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) 2024 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_core.db import properties
  13 from nominatim_db.tokenizer import factory
  14 from nominatim_core.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
  31     def test_setup_dummy_tokenizer(self, temp_db_conn):
  32         tokenizer = factory.create_tokenizer(self.config)
  33
  34         assert isinstance(tokenizer, DummyTokenizer)
  35         assert tokenizer.init_state == "new"
  36         assert (self.config.project_dir / 'tokenizer').is_dir()
  37
  38         assert properties.get_property(temp_db_conn, 'tokenizer') == 'dummy'
  39
  40
  41     def test_setup_tokenizer_dir_exists(self):
  42         (self.config.project_dir / 'tokenizer').mkdir()
  43
  44         tokenizer = factory.create_tokenizer(self.config)
  45
  46         assert isinstance(tokenizer, DummyTokenizer)
  47         assert tokenizer.init_state == "new"
  48
  49
  50     def test_setup_tokenizer_dir_failure(self):
  51         (self.config.project_dir / 'tokenizer').write_text("foo")
  52
  53         with pytest.raises(UsageError):
  54             factory.create_tokenizer(self.config)
  55
  56
  57     def test_load_tokenizer(self):
  58         factory.create_tokenizer(self.config)
  59
  60         tokenizer = factory.get_tokenizer_for_db(self.config)
  61
  62         assert isinstance(tokenizer, DummyTokenizer)
  63         assert tokenizer.init_state == "loaded"
  64
  65
  66     def test_load_repopulate_tokenizer_dir(self):
  67         factory.create_tokenizer(self.config)
  68
  69         self.config.project_dir = self.config.project_dir
  70
  71         factory.get_tokenizer_for_db(self.config)
  72         assert (self.config.project_dir / 'tokenizer').exists()
  73
  74
  75     def test_load_missing_property(self, temp_db_cursor):
  76         factory.create_tokenizer(self.config)
  77
  78         temp_db_cursor.execute("TRUNCATE TABLE nominatim_properties")
  79
  80         with pytest.raises(UsageError):
  81             factory.get_tokenizer_for_db(self.config)