How I created my twitter bots

bots_creation

First, I should say that I had help. I have a lovely friend and I'm not sure whether or not she'd want to be mentioned here, but she was there for me when I ran into headaches. This was a HECK of a lot of work to do on a Windows machine. I understand it's easier on Mac.

I created the bot using the scripts at this repository:

https://github.com/mispy/twitter_ebooks

I worked with these two tutorials to get me through the process when I got stuck:

http://blog.boodoo.co/how-to-make-an-_ebooks/

https://github.com/trommel/ebooks_tutorial

Read through BOTH of these before you get started. I found them near the end and they'd have saved me a LOT of headaches.

I used the Twitter export method using my entire archive. I've been updating with live tweets but that's a whole 'nother setup. What I found, personally, is that when attempting to download using the: ebooks archive method, I have to use my Ubuntu box vs. my Windows 7 (your experience may vary and I'm told Mac works as well as Ubuntu)

I host the bot on a free https://heroku.com/ dyno (pro tip, when you've got a Heroku repository set up, you have to actually go in and add a dyno to your account or it won't run!)

When it came to actually pushing the repository to heroku, I had to use these steps:

http://stackoverflow.com/a/18676454

and used Git Bash vs. the command line to do so. If I do so using the command line, it doesn't work for me.

Sufficeth to say that the command line is your friend if you try this. I may be able to provide a little help, but I'm not so great with Ruby that I feel terribly comfortable. What I will do, however, is paste below a version of the bots.rb file which my friend gave me permission to share and which makes my bot truly awesome. You'll have to fill in details of "your_bot" and the name of your tweets.model file (if it's not just tweets.model)

--

#!/usr/bin/env ruby

require 'set'

require 'twitter_ebooks'

Ebooks::Bot.new("your_bot") do |bot|
  # Consumer details come from registering an app at https://dev.twitter.com/
  # OAuth details can be fetched with https://github.com/marcel/twurl although dev.twitter.com is better for doing all. Also the keys below have been renamed to be careful.
  bot.consumer_key = ""
  bot.consumer_secret = ""
  bot.oauth_token = ""
  bot.oauth_token_secret = ""

  model = Ebooks::Model.load('model/tweets.model')

  # mstea_ebooks only randomly tweets a person once per session
  tweeted = Set.new()

  bot.on_message do |dm|
    length = dm[:text].length
    bot.reply(dm, model.make_response(dm[:text], 140 - length))
  end

  bot.on_follow do |user|
    bot.follow(user[:screen_name])
  end

  bot.on_mention do |tweet, meta|
    # Avoid infinite reply chains (very small chance of crosstalk)
    next if tweet[:user][:screen_name].include?('ebooks') && rand > 0.05

    length = tweet[:text].length + meta[:reply_prefix].length
    response = model.make_response(tweet[:text], 140 - length)
    bot.reply(tweet, meta[:reply_prefix] + response)
  end

  bot.on_timeline do |tweet, meta|
    next if tweet[:retweeted_status] || tweet[:text].start_with?('RT')
    next unless rand < 0.05
    # mstea_ebooks only randomly tweets a person once per run
    next if tweeted.include? tweet[:user][:screen_name]

    tweeted << tweet[:user][:screen_name]

    # Reply to a tweet in the bot's timeline
    positive_words = tweet[:text] + ' good happy fun great positive perfect'
    length = tweet[:text].length + meta[:reply_prefix].length
    response = model.make_response(positive_words, 140 - length)
    bot.reply(tweet, meta[:reply_prefix] + response)
  end

  bot.scheduler.every '1h' do
    bot.tweet(model.make_statement(140))
  end
end