X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/ab6f35d83a350ead4eb12b59c4cbc050e48f93ee..28ab2f6048eff33e6119271c9fd31852db64240a:/test/python/tokenizer/test_icu.py diff --git a/test/python/tokenizer/test_icu.py b/test/python/tokenizer/test_icu.py index 83668b39..d85a5b65 100644 --- a/test/python/tokenizer/test_icu.py +++ b/test/python/tokenizer/test_icu.py @@ -1,8 +1,15 @@ +# SPDX-License-Identifier: GPL-2.0-only +# +# This file is part of Nominatim. (https://nominatim.org) +# +# Copyright (C) 2022 by the Nominatim developer community. +# For a full list of authors see the git log. """ Tests for ICU tokenizer. """ import shutil import yaml +import itertools import pytest @@ -65,12 +72,15 @@ def analyzer(tokenizer_factory, test_config, monkeypatch, def _mk_analyser(norm=("[[:Punctuation:][:Space:]]+ > ' '",), trans=(':: upper()',), variants=('~gasse -> gasse', 'street => st', ), - sanitizers=[]): + sanitizers=[], with_housenumber=False): cfgstr = {'normalization': list(norm), 'sanitizers': sanitizers, 'transliteration': list(trans), 'token-analysis': [{'analyzer': 'generic', 'variants': [{'words': list(variants)}]}]} + if with_housenumber: + cfgstr['token-analysis'].append({'id': '@housenumber', + 'analyzer': 'housenumbers'}) (test_config.project_dir / 'icu_tokenizer.yaml').write_text(yaml.dump(cfgstr)) tok.loader = nominatim.tokenizer.icu_rule_loader.ICURuleLoader(test_config) @@ -394,7 +404,9 @@ class TestPlaceAddress: @pytest.fixture(autouse=True) def setup(self, analyzer, sql_functions): - with analyzer(trans=(":: upper()", "'🜵' > ' '")) as anl: + hnr = {'step': 'clean-housenumbers', + 'filter-kind': ['housenumber', 'conscriptionnumber', 'streetnumber']} + with analyzer(trans=(":: upper()", "'🜵' > ' '"), sanitizers=[hnr]) as anl: self.analyzer = anl yield anl @@ -440,13 +452,6 @@ class TestPlaceAddress: 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', @@ -553,3 +558,142 @@ class TestPlaceAddress: assert 'addr' not in info + +class TestPlaceHousenumberWithAnalyser: + + @pytest.fixture(autouse=True) + def setup(self, analyzer, sql_functions): + hnr = {'step': 'clean-housenumbers', + 'filter-kind': ['housenumber', 'conscriptionnumber', 'streetnumber']} + with analyzer(trans=(":: upper()", "'🜵' > ' '"), sanitizers=[hnr], with_housenumber=True) as anl: + self.analyzer = anl + yield anl + + + @pytest.fixture + def getorcreate_hnr_id(self, temp_db_cursor): + temp_db_cursor.execute("""CREATE OR REPLACE FUNCTION create_analyzed_hnr_id(norm_term TEXT, lookup_terms TEXT[]) + RETURNS INTEGER AS $$ + SELECT -nextval('seq_word')::INTEGER; $$ LANGUAGE SQL""") + + + def process_address(self, **kwargs): + return self.analyzer.process_place(PlaceInfo({'address': kwargs})) + + + 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) + + return set((t[2] for t in tokens)) + + + @pytest.mark.parametrize('hnr', ['123 a', '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_duplicates(self, getorcreate_hnr_id): + info = self.process_address(housenumber='134', + conscriptionnumber='134', + streetnumber='99a') + + assert set(info['hnr'].split(';')) == set(('134', '99 A')) + 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} + + +class TestUpdateWordTokens: + + @pytest.fixture(autouse=True) + def setup(self, tokenizer_factory, table_factory, placex_table, word_table): + table_factory('search_name', 'place_id BIGINT, name_vector INT[]') + self.tok = tokenizer_factory() + + + @pytest.fixture + def search_entry(self, temp_db_cursor): + place_id = itertools.count(1000) + + def _insert(*args): + temp_db_cursor.execute("INSERT INTO search_name VALUES (%s, %s)", + (next(place_id), list(args))) + + return _insert + + + @pytest.fixture(params=['simple', 'analyzed']) + def add_housenumber(self, request, word_table): + if request.param == 'simple': + def _make(hid, hnr): + word_table.add_housenumber(hid, hnr) + elif request.param == 'analyzed': + def _make(hid, hnr): + word_table.add_housenumber(hid, [hnr]) + + return _make + + + @pytest.mark.parametrize('hnr', ('1a', '1234567', '34 5')) + def test_remove_unused_housenumbers(self, add_housenumber, word_table, hnr): + word_table.add_housenumber(1000, hnr) + + assert word_table.count_housenumbers() == 1 + self.tok.update_word_tokens() + assert word_table.count_housenumbers() == 0 + + + def test_keep_unused_numeral_housenumbers(self, add_housenumber, word_table): + add_housenumber(1000, '5432') + + assert word_table.count_housenumbers() == 1 + self.tok.update_word_tokens() + assert word_table.count_housenumbers() == 1 + + + def test_keep_housenumbers_from_search_name_table(self, add_housenumber, word_table, search_entry): + add_housenumber(9999, '5432a') + add_housenumber(9991, '9 a') + search_entry(123, 9999, 34) + + assert word_table.count_housenumbers() == 2 + self.tok.update_word_tokens() + assert word_table.count_housenumbers() == 1 + + + def test_keep_housenumbers_from_placex_table(self, add_housenumber, word_table, placex_table): + add_housenumber(9999, '5432a') + add_housenumber(9990, '34z') + placex_table.add(housenumber='34z') + placex_table.add(housenumber='25432a') + + assert word_table.count_housenumbers() == 2 + self.tok.update_word_tokens() + assert word_table.count_housenumbers() == 1 + + + def test_keep_housenumbers_from_placex_table_hnr_list(self, add_housenumber, word_table, placex_table): + add_housenumber(9991, '9 b') + add_housenumber(9990, '34z') + placex_table.add(housenumber='9 a;9 b;9 c') + + assert word_table.count_housenumbers() == 2 + self.tok.update_word_tokens() + assert word_table.count_housenumbers() == 1