negation_detection.py 文件源码

def breakWithOutWhiteSpace(sentence):
    import re
    r = "\.\w+"
    sentences = []
    tmp = re.findall(r, sentence, re.X)
    places = [0]
    if len(tmp)>0:
        import enchant
        d = enchant.Dict("en_UK")
        for item in tmp:
            word = item[1:]
            if len(word)<2:
                if word.lower() in ['i','a']:
                    places.extend([m.start() for m in re.finditer(item, sentence)])
            else:
                if d.check(item[1:]):
                    places.extend([m.start() for m in re.finditer(item, sentence)])

    places = sorted(set(places))
    places.append(len(sentence)-1)
    i = 0
    if len(places)==2:
        return [sentence]

    start = 0
    while True:
        start = places[i]
        if start>0:
            start +=1
        end = places[i+1] + 1
        if end>len(sentence):
            end = len(sentence)-1
        sentences.append(sentence[start:end])
        i +=1
        if len(sentences)==len(places)-1:
            break

    return sentences