Giulio Canti

Posted on Dec 18, 2019 • Edited on Oct 27, 2021

Getting started with fp-ts: Reader

#functional #typescript

The purpose of the Reader monad is to avoid threading arguments through multiple functions in order to only get them where they are needed.

One of the ideas presented here is to use the Reader monad for dependency injection.

The first thing you need to know is that the type Reader<R, A> represents a function (r: R) => A

interface Reader<R, A> {
  (r: R): A
}

where R represents an "environment" needed for the computation (we can "read" from it) and A is the result.

Example

Let's say we have the following piece of code

const f = (b: boolean): string => (b ? 'true' : 'false')

const g = (n: number): string => f(n > 2)

const h = (s: string): string => g(s.length + 1)

console.log(h('foo')) // 'true'

What if we want to internationalise f? Well, we could add an additional parameter

interface Dependencies {
  i18n: {
    true: string
    false: string
  }
}

const f = (b: boolean, deps: Dependencies): string => (b ? deps.i18n.true : deps.i18n.false)

Now we have a problem though, g doesn't compile anymore

const g = (n: number): string => f(n > 2) // error: An argument for 'deps' was not provided

We must add an additional parameter to g as well

const g = (n: number, deps: Dependencies): string => f(n > 2, deps) // ok

We haven't finished yet, now it's h that doesn't compile, we must add an additional parameter to h as well

const h = (s: string, deps: Dependencies): string => g(s.length + 1, deps)

finally we can run h by providing an actual instance of the Dependencies interface

const instance: Dependencies = {
  i18n: {
    true: 'vero',
    false: 'falso'
  }
}

console.log(h('foo', instance)) // 'vero'

As you can see, h and g must have knowledge about f dependencies despite not using them.

Can we improve this part? Yes we can, we can move Dependencies from the parameter list to the return type.

`Reader`

Let's start by rewriting our functions, putting the deps parameter alone

const f = (b: boolean): ((deps: Dependencies) => string) => deps => (b ? deps.i18n.true : deps.i18n.false)

const g = (n: number): ((deps: Dependencies) => string) => f(n > 2)

const h = (s: string): ((deps: Dependencies) => string) => g(s.length + 1)

Note that (deps: Dependencies) => string is just Reader<Dependencies, string>

import { Reader } from 'fp-ts/Reader'

const f = (b: boolean): Reader<Dependencies, string> => deps => (b ? deps.i18n.true : deps.i18n.false)

const g = (n: number): Reader<Dependencies, string> => f(n > 2)

const h = (s: string): Reader<Dependencies, string> => g(s.length + 1)

console.log(h('foo')(instance)) // 'vero'

`ask`

What if we want to also inject the lower bound (2 in our example) in g? Let's add a new field to Dependencies first

export interface Dependencies {
  i18n: {
    true: string
    false: string
  }
  lowerBound: number
}

const instance: Dependencies = {
  i18n: {
    true: 'vero',
    false: 'falso'
  },
  lowerBound: 2
}

Now we can read lowerBound from the environment using ask

import { pipe } from 'fp-ts/function'
import { ask, chain, Reader } from 'fp-ts/Reader'

const g = (n: number): Reader<Dependencies, string> =>
  pipe(
    ask<Dependencies>(),
    chain(deps => f(n > deps.lowerBound))
  )

console.log(h('foo')(instance)) // 'vero'
console.log(h('foo')({ ...instance, lowerBound: 4 })) // 'falso'

p.s.

As a curiosity, Reader's map is (the usual) function composition

import { flow, pipe } from 'fp-ts/function'
import { map } from 'fp-ts/Reader'

const len = (s: string): number => s.length
const double = (n: number): number => n * 2
const gt2 = (n: number): boolean => n > 2

const composition = flow(len, double, gt2)
// equivalent to
const composition = pipe(len, map(double), map(gt2))

Top comments (18)

Pragmatic Maciej • Dec 18 '19 • Edited

Very nice functional concept. But I have small concern if the whole typing of that has/or not a big sense. I mean we really use here a HOF and the whole composition h . g . f result is a function with deps argument

const f = (b: boolean) => (deps: Dependencies) => (b ? deps.i18n.true : deps.i18n.false)
const g = (n: number) => f(n > 2)
const h = (s: string) => g(s.length + 1)

For the second, I see some benefit, but not see doing it like below as a bad thing (yes deps are passed explicitly, but is it any issue?)

const f = (b: boolean) => (deps: Dependencies) => (b ? deps.i18n.true : deps.i18n.false)
const g = (n: number) => (deps: Dependencies) => f(n > deps.lowerBound)(deps) 
// above explicit deps which we can use
const h = (s: string) => g(s.length + 1)

Not get me wrong, just thinking if the whole additional piping and typing is beneficial here.

