spaCy/spacy/tests/lang/sv/test_exceptions.py

59 lines
1.8 KiB
Python
Raw Normal View History

import pytest
SV_TOKEN_EXCEPTION_TESTS = [
2019-02-08 16:14:49 +03:00
(
"Smörsåsen används bl.a. till fisk",
["Smörsåsen", "används", "bl.a.", "till", "fisk"],
),
(
"Jag kommer först kl. 13 p.g.a. diverse förseningar",
["Jag", "kommer", "först", "kl.", "13", "p.g.a.", "diverse", "förseningar"],
),
(
"Anders I. tycker om ord med i i.",
["Anders", "I.", "tycker", "om", "ord", "med", "i", "i", "."],
),
]
2019-02-08 16:14:49 +03:00
@pytest.mark.parametrize("text,expected_tokens", SV_TOKEN_EXCEPTION_TESTS)
def test_sv_tokenizer_handles_exception_cases(sv_tokenizer, text, expected_tokens):
tokens = sv_tokenizer(text)
token_list = [token.text for token in tokens if not token.is_space]
assert expected_tokens == token_list
2019-02-08 16:14:49 +03:00
@pytest.mark.parametrize("text", ["driveru", "hajaru", "Serru", "Fixaru"])
def test_sv_tokenizer_handles_verb_exceptions(sv_tokenizer, text):
tokens = sv_tokenizer(text)
assert len(tokens) == 2
assert tokens[1].text == "u"
2019-02-08 16:14:49 +03:00
@pytest.mark.parametrize("text", ["bl.a", "m.a.o.", "Jan.", "Dec.", "kr.", "osv."])
def test_sv_tokenizer_handles_abbr(sv_tokenizer, text):
tokens = sv_tokenizer(text)
assert len(tokens) == 1
2019-02-08 16:14:49 +03:00
@pytest.mark.parametrize("text", ["Jul.", "jul.", "sön.", "Sön."])
def test_sv_tokenizer_handles_ambiguous_abbr(sv_tokenizer, text):
tokens = sv_tokenizer(text)
assert len(tokens) == 2
def test_sv_tokenizer_handles_exc_in_text(sv_tokenizer):
text = "Det är bl.a. inte meningen"
tokens = sv_tokenizer(text)
assert len(tokens) == 5
assert tokens[2].text == "bl.a."
def test_sv_tokenizer_handles_custom_base_exc(sv_tokenizer):
text = "Här är något du kan titta på."
tokens = sv_tokenizer(text)
assert len(tokens) == 8
assert tokens[6].text == ""
assert tokens[7].text == "."