Awesome

florida-keys

Pure functional accessor factories in TypeScript.

Install

npm install florida-keys

Testing

npm install
npm test

Summary of API

// import a factory function and accesssory functions from florida-keys
import { florida, and } from 'florida-keys';

// This is the type of data we want to access, filter and sort
interface Person {
  first: string;
  last: string;
  dob: Date;
}

// create a florida-keys accessor for a Person type
const f = florida<Person>();

// define some accessor functions
const first = f.k('first');
const dob = f.k('dob');
const fullname = f.g(d => `${d.last}, ${d.first}`);

// define some conditions
const gNames = first.i(0).eq('G');
const nineties = and(
  dob.gte(new Date('1/1/1990')),
  dob.lt(new Date('1/1/2000'))
);

// define a sorting function
const dobSort = dob.asc();

// The data
const people: Person[] = [
  { first: 'Gil', last: 'Wiggins', dob: new Date('8/1/1992') },
  ...
];

// returns an array
people
  .filter(and(gNames, nineties).$) // people with "G" names born in the 90's
  .sort(dobSort.$)                 // sorted by date of birth
  .map(fullname.$);                // their fullnames

Accessors

Example

const f = florida<Person>();
const first = f.k('first');

people
  .map(first.$);  // Returns a list of first names

Operators

Example

const f = florida<Person>();
const firstname = f.k('firstname');
const gills = firstname.eq('Gil');

people
  .filter(gills.$);  // returns a list of Gil's

Logicals

Example

const f = florida<Person>();
const dob = f.k('dob');
const nineties = and(
  dob.gte(new Date('1/1/1990')),
  dob.lt(new Date('1/1/2000'))
);

people
  .filter(nineties.$);  // returns a list of people born in the nineties

Sorting

Example

const f = florida<Person>();
const dobSort = f.k('dob').asc();

people
  .sort(dobSort.$);  // returns a list of people sorted by date of birth in ascending order

Why?

In TypeScript we often write accessor functions like this:

function(d: Person) { return d.first; }

or using arrow functions

(d: Person) => d.first

This simple function returns the value of the first key when an object is pass to it. For example in the map function:

people.map((d: Person) => d.first);

This is lightweight, simple, and readable. There is nothing wrong with it. Sometimes, however, in order to avoid repeating ourselves so we create a reusable accessor function like this:

const firstname = (d: Person) => d.first;
people.map(firstname);

Now imagine the object also has a dob key whose values are date objects. We may want to filter like this:

const firstname = (d: Person) => d.first;
const dob = (d: Person) => d.dob;
const dobFilter = (d: Person) => d.dob >= new Date('1/1/1990');

people
  .filter(dobFilter)
  .map(firstname);

However, this has a couple of drawbacks. First of all, you will need to create a new filter every time the date changes; also the Date constructor is called for every element in the people array. A better approach is an accessor function factory:

const createDobFilter = (date: Date) => d => d.dob >= date;
const dobFilter = createDobFilter(new Date('1/1/1990'));

people
  .filter(dobFilter)
  .map(firstname);

It's a little ugly but here the Date constructor is only called once and the createDobFilter function returns the accessor. An new accessor can be created any time by calling createDobFilter

Now what if we want to filter between two dates. We can do modify our accessor factory:

const createDobFilter = (a: Date, b: Date) => d => d.dob >= a && d.dob < b;

but let's say that you have multidimensional data where dates a and b change independently. You might be tempted to do something like this:

const createDobGteFilter = (date: Date) => d => d.dob >= date;
const createDobLteFilter = (date: Date) => d => d.dob < date;

const dobGteFilter = createDobGteFilter(new Date('1/1/1990'));
const dobLteFilter = createDobLteFilter(new Date('1/1/2000'));

const values = data
  .filter(dobGteFilter)
  .filter(dobLteFilter)
  .map(firstname);

Then let's add sorting by dob and returning the person's full name.

const createDobGteFilter = (date: Date) => d => d.year >= date;
const createDobLteFilter = (date: Date) => d => d.year < date;

const dob = (d: Person) => d.dob;
const dobGteFilter = createDobGteFilter(new Date('1/1/1990'));
const dobLteFilter = createDobLteFilter(new Date('1/1/2000'));
const fullname = (d: Person) => `${d.last}, ${d.first}`;

const dobSort = (a: Person, b: Person) => a.dob < b.dob ? -1 : a.dob > b.dob ? 1 : 0;

people
  .filter(dobGteFilter)
  .filter(dobLteFilter)
  .sort(dobSort)
  .map(fullname);

Ok, no we're getting ridiculous. The Date constructor is not that expensive. But you can imagine a situation where the values for filters could be very expensive. For example based on aggregated statistics or reading from the DOM, or, perhaps, we are interating over a lot of data

Ok, at this point let me introduce florida-keys. florida-keys has simply a set of shortcuts for all this. For example:

const f = florida<Person>();
const firstname = f.k('first');
people.map(firstname.$);

The value returned from f.k(key).$ in this case is simply the typed accessor function function(d: T) { return d[key]; }.

Interesting. How about this:

const f = florida<Person>();
const firstname = f.k('first');
const dobFilter = f.k('dob').gte(new Date('1/1/1990'));
people
  .filter(dobFilter.$)
  .map(firstname.$);

f.k(key).gte(somevalue) is essentially a shortcut for function(d: T) { return d['key'] >= somevalue; }.

It gets better:

const f = florida<Person>();
const firstname = f.k('first');
const dob = f.k('dob');

const nineties = and(
  dob.gte(new Date('1/1/1990')),
  dob.lt(new Date('1/1/2000'))
);

people
  .filter(nineties.$)
  .map(firstname.$);

or how about this:

const f = florida<Person>();
const firstname = f.k('first');
const dob = f.k('dob');
const fullname = f.g(d => `${d.last}, ${d.first}`);

const gNames = firstname.i(0).eq('G');
const nineties = and(
  dob.gte(new Date('1/1/1990')),
  dob.lt(new Date('1/1/2000'))
);

const filter = and(gNames, nineties);
const sort = dob.asc();

people
  .filter(filter.$)
  .sort(sort.$)
  .map(fullname.$);

Pretty neat?

Acknowledgments

Also see utatti/lens.ts

License

Copyright (c) 2014+ Jayson Harshbarger MIT