def test_baseform_normalizer(self):
tokenizer = Tokenizer(lang="en")
text = "goes playing"
tokens = tokenizer.transform([text])
normalized = tfm.BaseFormNormalizer().transform(tokens)[0]
self.assertTrue("go", normalized[0])
self.assertTrue("play", normalized[1])
def test_ja_tokenize(self):
tokenizer = Tokenizer(lang="ja")
text = "日本語の形態素解析、マジ卍。"
tokens = tokenizer.transform(text)
for t in tokens:
self.assertFalse(t.is_spacy)
self.assertTrue(t.is_ja)
self.assertTrue(t.surface)
self.assertTrue(t.base_form)
self.assertTrue(t.pos)
self.assertTrue(t.tag)
def test_split_tokenize(self):
tokenizer = Tokenizer(lang=None)
text = "Tom goes to a park that Mary is playing."
tokens = tokenizer.transform(text)
for t in tokens:
self.assertFalse(t.is_spacy)
self.assertFalse(t.is_ja)
self.assertTrue(t.surface)
self.assertTrue(t.base_form)
self.assertTrue(t.pos)
self.assertTrue(t.tag)
def test_setter_token(self):
vocab = Vocabulary()
text = "you are making the vocabulary"
words = Tokenizer(lang="en").transform([text])[0]
vocab.set(words)
vocab_size = len(words) + len(["_pad_", "_unk_", "_bos", "_eos"])
self.assertEqual(len(vocab.get()), vocab_size)
print(vocab.get())
def __init__(self,
tokenizer=None,
text_transformers=(),
token_transformers=(),
vocabulary=None,
other_transformers=()):
self.tokenizer = tokenizer
if isinstance(self.tokenizer, str):
self.tokenizer = Tokenizer(self.tokenizer)
self.text_transformers = list(text_transformers)
self.token_transformers = list(token_transformers)
self.vocabulary = vocabulary
self.other_transformers = list(other_transformers)
def test_baseform_normalizer(self):
tokenizer = Tokenizer(lang="en")
text = "five players of 3 state on 1,000 location"
tokens = tokenizer.transform([text])
normalized = tfm.NumberNormalizer().transform(tokens)[0]
self.assertEqual(2, len([t for t in normalized if t.surface == "0"]))
tokenizer = Tokenizer(lang="ja")
text = "百に一つの場所に2人の人がいる"
tokens = tokenizer.transform([text])
normalized = tfm.NumberNormalizer().transform(tokens)[0]
self.assertEqual(2, len([t for t in normalized if t.surface == "0"]))
def test_stopword_ja(self):
tokenizer = Tokenizer(lang="ja")
text = "わたしの形態素解析、マジ卍。"
tokens = tokenizer.transform([text])
filtered = tfm.StopwordFilter(lang="ja").transform(tokens)
self.assertTrue(1, len(tokens) - len(filtered))
def test_en_tokenize(self):
tokenizer = Tokenizer(lang="en")
text = "Tom goes to a park that Mary is playing."
tokens = tokenizer.transform([text])
filtered = tfm.StopwordFilter(lang="en").transform(tokens)
self.assertTrue(4, len(tokens) - len(filtered))
def test_word_frequency_filter_en(self):
tokenizer = Tokenizer(lang="en")
text = "Tom goes to a park that Mary is playing. Tom and Mary is playing tennis in the park."
tokens = tokenizer.transform([text])
filtered = tfm.WordFrequencyFilter(n=3, min_freq=1).fit_transform(tokens)[0]
self.assertTrue(6, len(filtered))
def test_word_frequency_filter_ja(self):
tokenizer = Tokenizer(lang="ja")
text = "わたしの形態素解析は楽しい。わたしはまだまだ。"
tokens = tokenizer.transform([text])
filtered = tfm.WordFrequencyFilter(n=1, min_freq=1).fit_transform(tokens)[0]
self.assertTrue(4, len(filtered))