Files
Nominatim/test/python/api/search/test_api_search_query.py
2023-05-21 09:30:57 +02:00

102 lines
3.7 KiB
Python

# SPDX-License-Identifier: GPL-3.0-or-later
#
# This file is part of Nominatim. (https://nominatim.org)
#
# Copyright (C) 2023 by the Nominatim developer community.
# For a full list of authors see the git log.
"""
Tests for tokenized query data structures.
"""
import pytest
from nominatim.api.search import query
class MyToken(query.Token):
def get_category(self):
return 'this', 'that'
def mktoken(tid: int):
return MyToken(3.0, tid, 1, 'foo', True)
@pytest.mark.parametrize('ptype,ttype', [('NONE', 'WORD'),
('AMENITY', 'QUALIFIER'),
('STREET', 'PARTIAL'),
('CITY', 'WORD'),
('COUNTRY', 'COUNTRY'),
('POSTCODE', 'POSTCODE')])
def test_phrase_compatible(ptype, ttype):
assert query.PhraseType[ptype].compatible_with(query.TokenType[ttype])
@pytest.mark.parametrize('ptype', ['COUNTRY', 'POSTCODE'])
def test_phrase_incompatible(ptype):
assert not query.PhraseType[ptype].compatible_with(query.TokenType.PARTIAL)
def test_query_node_empty():
qn = query.QueryNode(query.BreakType.PHRASE, query.PhraseType.NONE)
assert not qn.has_tokens(3, query.TokenType.PARTIAL)
assert qn.get_tokens(3, query.TokenType.WORD) is None
def test_query_node_with_content():
qn = query.QueryNode(query.BreakType.PHRASE, query.PhraseType.NONE)
qn.starting.append(query.TokenList(2, query.TokenType.PARTIAL, [mktoken(100), mktoken(101)]))
qn.starting.append(query.TokenList(2, query.TokenType.WORD, [mktoken(1000)]))
assert not qn.has_tokens(3, query.TokenType.PARTIAL)
assert not qn.has_tokens(2, query.TokenType.COUNTRY)
assert qn.has_tokens(2, query.TokenType.PARTIAL)
assert qn.has_tokens(2, query.TokenType.WORD)
assert qn.get_tokens(3, query.TokenType.PARTIAL) is None
assert qn.get_tokens(2, query.TokenType.COUNTRY) is None
assert len(qn.get_tokens(2, query.TokenType.PARTIAL)) == 2
assert len(qn.get_tokens(2, query.TokenType.WORD)) == 1
def test_query_struct_empty():
q = query.QueryStruct([])
assert q.num_token_slots() == 0
def test_query_struct_with_tokens():
q = query.QueryStruct([query.Phrase(query.PhraseType.NONE, 'foo bar')])
q.add_node(query.BreakType.WORD, query.PhraseType.NONE)
q.add_node(query.BreakType.END, query.PhraseType.NONE)
assert q.num_token_slots() == 2
q.add_token(query.TokenRange(0, 1), query.TokenType.PARTIAL, mktoken(1))
q.add_token(query.TokenRange(1, 2), query.TokenType.PARTIAL, mktoken(2))
q.add_token(query.TokenRange(1, 2), query.TokenType.WORD, mktoken(99))
q.add_token(query.TokenRange(1, 2), query.TokenType.WORD, mktoken(98))
assert q.get_tokens(query.TokenRange(0, 2), query.TokenType.WORD) == []
assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.WORD)) == 2
partials = q.get_partials_list(query.TokenRange(0, 2))
assert len(partials) == 2
assert [t.token for t in partials] == [1, 2]
assert q.find_lookup_word_by_id(4) == 'None'
assert q.find_lookup_word_by_id(99) == '[W]foo'
def test_query_struct_incompatible_token():
q = query.QueryStruct([query.Phrase(query.PhraseType.COUNTRY, 'foo bar')])
q.add_node(query.BreakType.WORD, query.PhraseType.COUNTRY)
q.add_node(query.BreakType.END, query.PhraseType.NONE)
q.add_token(query.TokenRange(0, 1), query.TokenType.PARTIAL, mktoken(1))
q.add_token(query.TokenRange(1, 2), query.TokenType.COUNTRY, mktoken(100))
assert q.get_tokens(query.TokenRange(0, 1), query.TokenType.PARTIAL) == []
assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.COUNTRY)) == 1