catalinmiron/claudebot

Totally not trying to poke fun at the word cloud. WalaoBot's highly evolved, French cousin.

ClaudeBot

Due to recent changes in Hubot and hubot-scripts, the documentation below may not be accurate. README.MD is pending major revisions. Suggestions and contributions are welcomed.

Totally not trying to poke fun at the word cloud. *cough CloudIRC* This is a clone of GitHub's Campfire bot, Hubot. He's pretty chill. It is also WalaoBot's highly evolved, French cousin. Oui, oui.

What do you call a French bot that has been attacked by a bear? Claude bot. *Ba dum ts*

This version is designed to protect and serve the great people of FyreChat IRC #sandbox (using Hubot IRC adapter); and to be deployed on Heroku.

The original, generated Hubot README can be found here. Additional documentation can be found on Hubot's GitHub.

Features

As ClaudeBot is somewhat biologically identical to Hubot, it derives Hubot's default functionality and script. Unlike its twin in its natural form however, ClaudeBot is also exposed to custom scripts from the hubot-scripts package. A full list of enabled scripts may be found in the hubot-scripts.json file in this repo.

Scripting

ClaudeBot is open to evolution. It may be extended to become more alpha, or mutated to become more primitive and/or defunct. The Scripting Guide will provide you with more information on how you can make this brainless slave more exciting or useless and boring.

You can modify part of its genetic makeup and/or add whatever functionality you want it to have via the ./scripts folder and the hubot-scripts.json file in this repo. Browse the hubot-scripts catalog for a list of available genes that you can (de)activate on-the-fly.

Master Ian's Commands

A full list of command may be found by visiting the bot via HTTP /RobotName/help or by calling the help command.

• hubot cdnjs search <query>: Returns the CDNJS URL for the first 5 front-end dependencies matching the search <query>
• hubot cdnjs fetch <dependency>: Returns the CDNJS URL for a specific front-end <dependency> (e.g. jQuery)
• hubot so <query>: Returns the first 5 questions on Stack Overflow matching the search <query> (stack command shortcut)
• hubot stack [on] [about] <query>: Returns the first 5 questions on a Stack Exchange <site> matching the search <query>
• hubot ttv featured: Returns the first 5 featured live streams on Twitch
• hubot ttv game <category>: Returns the first 5 live streams on Twitch in a game <category> (case-sensitive)
• hubot ttv search <query>: Returns the first 5 live streams on Twitch matching the search <query>
• hubot ttv stream <name>: Returns information about Twitch stream <name>
• hubot ttv top: Returns the top 5 games sorted by the number of current viewers on Twitch, most popular first
• hubot hb stream <name>: Returns information about hitbox stream <name>
• hubot wiki search <query>: Returns the first 5 Wikipedia articles matching the search <query>
• hubot wiki summary <article>: Returns a one-line description about <article>
• hubot face [me]: Returns a random, but cool ASCII face
• hubot guess <number>: Returns a feedback indicating how accurate your 4-digit <number> guess is
• hubot guess surrender: Restarts the game with a new number, duh!
• hubot guess tutorial: Tells you how to play this game
• hubot steam id [me] <custom URL>: Returns the Steam ID for the user under http://steamcommunity.com/id/<custom URL>
• hubot steam status <Steam ID|custom URL>: Returns <Steam ID> or <custom URL> community status
• hubot dota history <Steam ID|custom URL>: Returns metadata for the latest 5 game lobbies with <Steam ID> or <custom URL>
• hubot dota match <match ID> [<Steam ID|custom URL>]: Returns information about a particular <match ID>. Optionally, if <Steam ID> or <custom URL> is included, its match information will also be returned
• hubot mail <recipient> <message>: Sends a <message> to <recipient> when found available
• hubot unmail [<recipient>]: Deletes all mail sent by you. Optionally, if <recipient> is specified, all mail sent to <recipient> by you will be deleted
• hubot <search|google> <query>: Queries Google Search for <query> and returns the first 5 results
• hubot delete <key>: Removes <key> and all of its content from the local brain / persistence
• hubot brain save: Forces a save to Dweet.io
• hubot brain status: Returns the brain status

Steam Web API which powers all the Steam and Dota 2 commands requires an API key which you may set via STEAM_API_KEY.

Brain

ClaudeBot (like its counterpart, Hubot) has a short-term and a long-term memory store. Memories in its short-term store will not be persisted should it oh I don't know... die and come back alive? Its local brain will start with a blank slate. Thus, it has an external brain which serves as a more permanent storage medium for long-term memories.

Hubot's default, external brain redis-brain.coffee was surgically removed from hubot-scripts.json in favour of a custom engineered, cloud, key-value based brain solution powered by Dweet.io.

ClaudeBot uses dweet-brain.coffee for data storage / persistence. It is however, disabled unless the DWEET_THING environment variable is set. What is DWEET_THING you may ask? Well find out for yourself. It is also set to automatically transfer data from its short-term memory to its long-term memory every 30 minutes. You can tweak it by setting the delay (in seconds) via DWEET_AUTOSAVE. You can protect your data by purchasing a Dweet.io lock and setting its key via DWEET_KEY.

Usage

Testing (Locally)

You can make ClaudeBot run in your local shell environment by commanding the following:

% bin/hubot

Then, you can start interacting with ClaudeBot by typing hubot help (it has not quite absorbed the fact it is now called Claude, at least not locally). For a list of basic commands, refer to hubot's documentation.

Adapters

ClaudeBot - as mentioned earlier - was designed to be a slave to the great people of an IRC channel. Thus, to make it serve differently it needs to be commanded the following:

% bin/hubot -a irc

You may replace hubot-irc with your desired Hubot adapter.

Configuring

CaludeBot's genetic information may be set for different environments via environment variables (also known as config vars on Heroku).
To operate in an IRC environment however, the following environment variables must be set:

• HUBOT_IRC_SERVER
• HUBOT_IRC_ROOMS
• HUBOT_IRC_NICK

For a list of IRC configuration options, refer to hubot-irc's documentation.

Remember when I stated that ClaudeBot has not quite absorbed the fact it is now called Claude? Well, you can force its name by passing (-n|--name) NAME when running it. You may also set another name (alias) for it via (-l|--alias) ALIAS.

Oh, and you should probably tell it who its true masters are by setting HUBOT_AUTH_ADMIN (comma-separated like a boss).

Deployment

Once it is battle ready, send it off for a tour in Heroku:

% heroku create --stack cedar
% git push heroku master
% heroku ps:scale app=1

Upon deployment, it will take orders / commands from Procfile. You can change its name here and tell it how it should behave in a dyno.

You may deploy it elsewhere too.

Management

• Resurrect: heroku restart
• Battle report: heroku logs