mirror of
https://github.com/explosion/spaCy.git
synced 2025-02-06 06:30:35 +03:00
65 lines
1.4 KiB
Python
65 lines
1.4 KiB
Python
import pytest
|
|
from spacy.gold.new_example import NewExample as Example
|
|
from spacy.tokens import Doc
|
|
from spacy.vocab import Vocab
|
|
|
|
|
|
def test_Example_init_requires_doc_objects():
|
|
vocab = Vocab()
|
|
with pytest.raises(TypeError):
|
|
eg = Example(None, None)
|
|
with pytest.raises(TypeError):
|
|
eg = Example(Doc(vocab, words=["hi"]), None)
|
|
with pytest.raises(TypeError):
|
|
eg = Example(None, Doc(vocab, words=["hi"]))
|
|
|
|
|
|
|
|
def test_Example_from_dict_basic():
|
|
eg = Example.from_dict(
|
|
Doc(Vocab(), words=["hello", "world"]),
|
|
{
|
|
"words": ["hello", "world"]
|
|
}
|
|
)
|
|
assert isinstance(eg.x, Doc)
|
|
assert isinstance(eg.y, Doc)
|
|
|
|
|
|
@pytest.mark.parametrize("annots", [
|
|
{"words": ["ice", "cream"], "tags": ["NN", "NN"]},
|
|
])
|
|
def test_Example_from_dict_with_tags(annots):
|
|
vocab = Vocab()
|
|
predicted = Doc(vocab, words=annots["words"])
|
|
eg = Example.from_dict(predicted, annots)
|
|
for i, token in enumerate(eg.reference):
|
|
assert token.tag_ == annots["tags"][i]
|
|
|
|
|
|
"""
|
|
def test_Example_from_dict_with_entities(annots):
|
|
# TODO
|
|
pass
|
|
|
|
def test_Example_from_dict_with_parse(annots):
|
|
# TODO
|
|
pass
|
|
|
|
def test_Example_from_dict_with_morphology(annots):
|
|
# TODO
|
|
pass
|
|
|
|
def test_Example_from_dict_with_sent_start(annots):
|
|
# TODO
|
|
pass
|
|
|
|
def test_Example_from_dict_with_cats(annots):
|
|
# TODO
|
|
pass
|
|
|
|
def test_Example_from_dict_with_links(annots):
|
|
# TODO
|
|
pass
|
|
"""
|