summarize.py 文件源码

def run(self):
        users = {}
        for tweet_str in self.input().open('r'):
            tweet = json.loads(tweet_str)
            user = tweet['user']['screen_name']
            followers = int(tweet['user']['followers_count'])
            following = int(tweet['user']['friends_count'])
            if following > 0:
                r = followers / float(following)
                users[user] = r

        with self.output().open('w') as fp_counts:
            writer = csv.DictWriter(fp_counts, delimiter=',',
                                    quoting=csv.QUOTE_MINIMAL,
                                    fieldnames=['user', 'count'])
            writer.writeheader()
            for user, r in users.items():
                writer.writerow({'user': user, 'count': r})