def test_unicode_script_tokenizer_with_offsets2():
"""
Test UnicodeScriptTokenizer when para keep_whitespace=True and with_offsets=True
"""
unicode_script_strs2 = [["Welcome", " ", "to", " ", "Beijing", "!"],
["北京欢迎您", "！"], ["我喜欢", "English", "!"], [" "]]
dataset = ds.TextFileDataset(DATA_FILE, shuffle=False)
tokenizer = text.UnicodeScriptTokenizer(keep_whitespace=True,
with_offsets=True)
dataset = dataset.map(
input_columns=['text'],
output_columns=['token', 'offsets_start', 'offsets_limit'],
columns_order=['token', 'offsets_start', 'offsets_limit'],
operations=tokenizer)
tokens = []
expected_offsets_start = [[0, 7, 8, 10, 11, 18], [0, 15], [0, 9, 16], [0]]
expected_offsets_limit = [[7, 8, 10, 11, 18, 19], [15, 18], [9, 16, 17],
[2]]
count = 0
for i in dataset.create_dict_iterator():
token = text.to_str(i['token']).tolist()
tokens.append(token)
np.testing.assert_array_equal(i['offsets_start'],
expected_offsets_start[count])
np.testing.assert_array_equal(i['offsets_limit'],
expected_offsets_limit[count])
count += 1
logger.info("The out tokens is :", tokens)
assert unicode_script_strs2 == tokens
def test_unicode_script_tokenizer_default2():
"""
Test UnicodeScriptTokenizer when para keep_whitespace=True
"""
unicode_script_strs2 = [["Welcome", " ", "to", " ", "Beijing", "!"],
["北京欢迎您", "！"], ["我喜欢", "English", "!"], [" "]]
dataset = ds.TextFileDataset(DATA_FILE, shuffle=False)
tokenizer = text.UnicodeScriptTokenizer(keep_whitespace=True)
dataset = dataset.map(operations=tokenizer)
tokens = []
for i in dataset.create_dict_iterator():
token = text.to_str(i['text']).tolist()
tokens.append(token)
logger.info("The out tokens is :", tokens)
assert unicode_script_strs2 == tokens