Functional Programming: Functors

This content originally appeared on Bits and Pieces - Medium and was authored by Ahmad M. Hawwash

Functional Programming: Part 5 - Functors

This article is a part of a series that talks about “Functional Programming”

In the previous article in this series we discussed a very useful pattern; Composition and how we can use it to compose bigger systems out of smaller granular units. In this article we’ll be talking about Functors!

Table of content

• Functors
• A Problem
• The Maybe Functor
• Solving The Problem Using Maybe Functor
• Functors Laws
• Why To Use Functors?
• Why Don’t We See More Functors In Our Codebases?
• Quick Note
• Conclusion

Functors

(Note: we are talking about functors, not functions)

What is a Functor?

A functor is just a

1. Wrapper over a value
2. That gives us a mapping interface
3. And obeys the functor’s laws (we’ll talk about them later)

Examples on functors we use daily

1. Array
2. Promise

Yes Array and Promise are functors, but how?

Let’s reflect on the functor definition…

Array is a:

1. Wrapper over a list of items
2. That gives us map as a mapping interface
3. And obeys the functor’s laws

Example:

[1, 2, 3]   // the wrapped value
.map        // the mapping interface
(x => x * 2)

Promise is a:

1. Wrapper over any JS data type value
2. That gives us then as a mapping interface
3. And obeys the functor’s laws

Example:

const promise = new Promise((resolve, reject) => {
  resolve(
    { data: "Any JS type" } // the wrapped value
  ) 
});

promise
.then                       // the mapping interface
(response => console.log(response));

What’s the relationship between the Array (and Promise) and the Functor?

A Functor is a design pattern, while Array and Promise are data types.

Why am I telling you that Array and Promise are Functors?

To shutdown the fear of the idea of functors.

They are easy to understand, yet they have a powerful concept. And we use them daily without consciously knowing!

Why to use functors?

Great question! Please follow along…

A Problem

After talking a bit about functors and connecting them with our daily use, it’d be wise to dig deeper and build our own. In order to understand more the idea of a functor.

Let’s face a problem first…

Take a look at this piece of data

{
  products: [
    {
      name: 'building stuff',
      type: 'book',
      price: 22, 
      discount: 20
    },
  ]
}

Our task

To find the final price of the first product with discount. While if for any reason we encounter bad data, the fallback should be "Nothing” as string.

Algorithmic steps

1. Find first discounted product
2. Apply discount
3. Keep checking for data invalidity. Where if data is invalid for any reason, we return “Nothing” as a fallback

First try (the traditional way)

const isProductWithDiscount  = product  => !isNaN(product.discount)
const findFirstDiscounted    = products => products.find(isProductWithDiscount)
const calcPriceAfterDiscount = product  => product.price - product.discount;

const findFinalPrice = (data, fallbackValue) => {
  if(!data || !data.products) return fallbackValue;

  const discountedProd = findFirstDiscounted(data.products);
  if(!discountedProd) return fallbackValue;

  return calcPriceAfterDiscount(discountedProd);
}

findFinalPrice(data, "Nothing") // 2

Notes on our first try

The good things:

1. Granular logical units (isProductWithDiscount, findFirstDiscounted and calcPriceAfterDiscount)
2. We defended our logic against data invalidity

Things we can improve:

1. We’re defending too much. (Defensive programming is a must in any resilient software, but too much is too much. In our code, 50% of findFinalPrice function’s body is checking for data invalidity, which is considered too much defence)
2. fallbackValue is almost everywhere

Why are we concerned about these improvements? Because they are mentally consuming maintainers. Thus directly impacting DX (Developer Experience) negatively.

So there’s still room for improvements, right?

Let’s analyse the code and try to come up with a solution…

Code structure analysis: The parts we aim to improve are forming a pattern (defence and fallback) and they are actually intact and atomic! And that’s good!

Solution (Taking in mind analysis above): We should be able to abstract that pattern into a wrapper that could handle these corner cases for us.

Result: The wrapper will take care of the corner cases. While we’ll only have to take care of the business logic!

If we’re able to implement this, that’d be great indeed!

The Maybe Functor

Reminder: A functor is just a wrapper over a value and gives us a mapping interface (i.e. map) and obeys some laws.

As we discussed earlier. We only want a wrapper that abstracts away handling data (either bad or good).

So, the role of Maybe functor, is to wrap our data (potentially bad data), and handle the corner cases for us. How is that? Follow along…

Maybe Functor Implementation

function Maybe(value) {
  const isNothing = () => value === null || value === undefined;
  const map = (fn) => isNothing() ? Maybe() : Maybe(fn(value));
  const getValueOrFallback = 
    (fallbackValue) => isNothing() ? fallbackValue : value;

  return {
    map,
    getValueOrFallback,
  };
}

Implementation explanation:

1. isNothing: checks if the wrapped value in Maybe functor is invalid.
2. map: the mapping interface to our value wrapper, where we use this to apply our business logic lambdas to the wrapped value. (Note: that map returns new value in another Maybe instance, so we can keep doing .map().map().map…)
3. getValueOrFallback: returns the wrapped value or fallbacks to fallbackValue.

Let’s use Maybe:

• With good data:

Maybe('Ahmad')
  .map(x => x.substring(1))
  .getValueOrFallback('fallback') // 'hmad'

• With bad data:

Maybe(null)
  .map(x => x.substring(1))       // will not be executed
  .getValueOrFallback('fallback') // 'fallback'

Sounds good so far! Maybe functor cared for the corner cases for us! And “aborted” execution when data was invalid. While we only cared for business logic!

