Git Product home page Git Product logo

bitbot's Introduction

Bitbot

Gem Version Build Status Maintainability Test Coverage License

Bitbot is a lightweight Rack endpoint specifically intended for Slack webhooks. It can be mounted within a Rails app, or can be run as a standalone Rack app with the config.ru provided as an example.

You can write custom responders that take advantage of the logic within your larger application. Responders have support for custom routing and can utilize Wit.ai natural language processing.

For more complex responder examples, check out the bitbot-responders project.

Installation

gem "bitbot", github: "jejacks0n/bitbot"

Rails

Bitbot can run fine without Rails, but if you're using Rails, you can run the install generator. The generator will provide an initializer and mount the rack app within your routes -- be sure to update both the initializer and route if you change where it's mounted.

rails generate bitbot:install

Configuration

Bitbot requires being configured, but to simplify the README it's not included here, please check the config.ru for an example and configuration documentation.

The config.ru file is provided as a convenience and only serves as an example -- it is not included with the gem.

You can grab the config.ru and run the listener with rackup. Or if you've installed the generator, you can test your setup by starting your rails server and running (based on your configuration and port):

curl --data 'text=help+me&user_name=tester&channel=none&token=token' \
http://localhost:9292/rack-bitbot-webhook

You should get a JSON response back. If you don't, Bitbot is intentionally vague about what could've gone wrong, but the likely causes are that the token isn't correct, the request isn't a post, or that the username was the same as the bots (she doesn't respond to herself).

Setting up Slack

To get all of the configuration tokens and urls, you'll need to go to Slack and add the Incoming Webhooks, and Outgoing Webhooks integrations. You can get your incoming url, and outgoing token by doing this, which you can then set as environment variables and load them into your configuration.

When setting up the Outgoing Webhook integration you will need to know where you have configured the Rack endpoint so you can provide that as the url that will be used.

Adding Responders

There's a basic DSL for creating responders, which allows you to register help for the various commands, and define responder routes. Bitbot considers commands to be "routable", and so you can define them using route. Here's an example responder that specifies category, help and a single route. The category indicates grouping within the default help responder, but is somewhat arbitrary in it's meaning should you do something else with it.

class MyResponder < Bitbot::Responder
  category "Greetings"
  help "hi bot", description: "I'll respond with a greeting"

  route :say_hi_back, /^hi bot/i do
    respond_with("awesome! hi #{message.user_name}.")
  end
end

A route must be named, and provide a regexp matcher. Here's another example, but here we capture a value from the message.

In general a responder route will return a hash that's then sent back to the Slack request but additional messages can be announced from within the responder. As a general rule you should always use the respond_with method in your responder routes because it can determine if it should return the Hash, or make the announcement itself. You can also use the private_message, or public_message helper methods, which always announce and don't return a Hash.

route :echo, /^echo (.*)/i do |string|
  respond_with("heard #{message.user_name} say \"#{string}\" in #{message.channel}.")
end

Confirmations

Confirmations are included as a base feature, but need redis to work. Provide your own redis connection in the configuration and you can add confirmations (and more) to your responders. By default the configuration assumes redis is running locally, and is available at Redis.current -- otherwise it will try to connect to redis at the standard port.

route :say_hi_back?, /^hi bot/i do
  confirm("were you saying hi to me?", "yes") do
    respond_with("awesome! hi #{message.user_name}.")
  end
end

Wit.ai

We think Wit.ai is pretty rad for a bot setup, but it does take some work to get it trained and working the way you want. This is part of the fun, and part of the challenge.

To use Wit.ai in your responders, you need to require wit_ruby and include the Wit module in your responder. Then you can define intents, and which route they go to, as well as any entities that are within them. In the most complex form this would look something like the following.

Note: wit_ruby expects ENV["WIT_AI_TOKEN"] to be defined. read more

class MyResponder < Bitbot::Responder
  include Bitbot::Responder::Wit
  category "Greetings"
  help "hi bot my name is <name>", description: "I'll respond with a greeting"

  intent "greeting", :say_hi_back, entities: { contact: ->(e) { e['value'] } }
  route :say_hi_back, /^hi bot, my name is (.*)/i do |specified_name|
    respond_with("awesome! hi #{specified_name}, I'm bot.")
  end
end

Now if you train Wit to understand "Hello, I'm Jeremy Jackson", including the name portion as a wit/contact entity, it will make it through to the responder as the specified_name argument to the block. Again, this is a complex thing to setup and train, so have fun with it. You may also note that the route has a fallback regexp that allows using directly, even if Wit.ai wasn't able to determine what the intent was.

Worth mentioning, the proc that you see in the entities above doesn't need to be specified if all you want is the value, but if it's a proc it will call the proc with the entity hash. Some entities have complex structures, like duration, where you may want to pull out the seconds, instead of the number of minutes or hours that may have been provided. In those cases use duration: ->(e) { e['normalized']['value'] }, but in our above example, we could've just used contact: nil and the value would be pulled automatically for us.

Announcing

You can announce any message into any channel on Slack using the bot, for instance in a background job to have something happen on an action or predefined schedule. You must configure Bitbot's webhook_url by setting up an Incoming Webhook Integration on Slack before this will work however.

Bitbot.announce(text: "Hello all!", channel: "#general")

You can send private messages if you like as well.

Bitbot.announce(text: "Hello you!", channel: "@username")

You can also reuse any of the existing responder routes by having the responder handle the route directly. Obviously in these cases you must provide anything that that the might expect from the message, which always includes text, and may include common things like channel or user_name. Since responder routes can be pretty vague, and implement any number of things, you may have to provide additional information as well.

Since responders can make their own announcements, or return a hash you can use the Bitbot.announce method based on configuration.

Bitbot.announce(MyResponder.new.respond_to(text: "Hi bot", channel: "#general", user_name: "system"))
# or
MyResponder.new.respond_to(text: "Hi bot", channel: "#general", user_name: "system")

Regex can get hairy, so the respond_to method also accepts a block to return any additional context you may want to use inside your route.

MyResponder.new.respond_to(text: "Archive user", channel: "#admin", user_name: "system") { { id: 2 } }

route :archive, /^archive user/i do
  # access info from blockk
  more_info = context_block.call
  # do something with the info
  User.find(more_info[:id]).archive
  respond_with("You got it, User with id: #{more_info[:id]} has been archived.")
end

License

Licensed under the MIT License

Copyright 2019 jejacks0n

Make Code Not War


bitbot's People

Contributors

aka47 avatar brunchybit avatar dlinch avatar jejacks0n avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.