def test_should_not_tag_using_none_tag(self):
token_1 = SimpleToken(TOKEN_TEXT_1)
structured_document = SimpleStructuredDocument(
lines=[SimpleLine([token_1])])
annotate_structured_document_using_predictions(structured_document,
[NONE_TAG])
assert structured_document.get_tag(token_1,
scope=CRF_TAG_SCOPE) is None
def test_should_raise_error_if_token_props_do_not_match(self):
token_1 = SimpleToken(TOKEN_TEXT_1, bounding_box=BOUNDING_BOX)
structured_document = SimpleStructuredDocument(
SimplePage(lines=[SimpleLine([token_1])],
bounding_box=BOUNDING_BOX))
token_props_list = list(
structured_document_to_token_props(structured_document))
token_props_list[0]['text'] = TOKEN_TEXT_2
with pytest.raises(AssertionError):
annotate_structured_document_using_predictions(
structured_document, [TAG_1], token_props_list)
def test_should_tag_single_token_using_prediction_and_check_token_props(
self):
token_1 = SimpleToken(TOKEN_TEXT_1, bounding_box=BOUNDING_BOX)
structured_document = SimpleStructuredDocument(
SimplePage(lines=[SimpleLine([token_1])],
bounding_box=BOUNDING_BOX))
token_props_list = structured_document_to_token_props(
structured_document)
annotate_structured_document_using_predictions(structured_document,
[TAG_1],
token_props_list)
assert structured_document.get_tag(token_1,
scope=CRF_TAG_SCOPE) == TAG_1
def test_should_not_fail_with_empty_document(self):
structured_document = SimpleStructuredDocument()
annotate_structured_document_using_predictions(structured_document, [])
