2 Test for legacy tokenizer.
8 from nominatim.tokenizer import legacy_tokenizer
9 from nominatim.db import properties
10 from nominatim.errors import UsageError
13 def test_config(def_config, tmp_path):
14 def_config.project_dir = tmp_path / 'project'
15 def_config.project_dir.mkdir()
17 module_dir = tmp_path / 'module_src'
19 (module_dir / 'nominatim.so').write_text('TEST nomiantim.so')
21 def_config.lib_dir.module = module_dir
23 sqldir = tmp_path / 'sql'
25 (sqldir / 'tokenizer').mkdir()
26 (sqldir / 'tokenizer' / 'legacy_tokenizer.sql').write_text("SELECT 'a'")
27 (sqldir / 'words.sql').write_text("SELECT 'a'")
28 shutil.copy(str(def_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer_tables.sql'),
29 str(sqldir / 'tokenizer' / 'legacy_tokenizer_tables.sql'))
31 def_config.lib_dir.sql = sqldir
32 def_config.lib_dir.data = sqldir
38 def tokenizer_factory(dsn, tmp_path, property_table):
39 (tmp_path / 'tokenizer').mkdir()
42 return legacy_tokenizer.create(dsn, tmp_path / 'tokenizer')
48 def tokenizer_setup(tokenizer_factory, test_config, monkeypatch, sql_preprocessor):
49 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
50 tok = tokenizer_factory()
51 tok.init_new_db(test_config)
55 def analyzer(tokenizer_factory, test_config, monkeypatch, sql_preprocessor,
56 word_table, temp_db_with_extensions, tmp_path):
57 sql = tmp_path / 'sql' / 'tokenizer' / 'legacy_tokenizer.sql'
59 CREATE OR REPLACE FUNCTION getorcreate_housenumber_id(lookup_word TEXT)
60 RETURNS INTEGER AS $$ SELECT 342; $$ LANGUAGE SQL;
63 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
64 monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
65 tok = tokenizer_factory()
66 tok.init_new_db(test_config)
69 with tok.name_analyzer() as analyzer:
74 def make_standard_name(temp_db_cursor):
75 temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT)
76 RETURNS TEXT AS $$ SELECT '#' || lower(name) || '#'; $$ LANGUAGE SQL""")
80 def create_postcode_id(temp_db_cursor):
81 temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_postcode_id(postcode TEXT)
83 INSERT INTO word (word_token, word, class, type)
84 VALUES (' ' || postcode, postcode, 'place', 'postcode')
90 def make_keywords(temp_db_cursor, temp_db_with_extensions):
91 temp_db_cursor.execute(
92 """CREATE OR REPLACE FUNCTION make_keywords(names HSTORE)
93 RETURNS INTEGER[] AS $$ SELECT ARRAY[1, 2, 3] $$ LANGUAGE SQL""")
95 def test_init_new(tokenizer_factory, test_config, monkeypatch,
96 temp_db_conn, sql_preprocessor):
97 monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', 'xxvv')
98 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
100 tok = tokenizer_factory()
101 tok.init_new_db(test_config)
103 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) == 'xxvv'
105 outfile = test_config.project_dir / 'module' / 'nominatim.so'
107 assert outfile.exists()
108 assert outfile.read_text() == 'TEST nomiantim.so'
109 assert outfile.stat().st_mode == 33261
112 def test_init_module_load_failed(tokenizer_factory, test_config):
113 tok = tokenizer_factory()
115 with pytest.raises(UsageError):
116 tok.init_new_db(test_config)
119 def test_init_module_custom(tokenizer_factory, test_config,
120 monkeypatch, tmp_path, sql_preprocessor):
121 module_dir = (tmp_path / 'custom').resolve()
123 (module_dir/ 'nominatim.so').write_text('CUSTOM nomiantim.so')
125 monkeypatch.setenv('NOMINATIM_DATABASE_MODULE_PATH', str(module_dir))
126 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
128 tok = tokenizer_factory()
129 tok.init_new_db(test_config)
131 assert not (test_config.project_dir / 'module').exists()
134 def test_init_from_project(tokenizer_setup, tokenizer_factory):
135 tok = tokenizer_factory()
137 tok.init_from_project()
139 assert tok.normalization is not None
142 def test_update_sql_functions(sql_preprocessor, temp_db_conn,
143 tokenizer_factory, test_config, table_factory,
144 monkeypatch, temp_db_cursor):
145 monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '1133')
146 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
147 tok = tokenizer_factory()
148 tok.init_new_db(test_config)
151 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) == '1133'
153 table_factory('test', 'txt TEXT')
155 func_file = test_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer.sql'
156 func_file.write_text("""INSERT INTO test VALUES ('{{max_word_freq}}'),
157 ('{{modulepath}}')""")
159 tok.update_sql_functions(test_config)
161 test_content = temp_db_cursor.row_set('SELECT * FROM test')
162 assert test_content == set((('1133', ), (str(test_config.project_dir / 'module'), )))
165 def test_migrate_database(tokenizer_factory, test_config, temp_db_conn, monkeypatch):
166 monkeypatch.setattr(legacy_tokenizer, '_check_module', lambda m, c: None)
167 tok = tokenizer_factory()
168 tok.migrate_database(test_config)
170 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) is not None
171 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) is not None
173 outfile = test_config.project_dir / 'module' / 'nominatim.so'
175 assert outfile.exists()
