X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/296a66558ffb2ee2c43e653b00ff24097ee3525c..388ebcbae2b895279727edafebfb8ed794f0c4cc:/test/python/test_tokenizer_legacy.py diff --git a/test/python/test_tokenizer_legacy.py b/test/python/test_tokenizer_legacy.py index 44937904..58bf5a5e 100644 --- a/test/python/test_tokenizer_legacy.py +++ b/test/python/test_tokenizer_legacy.py @@ -1,6 +1,8 @@ """ Test for legacy tokenizer. """ +import shutil + import pytest from nominatim.tokenizer import legacy_tokenizer @@ -18,11 +20,23 @@ def test_config(def_config, tmp_path): def_config.lib_dir.module = module_dir + sqldir = tmp_path / 'sql' + sqldir.mkdir() + (sqldir / 'tokenizer').mkdir() + (sqldir / 'tokenizer' / 'legacy_tokenizer.sql').write_text("SELECT 'a'") + (sqldir / 'words.sql').write_text("SELECT 'a'") + shutil.copy(str(def_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer_tables.sql'), + str(sqldir / 'tokenizer' / 'legacy_tokenizer_tables.sql')) + + def_config.lib_dir.sql = sqldir + def_config.lib_dir.data = sqldir + return def_config @pytest.fixture -def tokenizer_factory(dsn, tmp_path, monkeypatch): +def tokenizer_factory(dsn, tmp_path, monkeypatch, property_table): + (tmp_path / 'tokenizer').mkdir() def _maker(): return legacy_tokenizer.create(dsn, tmp_path / 'tokenizer') @@ -30,13 +44,65 @@ def tokenizer_factory(dsn, tmp_path, monkeypatch): return _maker @pytest.fixture -def tokenizer_setup(tokenizer_factory, test_config, property_table, monkeypatch): +def tokenizer_setup(tokenizer_factory, test_config, monkeypatch, sql_preprocessor): monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None) tok = tokenizer_factory() tok.init_new_db(test_config) -def test_init_new(tokenizer_factory, test_config, property_table, monkeypatch, temp_db_conn): +@pytest.fixture +def analyzer(tokenizer_factory, test_config, monkeypatch, sql_preprocessor, + word_table, temp_db_with_extensions, tmp_path): + sql = tmp_path / 'sql' / 'tokenizer' / 'legacy_tokenizer.sql' + sql.write_text(""" + CREATE OR REPLACE FUNCTION getorcreate_housenumber_id(lookup_word TEXT) + RETURNS INTEGER AS $$ SELECT 342; $$ LANGUAGE SQL; + """) + + monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None) + monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();') + tok = tokenizer_factory() + tok.init_new_db(test_config) + monkeypatch.undo() + + with tok.name_analyzer() as analyzer: + yield analyzer + + +@pytest.fixture +def make_standard_name(temp_db_cursor): + temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION make_standard_name(name TEXT) + RETURNS TEXT AS $$ SELECT ' ' || name; $$ LANGUAGE SQL""") + + +@pytest.fixture +def create_postcode_id(table_factory, temp_db_cursor): + table_factory('out_postcode_table', 'postcode TEXT') + + temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_postcode_id(postcode TEXT) + RETURNS BOOLEAN AS $$ + INSERT INTO out_postcode_table VALUES (postcode) RETURNING True; + $$ LANGUAGE SQL""") + + +@pytest.fixture +def create_housenumbers(temp_db_cursor): + temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_housenumbers( + housenumbers TEXT[], + OUT tokens TEXT, OUT normtext TEXT) + AS $$ + SELECT housenumbers::TEXT, array_to_string(housenumbers, ';') + $$ LANGUAGE SQL""") + + +@pytest.fixture +def make_keywords(temp_db_cursor, temp_db_with_extensions): + temp_db_cursor.execute( + """CREATE OR REPLACE FUNCTION make_keywords(names HSTORE) + RETURNS INTEGER[] AS $$ SELECT ARRAY[1, 2, 3] $$ LANGUAGE SQL""") + +def test_init_new(tokenizer_factory, test_config, monkeypatch, + temp_db_conn, sql_preprocessor): monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', 'xxvv') monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None) @@ -52,15 +118,16 @@ def test_init_new(tokenizer_factory, test_config, property_table, monkeypatch, t assert outfile.stat().st_mode == 33261 -def test_init_module_load_failed(tokenizer_factory, test_config, property_table, monkeypatch, temp_db_conn): +def test_init_module_load_failed(tokenizer_factory, test_config, + monkeypatch, temp_db_conn): tok = tokenizer_factory() with pytest.raises(UsageError): tok.init_new_db(test_config) -def test_init_module_custom(tokenizer_factory, test_config, property_table, - monkeypatch, tmp_path): +def test_init_module_custom(tokenizer_factory, test_config, + monkeypatch, tmp_path, sql_preprocessor): module_dir = (tmp_path / 'custom').resolve() module_dir.mkdir() (module_dir/ 'nominatim.so').write_text('CUSTOM nomiantim.so') @@ -80,3 +147,152 @@ def test_init_from_project(tokenizer_setup, tokenizer_factory): tok.init_from_project() assert tok.normalization is not None + + +def test_update_sql_functions(sql_preprocessor, temp_db_conn, + tokenizer_factory, test_config, table_factory, + monkeypatch, temp_db_cursor): + monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '1133') + monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None) + tok = tokenizer_factory() + tok.init_new_db(test_config) + monkeypatch.undo() + + assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) == '1133' + + table_factory('test', 