knock72.py 文件源码

def make_feature(f, flag):
    with open(file_name) as f:
        feature = defaultdict(int)
        for i,line in enumerate(f):
            print(i)
            y,x = line.split('\t')
            y = int(y)
            words = x.split()
            for word in words:
                if stop_word_check(word) == False:
                    word = stem.PorterStemmer().stem(word)
                    if flag == 0:
                        feature[word] += 1
                    elif flag == 1:
                        if y == 1:
                            feature[word] += 1
                        elif y == -1:
                            feature[word] -= 1
    return feature