test/python/tokenizer/sanitizers/test_clean_housenumbers.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 the sanitizer that normalizes housenumbers.
   9 """
  10 import pytest
  11
  12 from nominatim.tokenizer.place_sanitizer import PlaceSanitizer
  13 from nominatim.indexer.place_info import PlaceInfo
  14
  15 @pytest.fixture
  16 def sanitize(request):
  17     sanitizer_args = {'step': 'clean-housenumbers'}
  18     for mark in request.node.iter_markers(name="sanitizer_params"):
  19         sanitizer_args.update({k.replace('_', '-') : v for k,v in mark.kwargs.items()})
  20
  21     def _run(**kwargs):
  22         place = PlaceInfo({'address': kwargs})
  23         _, address = PlaceSanitizer([sanitizer_args]).process_names(place)
  24
  25         return sorted([(p.kind, p.name) for p in address])
  26
  27     return _run
  28
  29
  30 def test_simple_number(sanitize):
  31     assert sanitize(housenumber='34') == [('housenumber', '34')]
  32
  33
  34 @pytest.mark.parametrize('number', ['1;2;3', '1,2,3', '1; 3 ,2',
  35                                     '2,,3,1', '1;2;3;;', ';3;2;1'])
  36 def test_housenumber_lists(sanitize, number):
  37     assert sanitize(housenumber=number) == \
  38            [('housenumber', '1'), ('housenumber', '2'), ('housenumber', '3')]
  39
  40
  41 @pytest.mark.sanitizer_params(filter_kind=('number', 'streetnumber'))
  42 def test_filter_kind(sanitize):
  43     assert sanitize(housenumber='34', number='4', badnumber='65') == \
  44             [('badnumber', '65'), ('housenumber', '34'), ('housenumber', '4')]