不使用其他LIB的情绪分析

Question

我试着评估一条推特的分数。除numpy外，不允许使用其他python库。以下是我尝试过的：

corpus = {}
with open("corpus.txt") as f:
    value=f.readlines()
    for line in value:
       (key, val) = line.strip("\n").split("\t")
       corpus[key] = int(val)
print(corpus)

{'limited': -1, 'suicidal': -2, 'pardon': 2, 'desirable': 2, 'protest': -2, 'lurking': -1, 'controversial': -2, 'hating': -3, 'ridiculous': -3, 'hate': -3,......} 

sub=('strength reassuring responsible reassuring elation elation dupe dupe')
sub=bag_of_words (sub)
print(sub)

{'dupe': 2, 'elation': 2, 'strength': 1, 'responsible': 1, 'reassuring': 2}

for k,v in sub.iteritems:
    v=v*corpus[k] if k in corpus
print(sub)

score=sum(sub.value())/len(sub)

理想的结果是：

score=(v1+v2...+vi)/i用于子

有人能帮我找出解决比赛和得分部分的代码吗？谢谢