def test_filter(self, file, reference_dict, num_present):
target = sample_dict_subset(reference_dict, num_present)
def condition(word):
return word in target
actual = dict(file.filter(condition))
assert_equiv_word2vec(actual, target)
def test_load(self, file, reference_dict):
target_word2vec = sample_dict_subset(reference_dict, 10)
words_to_load = list(target_word2vec.keys())
# All present words
word2vec = file.load(words_to_load)
assert_equiv_word2vec(word2vec, target_word2vec)
# One word missing
with pytest.raises(KeyError):
file.load(words_to_load + ['<<missing-word>>'])
def generate_find_case(word2vec: dict, num_present_words: int,
num_missing_words: int):
""" Generate a test case for EmbFile.find() """
target_dict = sample_dict_subset(word2vec, num_present_words)
present_words = list(target_dict.keys())
missing_words = generate_vocab(num_missing_words, forbidden_words=word2vec)
query_words = shuffled(present_words + list(missing_words))
return SimpleNamespace(
query_words=query_words, # words to find
target_dict=target_dict, # expected output word2vec dictionary
present_words=present_words, # subset of query_words in the file
missing_words=missing_words) # subset of query_words NOT in the file