def testGetAnswer(self):
context = u"America — Harvard, Yale" # "—": em-dash (\u2014)
# \u2013: en-dash
context_words = ["America", u"\u2013", "Harvard", ",", "Yale"]
word_answer_start = 0
word_answer_end = 2
output = tokenizer_util.get_answer(context, context_words,
word_answer_start, word_answer_end)
expected_output = u"America — Harvard"
self.assertEqual(expected_output, output)
def testGetAnswerBytes(self):
"""Test when context and context_words have byte sequences."""
# "—": em-dash (\u2014)
context = u"America — Harvard, Yale".encode("utf-8")
# \u2013: en-dash
context_words = [
"America", u"\u2013".encode("utf-8"), "Harvard", ",", "Yale"
]
word_answer_start = 0
word_answer_end = 2
output = tokenizer_util.get_answer(context,
context_words,
word_answer_start,
word_answer_end,
is_byte=True)
expected_output = u"America — Harvard".encode("utf-8")
self.assertEqual(expected_output, output)