test/python/tokenizer/sanitizers/test_clean_tiger_tags.py

   1 # SPDX-License-Identifier: GPL-2.0-only
   2 #
   3 # This file is part of Nominatim. (https://nominatim.org)
   4 #
   5 # Copyright (C) 2022 by the Nominatim developer community.
   6 # For a full list of authors see the git log.
   7 """
   8 Tests for sanitizer that clean up TIGER tags.
   9 """
  10 import pytest
  11
  12 from nominatim.tokenizer.place_sanitizer import PlaceSanitizer
  13 from nominatim.data.place_info import PlaceInfo
  14
  15 class TestCleanTigerTags:
  16
  17     @pytest.fixture(autouse=True)
  18     def setup_country(self, def_config):
  19         self.config = def_config
  20
  21
  22     def run_sanitizer_on(self, addr):
  23         place = PlaceInfo({'address': addr})
  24         _, outaddr = PlaceSanitizer([{'step': 'clean-tiger-tags'}], self.config).process_names(place)
  25
  26         return sorted([(p.name, p.kind, p.suffix) for p in outaddr])
  27
  28     @pytest.mark.parametrize('inname,outname', [('Hamilton, AL', 'Hamilton'),
  29                                                 ('Little, Borough, CA', 'Little, Borough')])
  30     def test_well_formatted(self, inname, outname):
  31         assert self.run_sanitizer_on({'tiger:county': inname})\
  32             == [(outname, 'county', 'tiger')]
  33
  34
  35     @pytest.mark.parametrize('name', ('Hamilton', 'Big, Road', ''))
  36     def test_badly_formatted(self, name):
  37         assert self.run_sanitizer_on({'tiger:county': name})\
  38             == [(name, 'county', 'tiger')]
  39
  40
  41     def test_unmatched(self):
  42         assert self.run_sanitizer_on({'tiger:country': 'US'})\
  43             == [('US', 'tiger', 'country')]