It looks like we achieved our goal! Let’s put it to work…

Solving the previous problem using Maybe functor

Ok, let’s try to solve the products problem using Maybe functor

const isProductWithDiscount  = product  => !isNaN(product.discount)
const findFirstDiscounted    = products => products.find(isProductWithDiscount)
const calcPriceAfterDiscount = product  => product.price - product.discount;

Maybe(data)
  .map((x) => x.products)
  .map(findFirstDiscounted)
  .map(calcPriceAfterDiscount)
  .getValueOrFallback('Nothing') // 2

We were able to improve the previous code using Maybe functor, where

1. We no more defend the code ourselves, but the Maybe functor does that for us.
2. We configured the fallbackValue only once.

Reflecting Maybe functor on functor’s definition

Maybe functor is just a

1. Wrapper over any JS data type value
2. That gives us map as a mapping interface
3. And obeys the functor’s laws

Functors Laws

• Identity law

When performing the mapping operation, if the values in the functor are mapped to themselves, the result will be an unmodified functor.

Example:

const w1 = Maybe(value)
const w2 = Maybe(value).map(v => v)
// w1 and w2 are equivalent

• Composition

If two sequential mapping operations are performed one after the other using two functions, the result should be the same as a single mapping operation with one function that is equivalent to applying the first function to the result of the second.

Example:

const w1 = Maybe(value).map(v => f(g(v))) 
const w2 = Maybe(value).map(v => g(v)).map(v => f(v))
// w1 and w2 are equivalent

Why To Use Functors?

1. **Abstraction of function application
2. Empowering functional composition
3. Reducing amount of defensive code (in case if we used Maybe functor)
4. Cleaner code structure (Did you notice how linear the logic is now? Amazing!)
5. Variables are more explicit about what we expect (That Maybe models a value that may or may not be present. That it models a “value” or “nothing”)

**What do we mean by “Abstraction of function application”?

That we pass a transformer function (i.e. x => x.products) to the mapping interface (i.e. map) of the wrapper (i.e. Maybe) and it knows how to take care of itself (by its internal implementation).

We don’t care about the wrapper’s implementation details it has inside (implementation details are hidden), and yet we know how to use the wrapper (Array or Promise) by using their mapping interfaces.

And that’s actually crucially important in the programming world. Where that lowers the bar on how much we -as programmers- need to understand in order to be able to get stuff done. They can be implemented in any language that supports higher-order functions (which is most of them these days). And that creates a language-agnostic vocabulary.

Why Don’t We See More Functors In Our Codebases?

Simply, because we’re not used to. Before .map (and .then) we were used to mutate arrays or loop through them manually. But once we discovered .map we started adapting it as the new transformation tool! I hope after understanding the value of functors, we start introducing them more into our daily coding routines as a very usual thing!

Quick Note

Maybe functor is just an example on functors. It’s not the only functor we have, there are more functors out there and for different purposes. We talked about the simplest here (The Maybe functor), so we can get a bit closer to the idea of functors!

Conclusion

Functor design pattern is a simple — yet very powerful pattern. We use it daily in different data types without knowing. Hopefully, we can recognise and appreciate functors a bit more and give them more space in our codebase, because they’ll clean code and give us more power; Due to the composability nature they have and abstractions they offer.

At the end, a functor is just a wrapper over a value that gives us a mapping interface, and obeys some laws.

Thanks a lot for taking time reading through this article ❤️ I’m cooking the next ones in the series. Please let me know what you think in the comments about this article or the series.

This is an article in a series of articles talking about Functional Programming

In this series of articles we cover the main concepts of Functional Programming. At the end of the series, you’ll be able to tackle problems in a more functional approach.

This series discusses:

0. A Brief Comparison Between Programming Paradigms

1. First Class functions
2. Pure functions
3. Currying
4. Composition
5. Functors (this article)
6. Monads

Resources

1. https://blog.bitsrc.io/functional-programming-composition-2e9b863d8bcb
2. https://wiki.haskell.org/Functor
3. https://marmelab.com/blog/2018/09/26/functional-programming-3-functor-redone.html
4. https://medium.com/swlh/beyond-monad-the-applicative-interface-9685d233c1b3
5. https://jrsinclair.com/articles/2016/marvellously-mysterious-javascript-maybe-monad/
6. https://blog.ploeh.dk/2018/03/26/the-maybe-functor/
7. https://jrsinclair.com/articles/2018/how-to-deal-with-dirty-side-effects-in-your-pure-functional-javascript/
8. https://stackoverflow.com/questions/74332025/what-does-abstraction-of-function-application-mean/
9. https://mostly-adequate.gitbook.io/mostly-adequate-guide/ch08

Functional Programming: Functors was originally published in Bits and Pieces on Medium, where people are continuing the conversation by highlighting and responding to this story.

This content originally appeared on Bits and Pieces - Medium and was authored by Ahmad M. Hawwash

Print Share Comment Cite Upload Translate Updates

APA

Ahmad M. Hawwash | Sciencx (2022-12-08T11:11:05+00:00) Functional Programming: Functors. Retrieved from https://www.scien.cx/2022/12/08/functional-programming-functors/

MLA

" » Functional Programming: Functors." Ahmad M. Hawwash | Sciencx - Thursday December 8, 2022, https://www.scien.cx/2022/12/08/functional-programming-functors/

HARVARD

Ahmad M. Hawwash | