ツイートのアーカイブからツイートの長さの中央値を求める

tweet_length_chuochi.rb

require 'json'

def entities_length(tweet)
  indices = []
  
  tweet['entities']['user_mentions'].each do |item|
    indices << item['indices']
  end
  
  tweet['entities']['urls'].each do |item|
    indices << item['indices']
  end
  
  return indices.map { |i,j| j-i }.inject(&:+) || 0
end

lengthes = [0] * 141 # lengthes[0] ~ lengthes[140]
Dir.glob('./tweets/*.js') do |path|
#Dir.glob('./tweets/2015_06.js') do |path|
  puts path
  open(path) do |io|
    io.gets
    tweets = JSON.parse(io.read)
    tweets.each do |tweet|
      next if tweet['retweeted_status']
#      next unless tweet['entities']['urls'].empty?
      lengthes[ tweet['text'].length - entities_length(tweet) ] += 1
    end
  end
end

lengthes.each_with_index do |count,length|
  print length,': ',count, "\n"
end

10.times { puts }

tweet_count = lengthes.inject(&:+)
half = tweet_count / 2

under_val = 0
lengthes.size.times do |i|
  under_val += lengthes[i]
  if under_val >= half
    puts i
    break
  end
end