python tweetstorm的简单脚本

Posted 2021-05-09

# -*- coding: utf-8 -*-


class Tweetstorm(object):

    MAX_TWEET_LENGTH = 134 # + 6 chars from initial string NN/NN = 140
    TWEET_MAXIMUM_PERCENTAGE_COMPLETE_SENTENCE = 0.6
    _separations_points = ['.', '?', '!', ':']
    _tweets = []

    def __init__(self, text):
        self.text = text.replace('\n', '')
        self._build()

    def __str__(self):
        return '\n'.join(self._tweets)

    def get_tweets(self):
        return self._tweets

    def _find_max_segment_from_dots(self, text_segment):
        segment = None, None
        biggest_size = 0
        for point in self._separations_points:
            result = text_segment.rfind(point)
            if result is not -1 and len(text_segment[:result]) > biggest_size:
                segment = text_segment[:result + 1].strip(), len(text_segment[:result + 1])
                biggest_size = segment[1]

        return segment

    def _find_max_segment_without_dots(self, text_segment):
        segment = ''
        for seg in text_segment.split(' '):
            if len(segment) + len(seg) > self.MAX_TWEET_LENGTH - 3:
                break
            segment += ' ' + seg

        return segment.strip() + '...', len(segment)

    def _build(self):
        initial_position = 0
        end_position = self.MAX_TWEET_LENGTH

        while initial_position < len(self.text):

            segment = self.text[initial_position:end_position]
            tweet_text, tweet_length = self._find_max_segment_from_dots(segment)

            if tweet_text is None or \
                    tweet_length / self.MAX_TWEET_LENGTH <= self.TWEET_MAXIMUM_PERCENTAGE_COMPLETE_SENTENCE:
                tweet_text, tweet_length = self._find_max_segment_without_dots(segment)

            self._tweets.append(tweet_text)
            initial_position += tweet_length
            end_position = initial_position + self.MAX_TWEET_LENGTH

        self._set_prefix()

        return self._tweets

    def _set_prefix(self):
        for idx, tweet in enumerate(self._tweets):
            self._tweets[idx] = '{}/{} {}'.format(idx + 1, len(self._tweets), tweet)


if __name__ == '__main__':
    import sys

    if len(sys.argv) < 2:
        print("Tweetstorm generator need an file as argument")
        sys.exit(1)

    try:
        with open(sys.argv[1], 'r') as f:
            text = f.read()
    except Exception as e:
        print(e)
        sys.exit(1)

    print(Tweetstorm(text))