176 assert outfile.read_text() == 'TEST nomiantim.so'
177 assert outfile.stat().st_mode == 33261
180 def test_normalize(analyzer):
181 assert analyzer.normalize('TEsT') == 'test'
184 def test_update_postcodes_from_db_empty(analyzer, table_factory, word_table,
186 table_factory('location_postcode', 'postcode TEXT',
187 content=(('1234',), ('12 34',), ('AB23',), ('1234',)))
189 analyzer.update_postcodes_from_db()
191 assert word_table.count() == 3
192 assert word_table.get_postcodes() == {'1234', '12 34', 'AB23'}
195 def test_update_postcodes_from_db_add_and_remove(analyzer, table_factory, word_table,
197 table_factory('location_postcode', 'postcode TEXT',
198 content=(('1234',), ('45BC', ), ('XX45', )))
199 word_table.add_postcode(' 1234', '1234')
200 word_table.add_postcode(' 5678', '5678')
202 analyzer.update_postcodes_from_db()
204 assert word_table.count() == 3
205 assert word_table.get_postcodes() == {'1234', '45BC', 'XX45'}
208 def test_update_special_phrase_empty_table(analyzer, word_table, make_standard_name):
209 analyzer.update_special_phrases([
210 ("König bei", "amenity", "royal", "near"),
211 ("Könige", "amenity", "royal", "-"),
212 ("könige", "amenity", "royal", "-"),
213 ("strasse", "highway", "primary", "in")
216 assert word_table.get_special() \
217 == set(((' #könig bei#', 'könig bei', 'amenity', 'royal', 'near'),
218 (' #könige#', 'könige', 'amenity', 'royal', None),
219 (' #strasse#', 'strasse', 'highway', 'primary', 'in')))
222 def test_update_special_phrase_delete_all(analyzer, word_table, make_standard_name):
223 word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
224 word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
226 assert word_table.count_special() == 2
228 analyzer.update_special_phrases([], True)
230 assert word_table.count_special() == 0
233 def test_update_special_phrases_no_replace(analyzer, word_table, make_standard_name):
234 word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
235 word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
237 assert word_table.count_special() == 2
239 analyzer.update_special_phrases([], False)
241 assert word_table.count_special() == 2
244 def test_update_special_phrase_modify(analyzer, word_table, make_standard_name):
245 word_table.add_special(' #foo#', 'foo', 'amenity', 'prison', 'in')
246 word_table.add_special(' #bar#', 'bar', 'highway', 'road', None)
248 assert word_table.count_special() == 2
250 analyzer.update_special_phrases([
251 ('prison', 'amenity', 'prison', 'in'),
252 ('bar', 'highway', 'road', '-'),
253 ('garden', 'leisure', 'garden', 'near')
256 assert word_table.get_special() \
257 == set(((' #prison#', 'prison', 'amenity', 'prison', 'in'),
258 (' #bar#', 'bar', 'highway', 'road', None),
259 (' #garden#', 'garden', 'leisure', 'garden', 'near')))
262 def test_add_country_names(analyzer, word_table, make_standard_name):
263 analyzer.add_country_names('de', ['Germany', 'Deutschland', 'germany'])
265 assert word_table.get_country() \
266 == {('de', ' #germany#'),
267 ('de', ' #deutschland#')}
270 def test_add_more_country_names(analyzer, word_table, make_standard_name):
271 word_table.add_country('fr', ' #france#')
272 word_table.add_country('it', ' #italy#')
273 word_table.add_country('it', ' #itala#')
275 analyzer.add_country_names('it', ['Italy', 'IT'])
277 assert word_table.get_country() \
278 == {('fr', ' #france#'),
284 def test_process_place_names(analyzer, make_keywords):
285 info = analyzer.process_place({'name' : {'name' : 'Soft bAr', 'ref': '34'}})
287 assert info['names'] == '{1,2,3}'
290 @pytest.mark.parametrize('pcode', ['12345', 'AB 123', '34-345'])
291 def test_process_place_postcode(analyzer, create_postcode_id, word_table, pcode):
292 analyzer.process_place({'address': {'postcode' : pcode}})
294 assert word_table.get_postcodes() == {pcode, }
297 @pytest.mark.parametrize('pcode', ['12:23', 'ab;cd;f', '123;836'])
298 def test_process_place_bad_postcode(analyzer, create_postcode_id, word_table, pcode):
299 analyzer.process_place({'address': {'postcode' : pcode}})
301 assert not word_table.get_postcodes()
304 class TestHousenumberName:
307 @pytest.fixture(autouse=True)
308 def setup_create_housenumbers(temp_db_cursor):
309 temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_housenumbers(
311 OUT tokens TEXT, OUT normtext TEXT)
313 SELECT housenumbers::TEXT, array_to_string(housenumbers, ';')
318 @pytest.mark.parametrize('hnr', ['123a', '1', '101'])
319 def test_process_place_housenumbers_simple(analyzer, hnr):
320 info = analyzer.process_place({'address': {'housenumber' : hnr}})
322 assert info['hnr'] == hnr
323 assert info['hnr_tokens'].startswith("{")
327 def test_process_place_housenumbers_lists(analyzer):
328 info = analyzer.process_place({'address': {'conscriptionnumber' : '1; 2;3'}})
330 assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
334 def test_process_place_housenumbers_duplicates(analyzer):
335 info = analyzer.process_place({'address': {'housenumber' : '134',
336 'conscriptionnumber' : '134',
337 'streetnumber' : '99a'}})
339 assert set(info['hnr'].split(';')) == set(('134', '99a'))