Exemplo n.º 1
0
 def test_convert_binary_bracketing(self):
     given = "( 0 ( ( 1 2 ) 3 ) ) )"
     expected_tokens = [str(x) for x in range(4)]
     expected_transitions = [0, 0, 0, 1, 0, 1, 1]
     tokens, transitions = util.ConvertBinaryBracketedSeq(given.split(' '))
     assert all(t == e for t, e in zip(tokens, expected_tokens))
     assert all(t == e for t, e in zip(transitions, expected_transitions))
Exemplo n.º 2
0
def load_data(path, lowercase=None):
    examples = []
    with open(path) as f:
        for pairID, line in enumerate(f):
            line = line.strip()
            label, s1, s2 = line.split('\t')
            tokens1, transitions1 = util.ConvertBinaryBracketedSeq(s1.split(' '))
            tokens2, transitions2 = util.ConvertBinaryBracketedSeq(s2.split(' '))

            example = {}
            example["label"] = label
            example["premise"] = s1
            example["premise_tokens"] = tokens1
            example["premise_transitions"] = transitions1
            example["hypothesis"] = s2
            example["hypothesis_tokens"] = tokens2
            example["hypothesis_transitions"] = transitions2
            example["example_id"] = str(pairID)

            examples.append(example)
    return examples
Exemplo n.º 3
0
def load_data(path, lowercase=None):
    examples = []
    with open(path) as f:
        for example_id, line in enumerate(f):
            line = line.strip()
            label, seq = line.split('\t')
            tokens, transitions = util.ConvertBinaryBracketedSeq(
                seq.split(' '))

            example = {}
            example["label"] = label
            example["sentence"] = seq
            example["tokens"] = tokens
            example["transitions"] = transitions
            example["example_id"] = str(example_id)

            examples.append(example)
    return examples