from nominatim.tokenizer.icu_rule_loader import ICURuleLoader
from nominatim.db import properties
+from mock_icu_word_table import MockIcuWordTable
+
+@pytest.fixture
+def word_table(temp_db_conn):
+ return MockIcuWordTable(temp_db_conn)
+
@pytest.fixture
def test_config(def_config, tmp_path):
sqldir.mkdir()
(sqldir / 'tokenizer').mkdir()
(sqldir / 'tokenizer' / 'legacy_icu_tokenizer.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'))
+ shutil.copy(str(def_config.lib_dir.sql / 'tokenizer' / 'icu_tokenizer_tables.sql'),
+ str(sqldir / 'tokenizer' / 'icu_tokenizer_tables.sql'))
def_config.lib_dir.sql = sqldir
monkeypatch.undo()
def _mk_analyser(norm=("[[:Punctuation:][:Space:]]+ > ' '",), trans=(':: upper()',),
- suffixes=('gasse', ), abbr=('street => st', )):
+ variants=('~gasse -> gasse', 'street => st', )):
cfgfile = tmp_path / 'analyser_test_config.yaml'
with cfgfile.open('w') as stream:
cfgstr = {'normalization' : list(norm),
'transliteration' : list(trans),
- 'compound_suffixes' : list(suffixes),
- 'abbreviations' : list(abbr)}
+ 'variants' : [ {'words': list(variants)}]}
yaml.dump(cfgstr, stream)
tok.naming_rules = ICUNameProcessorRules(loader=ICURuleLoader(cfgfile))
term_count INTEGER;
BEGIN
SELECT min(word_id) INTO full_token
- FROM word WHERE word = norm_term and class is null and country_code is null;
+ FROM word WHERE info->>'word' = norm_term and type = 'W';
IF full_token IS NULL THEN
full_token := nextval('seq_word');
- INSERT INTO word (word_id, word_token, word, search_name_count)
- SELECT full_token, ' ' || lookup_term, norm_term, 0 FROM unnest(lookup_terms) as lookup_term;
+ INSERT INTO word (word_id, word_token, type, info)
+ SELECT full_token, lookup_term, 'W',
+ json_build_object('word', norm_term, 'count', 0)
+ FROM unnest(lookup_terms) as lookup_term;
END IF;
FOR term IN SELECT unnest(string_to_array(unnest(lookup_terms), ' ')) LOOP
partial_tokens := '{}'::INT[];
FOR term IN SELECT unnest(partial_terms) LOOP
- SELECT min(word_id), max(search_name_count) INTO term_id, term_count
- FROM word WHERE word_token = term and class is null and country_code is null;
+ SELECT min(word_id), max(info->>'count') INTO term_id, term_count
+ FROM word WHERE word_token = term and type = 'w';
IF term_id IS NULL THEN
term_id := nextval('seq_word');
term_count := 0;
- INSERT INTO word (word_id, word_token, search_name_count)
- VALUES (term_id, term, 0);
+ INSERT INTO word (word_id, word_token, type, info)
+ VALUES (term_id, term, 'w', json_build_object('count', term_count));
END IF;
IF NOT (ARRAY[term_id] <@ partial_tokens) THEN
- partial_tokens := partial_tokens || term_id;
+ partial_tokens := partial_tokens || term_id;
END IF;
END LOOP;
END;
assert db_prop(legacy_icu_tokenizer.DBCFG_MAXWORDFREQ) is not None
+def test_init_word_table(tokenizer_factory, test_config, place_row, word_table):
+ place_row(names={'name' : 'Test Area', 'ref' : '52'})
+ place_row(names={'name' : 'No Area'})
+ place_row(names={'name' : 'Holzstrasse'})
+
+ 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)}
+
+
def test_init_from_project(monkeypatch, test_config, tokenizer_factory):
monkeypatch.setenv('NOMINATIM_TERM_NORMALIZATION', ':: lower();')
monkeypatch.setenv('NOMINATIM_MAX_WORD_FREQUENCY', '90300')
assert test_content == set((('1133', ), ))
-def test_make_standard_hnr(analyzer):
- with analyzer(abbr=('IV => 4',)) as anl:
- assert anl._make_standard_hnr('345') == '345'
- assert anl._make_standard_hnr('iv') == 'IV'
+def test_normalize_postcode(analyzer):
+ with analyzer() as anl:
+ anl.normalize_postcode('123') == '123'
+ anl.normalize_postcode('ab-34 ') == 'AB-34'
+ anl.normalize_postcode('38 Б') == '38 Б'
def test_update_postcodes_from_db_empty(analyzer, table_factory, word_table):
def test_update_special_phrase_empty_table(analyzer, word_table):
with analyzer() as anl:
anl.update_special_phrases([
- ("König bei", "amenity", "royal", "near"),
- ("Könige", "amenity", "royal", "-"),
+ ("König bei", "amenity", "royal", "near"),
+ ("Könige ", "amenity", "royal", "-"),
("street", "highway", "primary", "in")
], True)
assert word_table.get_special() \
