reorganize and complete tests around generic token analysis

[nominatim.git] / test / python / test_tokenizer_icu.py

diff --git a/test/python/test_tokenizer_icu.py b/test/python/test_tokenizer_icu.py
index 28c6ef7abb48e0c7192cb1cf3039a14e43259229..6a2f2f8bd04b405f7741aca6fbaa27cbe2ce113c 100644 (file)
--- a/test/python/test_tokenizer_icu.py
+++ b/test/python/test_tokenizer_icu.py
@@ -7,7 +7,6 @@ import yaml
 import pytest
 
 from nominatim.tokenizer import icu_tokenizer
-from nominatim.tokenizer.icu_name_processor import ICUNameProcessorRules
 from nominatim.tokenizer.icu_rule_loader import ICURuleLoader
 from nominatim.db import properties
 from nominatim.db.sql_preprocessor import SQLPreprocessor
@@ -68,11 +67,15 @@ def analyzer(tokenizer_factory, test_config, monkeypatch,
     monkeypatch.undo()
 
     def _mk_analyser(norm=("[[:Punctuation:][:Space:]]+ > ' '",), trans=(':: upper()',),
-                     variants=('~gasse -> gasse', 'street => st', )):
-        cfgstr = {'normalization' : list(norm),
-                   'transliteration' : list(trans),
-                   'variants' : [ {'words': list(variants)}]}
-        tok.naming_rules = ICUNameProcessorRules(loader=ICURuleLoader(cfgstr))
+                     variants=('~gasse -> gasse', 'street => st', ),
+                     sanitizers=[]):
+        cfgstr = {'normalization': list(norm),
+                  'sanitizers': sanitizers,
+                  'transliteration': list(trans),
+                  'token-analysis': [{'analyzer': 'generic',
+                                      'variants': [{'words': list(variants)}]}]}
+        (test_config.project_dir / 'icu_tokenizer.yaml').write_text(yaml.dump(cfgstr))
+        tok.loader = ICURuleLoader(test_config)
 
         return tok.name_analyzer()
 
@@ -166,9 +169,7 @@ def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
     tok.init_new_db(test_config)
 
     assert word_table.get_partial_words() == {('test', 1),
-                                              ('no', 1), ('area', 2),
-                                              ('holz', 1), ('strasse', 1),
-                                              ('str', 1)}
+                                              ('no', 1), ('area', 2)}
 
 
 def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
@@ -178,9 +179,9 @@ def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
     monkeypatch.undo()
 
     tok = tokenizer_factory()
-    tok.init_from_project()
+    tok.init_from_project(test_config)
 
-    assert tok.naming_rules is not None
+    assert tok.loader is not None
     assert tok.term_normalization == ':: lower();'
 
 
@@ -309,26 +310,23 @@ class TestPlaceNames:
 
     @pytest.fixture(autouse=True)
     def setup(self, analyzer, sql_functions):
-        with analyzer() as anl:
+        sanitizers = [{'step': 'split-name-list'},
+                      {'step': 'strip-brace-terms'}]
+        with analyzer(sanitizers=sanitizers) as anl:
             self.analyzer = anl
             yield anl
 
 
     def expect_name_terms(self, info, *expected_terms):
         tokens = self.analyzer.get_word_token_info(expected_terms)
-        print (tokens)
         for token in tokens:
             assert token[2] is not None, "No token for {0}".format(token)
 
         assert eval(info['names']) == set((t[2] for t in tokens))
 
 
-    def process_named_place(self, names, country_feature=None):
-        place = {'name': names}
-        if country_feature:
-            place['country_feature'] = country_feature
-
-        return self.analyzer.process_place(PlaceInfo(place))
+    def process_named_place(self, names):
+        return self.analyzer.process_place(PlaceInfo({'name': names}))
 
 
     def test_simple_names(self):
@@ -353,7 +351,13 @@ class TestPlaceNames:
 
 
     def test_country_name(self, word_table):
-        info = self.process_named_place({'name': 'Norge'}, country_feature='no')
+        place = PlaceInfo({'name' : {'name': 'Norge'},
+                           'country_code': 'no',
+                           'rank_address': 4,
+                           'class': 'boundary',
+                           'type': 'administrative'})
+
+        info = self.analyzer.process_place(place)
 
         self.expect_name_terms(info, '#norge', 'norge')
         assert word_table.get_country() == {('no', 'NORGE')}