hacknight5

Meet Misty, your new best friend / robot overlord

She's the greatest. Someone from marketing will hopefully fill this in shortly

About this repo

The goal is to provide a REST api that is somewhat similar to Misty's request/response cycle, and then provide an interface to "see" those actions happening on your virtual Misty via the browser.

Obviously not everything can be implemented / virtualized, and the first attempts are pretty basic. But, it's a simple model to contribute to / code against, allowing you to poke at a Misty locally even if your new best friend is out conquering the resistance foreign lands.

The code is separated into two npm-powered projects:

• misty - the Vue.js powered web interface to your virtualized Misty
• server - the stateful Node.js dev server that emulates Misty's API and communicates with the browser-Misty via websockets

You can ignore most of server's code, even if you want to implement missing / stubbed Misty features in the emulator. Misty's emulated commands can be found in the corresponding server/commands/[category]/[command].js file, and are bootstrapped with a context of the environment. Said context has some useful items on it, including the following:

• state - the current state of the virtualized Misty. See the top section of misty-state.js for an idea of what status she could be in.
• _ - we bundled in lodash as _ for helpful methods
• busEvent - API Commands leverage an event bus, and this is the event emitted when a command should be executed. Don't dig to deep (it's ugly still) but know that you've got the following args:
  • req - The HTTP request representing the Command
  • res - The HTTP response for the Command - it's your job to implement it correctly according to docs :)
  • args - The parsed arguments, after validation (more on that below)
• helpers - We reserved this namespace for utils to help. So far, it contains
  • timers - Because Misty's state needs to be simple JSON-friendly objects, we can't put Node's Timer construct on the state. And since setTimeout doesn't return numbers, we cooked this stupid utility up for easily tracking actions while they're happening. Instead of setTimeout, setInterval, clearXYZ, use timers.timeout(..), timers.interval(..), and timers.clear(id)

An example of a simple Command is this Drive implementation that just sets her velocities accordingly.

module.exports = ctx => {
  const {
    _,
    state,
    busEvent: { req, res, args }
  } = ctx

  state.drive = {
    linearVelocity: 0,
    angularVelocity: 0,
    ...state.drive,
    ..._.pick(args, 'linearVelocity,angularVelocity')
  }

  res.send({ message: 'ok' })
}

IMPORTANT NOTES

Some important things to note:

• If you update the misty-state object, the updated object will be sent over the websocket automagically.
• When the state comes to Vue, it's automatically bound for the UI to update
• This means, hopefully, any dev can be isolated to <command>.js and misty-state.js
• Because you have the websocket open to any clients separate from the rest API request, you can (and should) send the Rest API response ASAP while updating Misty's state afterwards via a timeout / interval / loop / carrier pidgeon

Getting Started

In terminal one:

• cd into server
• run npm install and wait for half the internet to download
• run npm start to start your server. You should see Listening on port 3000 if all goes to plan

In terminal two:

• cd into misty
• run npm install and wait for the other half of the internet to download
• run npm start to start your server. You should see the following if all goes to plan ``` App running at:
• Local: http://localhost:8080/
• Network: http://xxx.yyy.zzz.fml:8080/

At this point, hit http://localhost:8080 and you should be greated by the virtual Misty. We're still polishing the UI, but you can see what her state is below her simple picture. We're also working on a SVG-based render, which volunteers are welcome for.