deleted_tweets.rb

deleted_tweets.rb

require 'twitter'
require 'big_query'

# Twitter Gem のバグを回避
class HTTP::URI
  def port
    443 if self.https?
  end
end

# Twitter API 認証情報
CONSUMER_KEY        = ''
CONSUMER_SECRET     = ''
ACCESS_TOKEN        = ''
ACCESS_TOKEN_SECRET = ''

# Twitter Client の新規作成
streaming_client = Twitter::Streaming::Client.new do |config|
  config.consumer_key        = CONSUMER_KEY        
  config.consumer_secret     = CONSUMER_SECRET     
  config.access_token        = ACCESS_TOKEN        
  config.access_token_secret = ACCESS_TOKEN_SECRET 
end

# GCP API 認証情報
opts = {}
opts['client_id']     = ''
opts['service_email'] = ''
opts['key']           = ''
opts['project_id']    = ''
opts['dataset']       = ''

bq = BigQuery::Client.new(opts)
if bq.tables.empty?
  table_names = ['tweets', 'deleted_tweets']
  table_names.each do |table_name|
    table_name = table_name
    table_schema = { status_id:   { type: 'INTEGER' },
                     screen_name: { type: 'STRING' },
                     text:        { type: 'STRING' },
                     created_at:  { type: 'TIMESTAMP' },
                     json:        { type: 'STRING' },
                     deleted:     { type: 'INTEGER' } }
    bq.create_table(table_name, table_schema)
  end
end

streaming_client.user do |object|
  if object.kind_of? Twitter::Streaming::DeletedTweet
    status_id = object.id
    begin
      res = bq.query("SELECT * FROM [dataset.tweets] where status_id=#{status_id} LIMIT 1")
      res = res['rows'][0]['f']
      data = {
        'status_id'   => res[0]['v'],
        'screen_name' => res[1]['v'],
        'text'        => res[2]['v'],
        'json'        => res[4]['v']
      }
      puts data['screen_name'].to_s + ": " + data['text']
      bq.insert('deleted_tweets', data)
      puts "---削除されたツイートのデータを保存した---"
    rescue
      puts "---削除されたツイートのデータがなかった---"
    end
  end
  if object.kind_of? Twitter::Tweet
    data = {
      'status_id' => object.id,
      'screen_name' => object.user.screen_name,
      'text' => object.text,
      'created_at' => object.created_at.strftime("%Y-%m-%d %H:%M:%S"),
      'json' => object.to_h.to_json,
      'deleted' => 0 
    }
    bq.insert('tweets', data)
  end

end