def test_sentencier_en_new_lines(self):
"""
New lines are also considered as a separator.
"""
sentencizer = Sentencizer()
text = 'It is a sunny day!!!! When Andy comes back,\n' \
'we are going to the zoo.'
crafted_chunk_list = sentencizer.craft(text, 0)
self.assertEqual(len(crafted_chunk_list), 3)
def test_sentencier_en_float_numbers(self):
"""
Separators in float numbers, URLs, emails, abbreviations (like 'Mr.')
are not taking into account.
"""
sentencizer = Sentencizer()
text = 'With a 0.99 probability this sentence will be ' \
'tokenized in 2 sentences.'
crafted_chunk_list = sentencizer.craft(text, 0)
self.assertEqual(len(crafted_chunk_list), 2)
def test_sentencier_en_trim_spaces(self):
"""
Trimming all spaces at the beginning an end of the chunks.
Keeping extra spaces inside chunks.
Ignoring chunks with only spaces.
"""
sentencizer = Sentencizer()
buffer = b' This , text is... . Amazing !!'
chunks = [i["text"] for i in sentencizer.craft(buffer, 0)]
self.assertListEqual(chunks, ["This , text is", "Amazing"])
def test_sentencier_en_trim_spaces(self):
"""
Trimming all spaces at the beginning an end of the chunks.
Keeping extra spaces inside chunks.
Ignoring chunks with only spaces.
"""
sentencizer = Sentencizer()
text = ' This , text is... . Amazing !!'
chunks = [i['text'] for i in sentencizer.craft(text, 0)]
locs = [i['location'] for i in sentencizer.craft(text, 0)]
self.assertListEqual(chunks, ["This , text is...", "Amazing"])
self.assertEqual(text[locs[0][0]:locs[0][1]], ' This , text is...')
self.assertEqual(text[locs[1][0]:locs[1][1]], ' Amazing')
def validate(req):
self.assertEqual(req.docs[0].chunks[0].text, 'This , text is...')
self.assertEqual(req.docs[0].chunks[1].text, 'Amazing')
f = Flow().add(yaml_path='!Sentencizer')
with f:
f.index_lines([' This , text is... . Amazing !!'], output_fn=validate, callback_on_body=True)
def main(data_repeat=100, repeat=3):
"""
:param data_repeat: Number of times do you want to repeat the byte array.
:param repeat: Number of times to repeat the test to get a more exact average time.
"""
# Setup.
raw_bytes = data_loader.get_bytes(repeat=data_repeat)
sentencizer = Sentencizer()
# Test execution.
timer = timeit.Timer(lambda: sentencizer.craft(raw_bytes, 0))
times = timer.repeat(repeat, 1)
# Show results.
mb = len(raw_bytes) / 2**20
time_mean = sum(times) / repeat
print(f"{mb:.2f}MB read {repeat} times.")
print(f"Average per iteration: {time_mean:.6f}.")
def test_sentencier_en(self):
sentencizer = Sentencizer()
raw_bytes = b'It is a sunny day!!!! When Andy comes back, we are going to the zoo.'
crafted_chunk_list = sentencizer.craft(raw_bytes, 0)
self.assertEqual(len(crafted_chunk_list), 2)
def test_sentencier_cn(self):
sentencizer = Sentencizer()
raw_bytes = '今天是个大晴天！安迪回来以后，我们准备去动物园。'.encode('utf8')
crafted_chunk_list = sentencizer.craft(raw_bytes, 0)
self.assertEqual(len(crafted_chunk_list), 2)
def test_sentencier_cn(self):
sentencizer = Sentencizer()
text = '今天是个大晴天！安迪回来以后，我们准备去动物园。'
crafted_chunk_list = sentencizer.craft(text, 0)
self.assertEqual(len(crafted_chunk_list), 2)
def test_sentencier_cn(self):
sentencizer = Sentencizer()
text = '今天是个大晴天！安迪回来以后，我们准备去动物园。'
crafted_chunk_list = sentencizer.craft(text, 0)
# Sentencizer does not work for chinese because string.printable does not contain Chinese characters
self.assertEqual(len(crafted_chunk_list), 0)