Merge remote-tracking branch 'upstream/master'

[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 ed079269561ff8ba8a19b56ecaff840ec0bd0c93..5dbe292e38bcd73c869275ed225794adfaa2e76e 100644 (file)
--- a/test/python/test_tokenizer_icu.py
+++ b/test/python/test_tokenizer_icu.py
@@ -7,10 +7,10 @@ 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
+import nominatim.tokenizer.icu_rule_loader
 from nominatim.db import properties
 from nominatim.db.sql_preprocessor import SQLPreprocessor
+from nominatim.indexer.place_info import PlaceInfo
 
 from mock_icu_word_table import MockIcuWordTable
 
@@ -67,11 +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 = nominatim.tokenizer.icu_rule_loader.ICURuleLoader(test_config)
 
         return tok.name_analyzer()
 
@@ -147,16 +151,15 @@ def getorcreate_hnr_id(temp_db_cursor):
                                 SELECT -nextval('seq_word')::INTEGER; $$ LANGUAGE SQL""")
 
 
-def test_init_new(tokenizer_factory, test_config, monkeypatch, db_prop):
-    monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
-
+def test_init_new(tokenizer_factory, test_config, db_prop):
     tok = tokenizer_factory()
     tok.init_new_db(test_config)
 
-    assert db_prop(icu_tokenizer.DBCFG_TERM_NORMALIZATION) == ':: lower();'
+    assert db_prop(nominatim.tokenizer.icu_rule_loader.DBCFG_IMPORT_NORM_RULES) \
+            .startswith(':: lower ();')
 
 
-def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
+def test_init_word_table(tokenizer_factory, test_config, place_row, temp_db_cursor):
     place_row(names={'name' : 'Test Area', 'ref' : '52'})
     place_row(names={'name' : 'No Area'})
     place_row(names={'name' : 'Holzstrasse'})
@@ -164,23 +167,17 @@ def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
     tok = tokenizer_factory()
     tok.init_new_db(test_config)
 
-    assert word_table.get_partial_words() == {('test', 1),
-                                              ('no', 1), ('area', 2),
-                                              ('holz', 1), ('strasse', 1),
-                                              ('str', 1)}
+    assert temp_db_cursor.table_exists('word')
 
 
-def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
-    monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
+def test_init_from_project(test_config, tokenizer_factory):
     tok = tokenizer_factory()
     tok.init_new_db(test_config)
-    monkeypatch.undo()
 
     tok = tokenizer_factory()
-    tok.init_from_project()
+    tok.init_from_project(test_config)
 
-    assert tok.naming_rules is not None
-    assert tok.term_normalization == ':: lower();'
+    assert tok.loader is not None
 
 
 def test_update_sql_functions(db_prop, temp_db_cursor,
@@ -308,44 +305,54 @@ 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):
+        return self.analyzer.process_place(PlaceInfo({'name': names}))
+
+
     def test_simple_names(self):
-        info = self.analyzer.process_place({'name': {'name': 'Soft bAr', 'ref': '34'}})
+        info = self.process_named_place({'name': 'Soft bAr', 'ref': '34'})
 
         self.expect_name_terms(info, '#Soft bAr', '#34', 'Soft', 'bAr', '34')
 
 
     @pytest.mark.parametrize('sep', [',' , ';'])
     def test_names_with_separator(self, sep):
-        info = self.analyzer.process_place({'name': {'name': sep.join(('New York', 'Big Apple'))}})
+        info = self.process_named_place({'name': sep.join(('New York', 'Big Apple'))})
 
         self.expect_name_terms(info, '#New York', '#Big Apple',
                                      'new', 'york', 'big', 'apple')
 
 
     def test_full_names_with_bracket(self):
-        info = self.analyzer.process_place({'name': {'name': 'Houseboat (left)'}})
+        info = self.process_named_place({'name': 'Houseboat (left)'})
 
         self.expect_name_terms(info, '#Houseboat (left)', '#Houseboat',
                                      'houseboat', 'left')
 
 
     def test_country_name(self, word_table):
-        info = self.analyzer.process_place({'name': {'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')}
@@ -361,7 +368,7 @@ class TestPlaceAddress:
 
 
     def process_address(self, **kwargs):
-        return self.analyzer.process_place({'address': kwargs})
+        return self.analyzer.process_place(PlaceInfo({'address': kwargs}))
 
 
     def name_token_set(self, *expected_terms):