]> git.openstreetmap.org Git - nominatim.git/blob - test/python/tokenizer/test_factory.py
Merge remote-tracking branch 'upstream/master'
[nominatim.git] / test / python / tokenizer / test_factory.py
1 # SPDX-License-Identifier: GPL-3.0-or-later
2 #
3 # This file is part of Nominatim. (https://nominatim.org)
4 #
5 # Copyright (C) 2025 by the Nominatim developer community.
6 # For a full list of authors see the git log.
7 """
8 Tests for creating new tokenizers.
9 """
10 import pytest
11
12 from nominatim_db.db import properties
13 from nominatim_db.tokenizer import factory
14 from nominatim_db.errors import UsageError
15 from dummy_tokenizer import DummyTokenizer
16
17
18 def test_setup_bad_tokenizer_name(project_env, monkeypatch):
19     monkeypatch.setenv('NOMINATIM_TOKENIZER', 'dummy')
20
21     with pytest.raises(UsageError):
22         factory.create_tokenizer(project_env)
23
24
25 class TestFactory:
26     @pytest.fixture(autouse=True)
27     def init_env(self, project_env, property_table, tokenizer_mock):
28         self.config = project_env
29
30     def test_setup_dummy_tokenizer(self, temp_db_conn):
31         tokenizer = factory.create_tokenizer(self.config)
32
33         assert isinstance(tokenizer, DummyTokenizer)
34         assert tokenizer.init_state == "new"
35         assert (self.config.project_dir / 'tokenizer').is_dir()
36
37         assert properties.get_property(temp_db_conn, 'tokenizer') == 'dummy'
38
39     def test_setup_tokenizer_dir_exists(self):
40         (self.config.project_dir / 'tokenizer').mkdir()
41
42         tokenizer = factory.create_tokenizer(self.config)
43
44         assert isinstance(tokenizer, DummyTokenizer)
45         assert tokenizer.init_state == "new"
46
47     def test_setup_tokenizer_dir_failure(self):
48         (self.config.project_dir / 'tokenizer').write_text("foo")
49
50         with pytest.raises(UsageError):
51             factory.create_tokenizer(self.config)
52
53     def test_load_tokenizer(self):
54         factory.create_tokenizer(self.config)
55
56         tokenizer = factory.get_tokenizer_for_db(self.config)
57
58         assert isinstance(tokenizer, DummyTokenizer)
59         assert tokenizer.init_state == "loaded"
60
61     def test_load_repopulate_tokenizer_dir(self):
62         factory.create_tokenizer(self.config)
63
64         self.config.project_dir = self.config.project_dir
65
66         factory.get_tokenizer_for_db(self.config)
67         assert (self.config.project_dir / 'tokenizer').exists()
68
69     def test_load_missing_property(self, temp_db_cursor):
70         factory.create_tokenizer(self.config)
71
72         temp_db_cursor.execute("TRUNCATE TABLE nominatim_properties")
73
74         with pytest.raises(UsageError):
75             factory.get_tokenizer_for_db(self.config)