Giulio Canti • Dec 18 '19

I have small concern if the whole typing of that has/or not a big sense

Well, you are just writing a bunch of functions returning Reader, whether or not you add an explicit type. The point is making the dependencies as implicit as possibile, the key is to put them as a last argument. Note that often people put the dependencies as a first argument

const f = (deps: Dependencies) => (b: boolean) => (b ? deps.i18n.true : deps.i18n.false)

making composition more difficult.

For the second...

all the following gs are equivalent (so feel free to choose the style you prefer)

const g = (n: number) => (deps: Dependencies) => pipe(deps, f(n > deps.lowerBound))

const g = (n: number): Reader<Dependencies, string> => deps => pipe(deps, f(n > deps.lowerBound))

const g = (n: number): Reader<Dependencies, string> => deps => f(n > deps.lowerBound)(deps)

const g = (n: number): Reader<Dependencies, string> =>
  pipe(
    ask<Dependencies>(),
    chain(deps => f(n > deps.lowerBound))
  )

Danilo Spinelli • Dec 28 '19

I'm used to put deps as the first argument since this lets you partially apply f:

const f = (deps: Dependencies) => (b: boolean): string => (...)

Pragmatic Maciej • Dec 18 '19

Yes the whole idea has a lot of sense, and I mean adding argument after not before. Thanks for clarifying!

Michał Kaczanowicz • Dec 18 '19

I'd say there'e no magic in here - in FP all functions have to be referentially transparent at the end of the day i.e. all arguments have to be passed directly! What I think is crucial in here is the conceptual "turning inside-out" (or rather "bottom-up", should I say) - instead of passing props/arguments down multiple levels, we pass a function consuming these props (Reader) multiple levels up, in an almost implicit way. When the Reader reaches "the surface", we can pass its dependencies directly.

Pragmatic Maciej • Dec 18 '19

Yes I see this in that way also. The whole idea here is in dependency argument in the end and not in the beginning.

In my comment I was more addressing the whole additional idealogy over that, like special typings and using some pipe, ask, chain. I see these as only a fog and complexity over really simple but powerful concept.

Giulio Canti • Dec 19 '19

There's a benefit in using the monadic interface though: software is written in a uniform style, regardless of the effect.

You can think of Reader as an effect, muck like Either or Task.

Let's say you have the following snippet

import * as E from 'fp-ts/lib/Either'
import { pipe } from 'fp-ts/lib/pipeable'

declare function f(s: string): E.Either<Error, number>
declare function g(n: number): boolean
declare function h(b: boolean): E.Either<Error, Date>

// composing `f`, `g`, and `h` -------------v---------v-----------v
const result = pipe(E.right('foo'), E.chain(f), E.map(g), E.chain(h))

const pointFreeVersion = flow(f, E.map(g), E.chain(h))

and at some point you must refactor f to

import * as RE from 'fp-ts/lib/ReaderEither'

interface Dependencies {
  foo: string
}

declare function f(s: string): RE.ReaderEither<Dependencies, Error, number>

result and pointFreeVersion must be refactored as well, fortunately you can use ReaderEither's monadic interface

// before 
const result = pipe(E.right('foo'), E.chain(f), E.map(g), E.chain(h))

const pointFreeVersion = flow(f, E.map(g), E.chain(h))

// after
const result = pipe(
  RE.right('foo'),
  RE.chain(f),
  RE.map(g),
  RE.chain(b => RE.fromEither(h(b)))
)

const pointFreeVersion = flow(
  f,
  RE.map(g),
  RE.chain(b => RE.fromEither(h(b)))
)

Michał Kaczanowicz • Dec 19 '19

There's a benefit in using the monadic interface though: software is written in a uniform style, regardless of the effect.

I was just about to say it: I think main benefit is (ironically) readability - by saying explicitly Reader<Dependencies, string> you clearly communicate your intent (assuming others know the concept as well, of course 😄).

Giulio Canti • Dec 19 '19

That's right, you write programs by composing kleisli arrows A -> M<B>, for some effect M.

