test/python/tokenizer/sanitizers/test_strip_brace_terms.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 the sanitizer that handles braced suffixes.
   9 """
  10 import pytest
  11
  12 from nominatim_db.tokenizer.place_sanitizer import PlaceSanitizer
  13 from nominatim_db.data.place_info import PlaceInfo
  14
  15 class TestStripBrace:
  16
  17     @pytest.fixture(autouse=True)
  18     def setup_country(self, def_config):
  19         self.config = def_config
  20
  21     def run_sanitizer_on(self, **kwargs):
  22         place = PlaceInfo({'name': kwargs})
  23         name, _ = PlaceSanitizer([{'step': 'strip-brace-terms'}], self.config).process_names(place)
  24
  25         return sorted([(p.name, p.kind, p.suffix) for p in name])
  26
  27
  28     def test_no_braces(self):
  29         assert self.run_sanitizer_on(name='foo', ref='23') == [('23', 'ref', None),
  30                                                                ('foo', 'name', None)]
  31
  32
  33     def test_simple_braces(self):
  34         assert self.run_sanitizer_on(name='Halle (Saale)', ref='3')\
  35           == [('3', 'ref', None), ('Halle', 'name', None), ('Halle (Saale)', 'name', None)]
  36         assert self.run_sanitizer_on(name='ack ( bar')\
  37           == [('ack', 'name', None), ('ack ( bar', 'name', None)]
  38
  39
  40     def test_only_braces(self):
  41         assert self.run_sanitizer_on(name='(maybe)') == [('(maybe)', 'name', None)]
  42
  43
  44     def test_double_braces(self):
  45         assert self.run_sanitizer_on(name='a((b))') == [('a', 'name', None),
  46                                                         ('a((b))', 'name', None)]
  47         assert self.run_sanitizer_on(name='a (b) (c)') == [('a', 'name', None),
  48                                                            ('a (b) (c)', 'name', None)]
  49
  50
  51 def test_no_names(def_config):
  52     place = PlaceInfo({'address': {'housenumber': '3'}})
  53     name, address = PlaceSanitizer([{'step': 'strip-brace-terms'}], def_config).process_names(place)
  54
  55     assert not name
  56     assert len(address) == 1