- == {(' KÖNIG BEI', 'könig bei', 'amenity', 'royal', 'near'),
- (' KÖNIGE', 'könige', 'amenity', 'royal', None),
- (' STREET', 'street', 'highway', 'primary', 'in')}
+ == {('KÖNIG BEI', 'König bei', 'amenity', 'royal', 'near'),
+ ('KÖNIGE', 'Könige', 'amenity', 'royal', None),
+ ('STREET', 'street', 'highway', 'primary', 'in')}
def test_update_special_phrase_delete_all(analyzer, word_table):
- word_table.add_special(' FOO', 'foo', 'amenity', 'prison', 'in')
- word_table.add_special(' BAR', 'bar', 'highway', 'road', None)
+ word_table.add_special('FOO', 'foo', 'amenity', 'prison', 'in')
+ word_table.add_special('BAR', 'bar', 'highway', 'road', None)
assert word_table.count_special() == 2
def test_update_special_phrases_no_replace(analyzer, word_table):
- word_table.add_special(' FOO', 'foo', 'amenity', 'prison', 'in')
- word_table.add_special(' BAR', 'bar', 'highway', 'road', None)
+ word_table.add_special('FOO', 'foo', 'amenity', 'prison', 'in')
+ word_table.add_special('BAR', 'bar', 'highway', 'road', None)
assert word_table.count_special() == 2
def test_update_special_phrase_modify(analyzer, word_table):
- word_table.add_special(' FOO', 'foo', 'amenity', 'prison', 'in')
- word_table.add_special(' BAR', 'bar', 'highway', 'road', None)
+ word_table.add_special('FOO', 'foo', 'amenity', 'prison', 'in')
+ word_table.add_special('BAR', 'bar', 'highway', 'road', None)
assert word_table.count_special() == 2
], True)
assert word_table.get_special() \
- == {(' PRISON', 'prison', 'amenity', 'prison', 'in'),
- (' BAR', 'bar', 'highway', 'road', None),
- (' GARDEN', 'garden', 'leisure', 'garden', 'near')}
+ == {('PRISON', 'prison', 'amenity', 'prison', 'in'),
+ ('BAR', 'bar', 'highway', 'road', None),
+ ('GARDEN', 'garden', 'leisure', 'garden', 'near')}
+
+
+def test_add_country_names_new(analyzer, word_table):
+ with analyzer() as anl:
+ anl.add_country_names('es', {'name': 'Espagña', 'name:en': 'Spain'})
+
+ assert word_table.get_country() == {('es', 'ESPAGÑA'), ('es', 'SPAIN')}
+
+
+def test_add_country_names_extend(analyzer, word_table):
+ word_table.add_country('ch', 'SCHWEIZ')
+
+ with analyzer() as anl:
+ anl.add_country_names('ch', {'name': 'Schweiz', 'name:fr': 'Suisse'})
+
+ assert word_table.get_country() == {('ch', 'SCHWEIZ'), ('ch', 'SUISSE')}
class TestPlaceNames:
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)
def test_simple_names(self):
- info = self.analyzer.process_place({'name' : {'name' : 'Soft bAr', 'ref': '34'}})
+ info = self.analyzer.process_place({'name': {'name': 'Soft bAr', 'ref': '34'}})
- self.expect_name_terms(info, '#Soft bAr', '#34','Soft', 'bAr', '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.analyzer.process_place({'name': {'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.analyzer.process_place({'name': {'name': 'Houseboat (left)'}})
self.expect_name_terms(info, '#Houseboat (left)', '#Houseboat',
'houseboat', 'left')
-@pytest.mark.parametrize('pcode', ['12345', 'AB 123', '34-345'])
-def test_process_place_postcode(analyzer, word_table, pcode):
- with analyzer() as anl:
- anl.process_place({'address': {'postcode' : pcode}})
+ def test_country_name(self, word_table):
+ info = self.analyzer.process_place({'name': {'name': 'Norge'},
+ 'country_feature': 'no'})
- assert word_table.get_postcodes() == {pcode, }
+ self.expect_name_terms(info, '#norge', 'norge')
+ assert word_table.get_country() == {('no', 'NORGE')}
-@pytest.mark.parametrize('pcode', ['12:23', 'ab;cd;f', '123;836'])
-def test_process_place_bad_postcode(analyzer, word_table, pcode):
- with analyzer() as anl:
- anl.process_place({'address': {'postcode' : pcode}})
+class TestPlaceAddress:
- assert not word_table.get_postcodes()
+ @pytest.fixture(autouse=True)
+ def setup(self, analyzer, getorcreate_full_word):
+ with analyzer(trans=(":: upper()", "'🜵' > ' '")) as anl:
+ self.analyzer = anl
+ yield anl
-@pytest.mark.parametrize('hnr', ['123a', '1', '101'])
-def test_process_place_housenumbers_simple(analyzer, hnr, getorcreate_hnr_id):
- with analyzer() as anl:
- info = anl.process_place({'address': {'housenumber' : hnr}})
+ def process_address(self, **kwargs):