'txt TEXT') + + func_file = test_config.lib_dir.sql / 'tokenizer' / 'legacy_tokenizer.sql' + func_file.write_text("""INSERT INTO test VALUES ('{{max_word_freq}}'), + ('{{modulepath}}')""") + + tok.update_sql_functions(test_config) + + test_content = temp_db_cursor.row_set('SELECT * FROM test') + assert test_content == set((('1133', ), (str(test_config.project_dir / 'module'), ))) + + +def test_migrate_database(tokenizer_factory, test_config, temp_db_conn, monkeypatch): + monkeypatch.setattr(legacy_tokenizer, '_check_module' , lambda m, c: None) + tok = tokenizer_factory() + tok.migrate_database(test_config) + + assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_MAXWORDFREQ) is not None + assert properties.get_property(temp_db_conn, legacy_tokenizer.DBCFG_NORMALIZATION) is not None + + outfile = test_config.project_dir / 'module' / 'nominatim.so' + + assert outfile.exists() + assert outfile.read_text() == 'TEST nomiantim.so' + assert outfile.stat().st_mode == 33261 + + +def test_normalize(analyzer): + assert analyzer.normalize('TEsT') == 'test' + + +def test_add_postcodes_from_db(analyzer, table_factory, temp_db_cursor, + create_postcode_id): + table_factory('location_postcode', 'postcode TEXT', + content=(('1234',), ('12 34',), ('AB23',), ('1234',))) + + analyzer.add_postcodes_from_db() + + assert temp_db_cursor.row_set("SELECT * from out_postcode_table") \ + == set((('1234', ), ('12 34', ), ('AB23',))) + + +def test_update_special_phrase_empty_table(analyzer, word_table, temp_db_cursor, + make_standard_name): + analyzer.update_special_phrases([ + ("König bei", "amenity", "royal", "near"), + ("Könige", "amenity", "royal", "-"), + ("strasse", "highway", "primary", "in") + ]) + + assert temp_db_cursor.row_set("""SELECT word_token, word, class, type, operator + FROM word WHERE class != 'place'""") \ + == set(((' könig bei', 'könig bei', 'amenity', 'royal', 'near'), + (' könige', 'könige', 'amenity', 'royal', None), + (' strasse', 'strasse', 'highway', 'primary', 'in'))) + + +def test_update_special_phrase_delete_all(analyzer, word_table, temp_db_cursor, + make_standard_name): + temp_db_cursor.execute("""INSERT INTO word (word_token, word, class, type, operator) + VALUES (' foo', 'foo', 'amenity', 'prison', 'in'), + (' bar', 'bar', 'highway', 'road', null)""") + + assert 2 == temp_db_cursor.scalar("SELECT count(*) FROM word WHERE class != 'place'""") + + analyzer.update_special_phrases([]) + + assert 0 == temp_db_cursor.scalar("SELECT count(*) FROM word WHERE class != 'place'""") + + +def test_update_special_phrase_modify(analyzer, word_table, temp_db_cursor, + make_standard_name): + temp_db_cursor.execute("""INSERT INTO word (word_token, word, class, type, operator) + VALUES (' foo', 'foo', 'amenity', 'prison', 'in'), + (' bar', 'bar', 'highway', 'road', null)""") + + assert 2 == temp_db_cursor.scalar("SELECT count(*) FROM word WHERE class != 'place'""") + + analyzer.update_special_phrases([ + ('prison', 'amenity', 'prison', 'in'), + ('bar', 'highway', 'road', '-'), + ('garden', 'leisure', 'garden', 'near') + ]) + + assert temp_db_cursor.row_set("""SELECT word_token, word, class, type, operator + FROM word WHERE class != 'place'""") \ + == set(((' prison', 'prison', 'amenity', 'prison', 'in'), + (' bar', 'bar', 'highway', 'road', None), + (' garden', 'garden', 'leisure', 'garden', 'near'))) + + +def test_process_place_names(analyzer, make_keywords): + + info = analyzer.process_place({'name' : {'name' : 'Soft bAr', 'ref': '34'}}) + + assert info['names'] == '{1,2,3}' + + +@pytest.mark.parametrize('pc', ['12345', 'AB 123', '34-345']) +def test_process_place_postcode(analyzer, temp_db_cursor, create_postcode_id, pc): + + info = analyzer.process_place({'address': {'postcode' : pc}}) + + assert temp_db_cursor.row_set("SELECT * from out_postcode_table") \ + == set(((pc, ),)) + + +@pytest.mark.parametrize('pc', ['12:23', 'ab;cd;f', '123;836']) +def test_process_place_bad_postcode(analyzer, temp_db_cursor, create_postcode_id, + pc): + + info = analyzer.process_place({'address': {'postcode' : pc}}) + + assert 0 == temp_db_cursor.scalar("SELECT count(*) from out_postcode_table") + + +@pytest.mark.parametrize('hnr', ['123a', '1', '101']) +def test_process_place_housenumbers_simple(analyzer, create_housenumbers, hnr): + info = analyzer.process_place({'address': {'housenumber' : hnr}}) + + assert info['hnr'] == hnr + assert info['hnr_tokens'].startswith("{") + + +def test_process_place_housenumbers_lists(analyzer, create_housenumbers): + info = analyzer.process_place({'address': {'conscriptionnumber' : '1; 2;3'}}) + + assert set(info['hnr'].split(';')) == set(('1', '2', '3')) + + +def test_process_place_housenumbers_duplicates(analyzer, create_housenumbers): + info = analyzer.process_place({'address': {'housenumber' : '134', + 'conscriptionnumber' : '134', + 'streetnumber' : '99a'}}) + + assert set(info['hnr'].split(';')) == set(('134', '99a'))