def test_weird_example(self):
sentences = [u'Copyright \xa9 *|CURRENT_YEAR|* *|LIST:COMPANY|*, All rights reserved.', u'*|IFNOT:ARCHIVE_PAGE|* *|LIST:DESCRIPTION|* Our mailing address is:*|HTML:LIST_ADDRESS_HTML|* *|END:IF|* unsubscribe from this list update subscription preferences*|IF:REWARDS|* *|HTML:REWARDS|* *|END:IF|*']
skeleton = u'{0}{1}'
truth_indexes = [0,len(sentences[0])]
indexes = _get_sentences_start_indexes(sentences, skeleton)
self.assertEqual(truth_indexes,indexes)
def test_canonical_cases(self):
canonical_cases = [
[["A","B"],u'{0}{1}',[0,1]],
[["A","B"],u'{0} {1}',[0,2]],
[["A","B"],u' {0}{1}',[1,2]],
[["A","B"],u' {0} {1}',[1,3]],
]
for i, (sentences,skeleton, truth) in enumerate(canonical_cases):
with self.subTest(i=i):
indexes = _get_sentences_start_indexes(sentences, skeleton)
self.assertEqual(truth,indexes)
