Jigawatt NPM | npm.io

jigawatt

Influential's Functional, Promise-based Express Middleware

Installation

npm install jigawatt --save

Using Jigawatt

First, require the module.

const JW = require('jigawatt')

For this walkthrough, we'll also be using Ramda.

const R = require('ramda')

Basic Usage Example

For this example, we have a poll and poll responses stored in a database.

Jigawatt middleware functions, such as getPollResults in this case, are created to sanitize, validate, and/or normalize the incoming data (awesomize), then use that awesomized data to query a database (io), and lastly, format the results to our liking and return the output (transform).

const getPollResults = {
  awesomize: (v) => ({
    pollId : {
      read     : R.path([ 'params', 'pollId' ])
    , sanitize : [ R.toLower ]
    , validate : [ v.required ]
    }
  })

, io: (req, data) => ({
    poll      : db.Poll.findOne({ _id: data.pollId })
  , responses : db.Vote.find({ poll_id: data.pollId })
  })

, transform: (req, data) => ({
    poll    : data.poll.title
  , results : tallyVotes(
                getCity(data.poll)
              , getAnswer(data.responses)
              )
  })
}

Our Jigawatt middleware can then handle a given Express route. Consider the following endpoint:

...

app.get('/poll/:pollId', JW(getPollResults))

Example output:

{
	"poll": "What is your favorite city?",
	"results": [ { "Juneau": 2 }
             , { "Vladivostok": 3 }
             , { "Redding": 1	}
             , { "Wilmington": 0 }
             , { "Galveston": 2	}
             ]
}

This might be a bit overwhelming at first sight, so let's break it down in a bit more detail...

Middleware Structure

Jigawatt Middleware expects at least one of the following three properties:

awesomize

awesomize :: Validator -> Request -> Object

Normalize/Sanitize/Validate an object
For more detailed documentation about awesomize, visit the awesomize documentation

Awesomize has four components, all of which are optional: read -> sanitize -> validate -> normalize

The read component has access to the entire object passed to the awesomize function. Here, we'll use Ramda to target a specific value of the object passed:

awesomize: (v) => ({
  id : {
    read : R.path([ 'order', 'id' ])

...

sanitize is an awesomize component that can manipulate the data before it is validated.

awesomize: (v) => ({
  id : {
    read : R.path([ 'order', 'id' ])
  , sanitize : [ R.drop(2) ]

...

validate is a validator that is passed to our function as v. Awesomize's validate component has a few built-in validator methods such as required, isInt, isFunction, etc...

awesomize: (v) => ({
  id : {
    read : R.path([ 'order', 'id' ])
  , validate : [ v.required, v.isInt ]

...

We can also chain validation methods, as seen above. For more info on awesomize validators, visit the documentation.

As a last note about validate, we can create our own custom validator functions as well:

const isCorrectLength = (str) => R.equals(24, str.length)

...

awesomize: (v) => ({
  pollId : {
    sanitize : [ R.toLower ]
  , validate : [ isCorrectLength ]
  }

...

normalize is the last awesomize component, called after the data has been validated.

awesomize: (v) => ({
  id : {
    read      : R.path([ 'order', 'id' ])
  , validate  : [ v.required, v.isInt ]
  , normalize : [ R.inc ]
  }
})

A complete awesomize function can awesomize more than one value:

awesomize: (v) => ({
  id : {
    read      : R.path([ 'order', 'id' ])
  , validate  : [ v.required, v.isInt ]
  }
, product  : { sanitize : [ R.trim ] }
, customer : { validate : [ v.required ] }
})

io

io :: Request -> Data -> Object

io's primary use is to fetch data using the information passed to it from the awesomize function. In the example below, we have io making two calls to two separate database tables. Once resolved, io will pass the data fetched from the database along to the transform method.

io: (req, data) => ({
    poll      : db.Poll.findOne({ _id: data.pollId })
  , responses : db.Vote.find({ poll_id: data.pollId })
  })

transform

transform :: Request -> Data -> Object

transform is used to structure the incoming data in a unique way. Remember that transform is optional, and if omitted, the Jigawatt middleware simply returns the raw results.

const getAnswer = (arr) => R.compose(
  R.map(R.dec)
, R.pluck('answer')
)(arr)

const getCity = (obj) => R.map(
  R.replace(/\,.*$/, '')
, obj.questions
)

const tallyVotes = (options, responses) => {
  // Tally total vote for a specific city
  return R.map((str) => {
    let ind = R.indexOf(str, options)

    let votes = R.compose(
      R.length
    , R.filter(R.equals(ind))
    )(responses)

    return R.assoc(str, votes, {})
  }, options)
}

...

transform: (req, data) => ({
    poll    : data.poll.title
  , results : tallyVotes(
                getCity(data.poll)
              , getAnswer(data.responses)
              )
  })

Summary

Our Jigawatt middleware was used to take an incoming ID, query two separate database tables for that ID, and format the results to our liking. The final output of our Jigawatt middleware looks like this:

{
	"poll": "What is your favorite city?",
	"results": [ { "Juneau": 2 }
             , { "Vladivostok": 3 }
             , { "Redding": 1	}
             , { "Wilmington": 0 }
             , { "Galveston": 2	}
             ]
}

Extras

Promisify

If you would like to use a Jigawatt middleware as a promise, you can use the JW.promisify method:

const getSingleVote = {
  awesomize: (v) => ...
, io: (req, data) => ...
, transform: (req, data) => ...
}

...

const voteDetails = JW.promisify(getSingleVote)

...
voteDetails(data).then((result) => // do with the data what you will

Pipe

JW.pipe can be used to chain multiple Jigawatt middleware together into a single unit:

const combinedJigawatts = JW.pipe(getPollResults, getSingleVote, ...)

app.get('/poll/:pollId', JW(combinedJigawatts))

Branch

JW.branch should be given a predicate, and two Jigawatt middlewares. If the predicate function returns true, the first Jigawatt middleware is called. If false, the latter is called:

const fetchUserDetails = JW(
  JW.branch(
    isAdmin         // Predicate
  , showAllData     // Called if predicate returns true
  , showMinimalData // Called if predicate returns false
  )
)

Contribute

If you find issues with Jigawatt, we encourage you to open an issue ticket on the issues page. Please open a ticket on the issues page before submitting any pull requests!