+ return self.analyzer.process_place({'address': kwargs})
- assert info['hnr'] == hnr.upper()
- assert info['hnr_tokens'] == "{-1}"
+ def name_token_set(self, *expected_terms):
+ tokens = self.analyzer.get_word_token_info(expected_terms)
+ for token in tokens:
+ assert token[2] is not None, "No token for {0}".format(token)
-def test_process_place_housenumbers_lists(analyzer, getorcreate_hnr_id):
- with analyzer() as anl:
- info = anl.process_place({'address': {'conscriptionnumber' : '1; 2;3'}})
+ return set((t[2] for t in tokens))
- assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
- assert info['hnr_tokens'] == "{-1,-2,-3}"
+ @pytest.mark.parametrize('pcode', ['12345', 'AB 123', '34-345'])
+ def test_process_place_postcode(self, word_table, pcode):
+ self.process_address(postcode=pcode)
-def test_process_place_housenumbers_duplicates(analyzer, getorcreate_hnr_id):
- with analyzer() as anl:
- info = anl.process_place({'address': {'housenumber' : '134',
- 'conscriptionnumber' : '134',
- 'streetnumber' : '99a'}})
+ assert word_table.get_postcodes() == {pcode, }
+
+
+ @pytest.mark.parametrize('pcode', ['12:23', 'ab;cd;f', '123;836'])
+ def test_process_place_bad_postcode(self, word_table, pcode):
+ self.process_address(postcode=pcode)
+
+ assert not word_table.get_postcodes()
+
+
+ @pytest.mark.parametrize('hnr', ['123a', '1', '101'])
+ def test_process_place_housenumbers_simple(self, hnr, getorcreate_hnr_id):
+ info = self.process_address(housenumber=hnr)
+
+ assert info['hnr'] == hnr.upper()
+ assert info['hnr_tokens'] == "{-1}"
+
+
+ def test_process_place_housenumbers_lists(self, getorcreate_hnr_id):
+ info = self.process_address(conscriptionnumber='1; 2;3')
+
+ assert set(info['hnr'].split(';')) == set(('1', '2', '3'))
+ assert info['hnr_tokens'] == "{-1,-2,-3}"
+
+
+ def test_process_place_housenumbers_duplicates(self, getorcreate_hnr_id):
+ info = self.process_address(housenumber='134',
+ conscriptionnumber='134',
+ streetnumber='99a')
+
+ assert set(info['hnr'].split(';')) == set(('134', '99A'))
+ assert info['hnr_tokens'] == "{-1,-2}"
+
+
+ def test_process_place_housenumbers_cached(self, getorcreate_hnr_id):
+ info = self.process_address(housenumber="45")
+ assert info['hnr_tokens'] == "{-1}"
+
+ info = self.process_address(housenumber="46")
+ assert info['hnr_tokens'] == "{-2}"
+
+ info = self.process_address(housenumber="41;45")
+ assert eval(info['hnr_tokens']) == {-1, -3}
+
+ info = self.process_address(housenumber="41")
+ assert eval(info['hnr_tokens']) == {-3}
+
+
+ def test_process_place_street(self):
+ info = self.process_address(street='Grand Road')
+
+ assert eval(info['street']) == self.name_token_set('#GRAND ROAD')
+
+
+ def test_process_place_street_empty(self):
+ info = self.process_address(street='🜵')
+
+ assert 'street' not in info
+
+
+ def test_process_place_place(self):
+ info = self.process_address(place='Honu Lulu')
+
+ assert eval(info['place_search']) == self.name_token_set('#HONU LULU',
+ 'HONU', 'LULU')
+ assert eval(info['place_match']) == self.name_token_set('#HONU LULU')
+
+
+ def test_process_place_place_empty(self):
+ info = self.process_address(place='🜵')
+
+ assert 'place_search' not in info
+ assert 'place_match' not in info
+
+
+ def test_process_place_address_terms(self):
+ info = self.process_address(country='de', city='Zwickau', state='Sachsen',
+ suburb='Zwickau', street='Hauptstr',
+ full='right behind the church')
+
+ city_full = self.name_token_set('#ZWICKAU')
+ city_all = self.name_token_set('#ZWICKAU', 'ZWICKAU')
+ state_full = self.name_token_set('#SACHSEN')
+ state_all = self.name_token_set('#SACHSEN', 'SACHSEN')
+
+ result = {k: [eval(v[0]), eval(v[1])] for k,v in info['addr'].items()}
+
+ assert result == {'city': [city_all, city_full],
+ 'suburb': [city_all, city_full],
+ 'state': [state_all, state_full]}
+
+
+ def test_process_place_address_terms_empty(self):
+ info = self.process_address(country='de', city=' ', street='Hauptstr',
+ full='right behind the church')
+
+ assert 'addr' not in info
- assert set(info['hnr'].split(';')) == set(('134', '99A'))
- assert info['hnr_tokens'] == "{-1,-2}"
projects / nominatim.git / blobdiff