So Reader (or ReaderEither, ReaderTaskEither, Option, Either, etc...) is just one of the possible effects

massimilianorango • Apr 24 '20

Wow, this looks great! 😍
I'm having trouble understanding how the Reader monad can help in more complex examples though.
Take a look at this for instance: (functions with the "2" suffix are the ones that I would write having dependencies as the first parameter)

export interface Dependencies {
  readonly i18n: {
    readonly true: string
    readonly false: string
  }
  readonly lowerBound: number
}

export interface OtherDependencies {
  readonly semicolon: boolean
}

// Transform:
const transform = (a: string) => (deps: OtherDependencies) =>
  `myString${deps.semicolon ? ':' : ''} ${a}`

const transform2 = (deps: OtherDependencies) => (a: string) =>
  `myString${deps.semicolon ? ':' : ''} ${a}`

// AnotherTransform:
const anotherTransform = (a: string) => (deps: OtherDependencies) =>
  `${a}${deps.semicolon ? ':' : ''} myString`

const anotherTransform2 = (deps: OtherDependencies) => (a: string) =>
  `${a}${deps.semicolon ? ':' : ''} myString`

// F:
const f = (b: boolean) => (deps: Dependencies) =>
  pipe(
    b,
    ifElse(equals(true), always(deps.i18n.true), always(deps.i18n.false))
  )

const f2 = (deps: Dependencies) =>
  ifElse(equals(true), always(deps.i18n.true), always(deps.i18n.false))

// G:
const g = (n: number) => (deps: Dependencies & OtherDependencies) =>
  pipe(
    n > 2,
    f,
    ff => ff(deps),
    s => transform(s)(deps),
    s => anotherTransform(s)(deps)
  )

const g2 = (deps: Dependencies & OtherDependencies) =>
  flow(
    (n: number) => n > 2,
    f2(deps),
    transform2(deps),
    anotherTransform2(deps)
  )

In particular, notice how:

f needs to make b explicit even if could be implicit (like it is in f2)
all the functions in the g pipe need to take the result from the previous computation and pass it as the first parameter to the transform functions, while this is implicit in g2. This leads to less readable code and harder composition. Is there something I'm missing?
Is it generally right to merge all the dependencies of the functions used by g like I did with Dependencies & OtherDependencies?

Giulio Canti • May 11 '20

import { pipe } from 'fp-ts/lib/pipeable' // v2.6
import { chainW, Reader } from 'fp-ts/lib/Reader'

export interface Dependencies {
  readonly i18n: {
    readonly true: string
    readonly false: string
  }
  readonly lowerBound: number
}

export interface OtherDependencies {
  readonly semicolon: boolean
}

declare function transform(a: string): Reader<OtherDependencies, string>
declare function anotherTransform(a: string): Reader<OtherDependencies, string>
declare function f(b: boolean): Reader<Dependencies, string>

const g = (n: number) => pipe(f(n > 2), chainW(transform), chainW(anotherTransform))

massimilianorango • May 11 '20

That's exactly what I needed. I saw you just released chainW in 2.6.0. Great job thanks! :)

Toni Ruottu • Jan 12 '20

Do we have a standard name for (r: Reader<R,A>) => A?

import { Reader } from 'fp-ts/lib/Reader';

type Provider<R> = <A>(r: Reader<R,A>) => A

type Name = string
type Age = number

type Customer = {
  name: Name,
  age: Age,
}

const john: Provider<Customer> = (r) => r({
  name: 'John Doe',
  age: 42,
})

const getAge: Reader<Customer, Age> = ({age}) => age

const ageJohn: Age = john(getAge);

Giulio Canti • Jan 13 '20

Yoneda? Provider<R> is isomorphic to R

λ fp-apprentice • Jul 13 '21

Whoops, my brain is starting to melt, but I like all of this. :)

λ fp-apprentice • Jul 13 '21

Wait, wait, I get it now

Eric Haynes • Jul 22 '23

I know this is an older post, I'm having trouble understanding how this fixes the original issue.

As you can see, h and g must have knowledge about f dependencies despite not using them.

That's true after the refactor as well. Dependencies of f are part of g and h signatures. The way the example is written, the problem is compounded when g has a dependency, because now f depends on parameters of g as well. They could be distinct types, but dependencies of g must be a superset of dependencies of f, e.g.

type FDependencies = { /* ... */ }
type GDependencies = { lowerBound: number } & FDependencies

I understand the desire to curry them so that we can pipe things together, but we haven't lessened the pain of refactoring here at all.

Even if g did not directly call f but was instead piped to it, would it not be better suited to have a flavor of flow or pipe that would automatically coalesce the dependencies?