki are not written in english
+        p_operator = p_operator.strip().lower()
         self.p_operator = '-' if p_operator not in ('near', 'in') else p_operator

test/python/test_tools_sp_wiki_loader.py

@@ -30,7 +30,7 @@ def test_parse_xml(sp_wiki_loader, xml_wiki_content):
         Should return the right SpecialPhrase objects.
     """
     phrases = sp_wiki_loader.parse_xml(xml_wiki_content)
-    assert check_phrases_content(phrases)
+    check_phrases_content(phrases)
 
 
 def test_next(sp_wiki_loader):
@@ -40,15 +40,29 @@ def test_next(sp_wiki_loader):
         the 'en' special phrases.
     """
     phrases = next(sp_wiki_loader)
-    assert check_phrases_content(phrases)
+    check_phrases_content(phrases)
 
 def check_phrases_content(phrases):
     """
         Asserts that the given phrases list contains
         the right phrases of the 'en' special phrases.
     """
-    return  len(phrases) > 1 \
+    assert set((p.p_label, p.p_class, p.p_type, p.p_operator) for p in phrases) ==\
-            and any(p.p_label == 'Embassies' and p.p_class == 'amenity' and p.p_type == 'embassy'
+              {('Zip Line', 'aerialway', 'zip_line', '-'),
-                    and p.p_operator == '-' for p in phrases) \
+               ('Zip Lines', 'aerialway', 'zip_line', '-'),
-            and any(p.p_label == 'Zip Line' and p.p_class == 'aerialway' and p.p_type == 'zip_line'
+               ('Zip Line in', 'aerialway', 'zip_line', 'in'),
-                    and p.p_operator == '-' for p in phrases)
+               ('Zip Lines in', 'aerialway', 'zip_line', 'in'),
+               ('Zip Line near', 'aerialway', 'zip_line', 'near'),
+               ('Animal shelter', 'amenity', 'animal_shelter', '-'),
+               ('Animal shelters', 'amenity', 'animal_shelter', '-'),
+               ('Animal shelter in', 'amenity', 'animal_shelter', 'in'),
+               ('Animal shelters in', 'amenity', 'animal_shelter', 'in'),
+               ('Animal shelter near', 'amenity', 'animal_shelter', 'near'),
+               ('Animal shelters near', 'amenity', 'animal_shelter', 'near'),
+               ('Drinking Water near', 'amenity', 'drinking_water', 'near'),
+               ('Water', 'amenity', 'drinking_water', '-'),
+               ('Water in', 'amenity', 'drinking_water', 'in'),
+               ('Water near', 'amenity', 'drinking_water', 'near'),
+               ('Embassy', 'amenity', 'embassy', '-'),
+               ('Embassys', 'amenity', 'embassy', '-'),
+               ('Embassies', 'amenity', 'embassy', '-')}

test/testdata/special_phrases_test_content.txt

@@ -70,9 +70,9 @@
 <model>wikitext</model>
 <format>text/x-wiki</format>
 <text bytes="158218" sha1="cst5x7tt58izti1pxzgljf27tx8qjcj" xml:space="preserve">
-== en == {| class="wikitable sortable" |- ! Word / Phrase !! Key !! Value !! Operator !! Plural |- | Zip Line || aerialway || zip_line || - || N |- | Zip Lines || aerialway || zip_line || - || Y |- | Zip Line in || aerialway || zip_line || in || N |- | Zip Lines in || aerialway || zip_line || in || Y |- | Zip Line near || aerialway || zip_line || near || N |- | Animal shelter || amenity || animal_shelter || - || N |- | Animal shelters || amenity || animal_shelter || - || Y |- | Animal shelter in || amenity || animal_shelter || in || N |- | Animal shelters in || amenity || animal_shelter || in || Y |- | Animal shelter near || amenity || animal_shelter || near|| N |- | Animal shelters near || amenity || animal_shelter || near|| Y |- | Drinking Water near || amenity || drinking_water || near || N |- | Water || amenity || drinking_water || - || N |- | Water in || amenity || drinking_water || in || N |- | Water near || amenity || drinking_water || near || N |- | Embassy || amenity || embassy || - || N |- | Embassys || amenity || embassy || - || Y |- | Embassies || amenity || embassy || - || Y |- |Coworkings near |amenity |coworking_space |near |Y |} [[Category:Word list]]
+== en == {| class="wikitable sortable" |- ! Word / Phrase !! Key !! Value !! Operator !! Plural |- | Zip Line || aerialway || zip_line || - || N |- | Zip Lines || aerialway || zip_line || - || Y |- | Zip Line in || aerialway || zip_line || in || N |- | Zip Lines in || aerialway || zip_line || in || Y |- | Zip Line near || aerialway || zip_line || near || N |- | Animal shelter || amenity || animal_shelter || - || N |- | Animal shelters || amenity || animal_shelter || - || Y |- | Animal shelter in || amenity || animal_shelter || in || N |- | Animal shelters in || amenity || animal_shelter || in || Y |- | Animal shelter near || amenity || animal_shelter || near|| N |- | Animal shelters near || amenity || animal_shelter || NEAR|| Y |- | Drinking Water near || amenity || drinking_water || near || N |- | Water || amenity || drinking_water || - || N |- | Water in || amenity || drinking_water || In || N |- | Water near || amenity || drinking_water || near || N |- | Embassy || amenity || embassy || - || N |- | Embassys || amenity || embassy || - || Y |- | Embassies || amenity || embassy || - || Y |- |Coworkings near |amenity |coworking_space |near |Y |} [[Category:Word list]]
 </text>
 <sha1>cst5x7tt58izti1pxzgljf27tx8qjcj</sha1>
 </revision>
 </page>
-</mediawiki>
+</mediawiki>