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 $$ SELECT ' ' || name; $$ 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 create_housenumbers(temp_db_cursor):
91 temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_housenumbers(
93 OUT tokens TEXT, OUT normtext TEXT)
95 SELECT housenumbers::TEXT, array_to_string(housenumbers, ';')
100 def make_keywords(temp_db_cursor, temp_db_with_extensions):
101 temp_db_cursor.execute(
102 """CREATE OR REPLACE FUNCTION make_keywords(names HSTORE)
103 RETURNS INTEGER[] AS $$ SELECT ARRAY[1, 2, 3] $$ LANGUAGE SQL""")
105 def test_init_new(tokenizer_factory, test_config, monkeypatch,
106 temp_db_conn, sql_preprocessor):
107 monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', 'xxvv')
108 monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None)
110 tok = tokenizer_factory()
111 tok.init_new_db(test_config)
113 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) == 'xxvv'
115 outfile = test_config.project_dir / 'module' / 'nominatim.so'
117 assert outfile.exists()
118 assert outfile.read_text() == 'TEST nomiantim.so'
119 assert outfile.stat().st_mode == 33261
122 def test_init_module_load_failed(tokenizer_factory, test_config,
123 monkeypatch, temp_db_conn):
124 tok = tokenizer_factory()
126 with pytest.raises(UsageError):
127 tok.init_new_db(test_config)
130 def test_init_module_custom(tokenizer_factory, test_config,
131 monkeypatch, tmp_path, sql_preprocessor):
132 module_dir = (tmp_path / 'custom').resolve()
134 (module_dir/ 'nominatim.so').write_text('CUSTOM nomiantim.so')
136 monkeypatch.setenv('NOMINATIM_DATABASE_MODULE_PATH', str(module_dir))
137 monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None)
139 tok = tokenizer_factory()
140 tok.init_new_db(test_config)
142 assert not (test_config.project_dir / 'module').exists()
145 def test_init_from_project(tokenizer_setup, tokenizer_factory):
146 tok = tokenizer_factory()
148 tok.init_from_project()
150 assert tok.normalization is not None
153 def test_update_sql_functions(sql_preprocessor, temp_db_conn,
154 tokenizer_factory, test_config, table_factory,
155 monkeypatch, temp_db_cursor):
156 monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '1133')
157 monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None)
158 tok = tokenizer_factory()
159 tok.init_new_db(test_config)
162 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) == '1133'
164 table_factory('test', 'txt TEXT')
166 func_file = test_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer.sql'
167 func_file.write_text("""INSERT INTO test VALUES ('{{max_word_freq}}'),
168 ('{{modulepath}}')""")
170 tok.update_sql_functions(test_config)
172 test_content = temp_db_cursor.row_set('SELECT * FROM test')
173 assert test_content == set((('1133', ), (str(test_config.project_dir / 'module'), )))
176 def test_migrate_database(tokenizer_factory, test_config, temp_db_conn, monkeypatch):
177 monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None)
178 tok = tokenizer_factory()
179 tok.migrate_database(test_config)
181 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) is not None
182 assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) is not None
184 outfile = test_config.project_dir / 'module' / 'nominatim.so'
186 assert outfile.exists()
187 assert outfile.read_text() == 'TEST nomiantim.so'
188 assert outfile.stat().st_mode == 33261
191 def test_normalize(analyzer):
192 assert analyzer.normalize('TEsT') == 'test'
195 def test_update_postcodes_from_db_empty(analyzer, table_factory, word_table,
197 table_factory('location_postcode', 'postcode TEXT',
198 content=(('1234',), ('12 34',), ('AB23',), ('1234',)))
200 analyzer.update_postcodes_from_db()
202 assert word_table.count() == 3
203 assert word_table.get_postcodes() == {'1234', '12 34', 'AB23'}
206 def test_update_postcodes_from_db_add_and_remove(analyzer, table_factory, word_table,
208 table_factory('location_postcode', 'postcode TEXT',
209 content=(('1234',), ('45BC', ), ('XX45', )))
210 word_table.add_postcode(' 1234', '1234')
211 word_table.add_postcode(' 5678', '5678')
213 analyzer.update_postcodes_from_db()
215 assert word_table.count() == 3
216 assert word_table.get_postcodes() == {'1234', '45BC', 'XX45'}
219 def test_update_special_phrase_empty_table(analyzer, word_table, make_standard_name):
220 analyzer.update_special_phrases([
221 ("König bei", "amenity", "royal", "near"),
222 ("Könige", "amenity", "royal", "-"),
223 ("strasse", "highway", "primary", "in")
226 assert word_table.get_special() \
227 == set(((' könig bei', 'könig bei', 'amenity', 'royal', 'near'),
228 (' könige', 'könige', 'amenity', 'royal', None),
229 (' strasse', 'strasse', 'highway', 'primary', 'in')))
232 def test_update_special_phrase_delete_all(analyzer, word_table, temp_db_cursor,
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([])
241 assert word_table.count_special() == 0
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_process_place_names(analyzer, make_keywords):
264 info = analyzer.process_place({'name' : {'name' : 'Soft bAr', 'ref': '34'}})
266 assert info['names'] == '{1,2,3}'
269 @pytest.mark.parametrize('pc', ['12345', 'AB 123', '34-345'])
270 def test_process_place_postcode(analyzer, create_postcode_id, word_table, pc):
271 info = analyzer.process_place({'address': {'postcode' : pc}})
273 assert word_table.get_postcodes() == {pc, }
276 @pytest.mark.parametrize('pc', ['12:23', 'ab;cd;f', '123;836'])
277 def test_process_place_bad_postcode(analyzer, create_postcode_id, word_table, pc):
278 info = analyzer.process_place({'address': {'postcode' : pc}})
280 assert not word_table.get_postcodes()
283 @pytest.mark.parametrize('hnr', ['123a', '1', '101'])
284 def test_process_place_housenumbers_simple(analyzer, create_housenumbers, hnr):
285 info = analyzer.process_place({'address': {'housenumber' : hnr}})
287 assert info['hnr'] == hnr
288 assert info['hnr_tokens'].startswith("{")
291 def test_process_place_housenumbers_lists(analyzer, create_housenumbers):
292 info = analyzer.process_place({'address': {'conscriptionnumber' : '1; 2;3'}})
294 assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
297 def test_process_place_housenumbers_duplicates(analyzer, create_housenumbers):
298 info = analyzer.process_place({'address': {'housenumber' : '134',
299 'conscriptionnumber' : '134',
300 'streetnumber' : '99a'}})
302 assert set(info['hnr'].split(';')) == set(('134', '99a'))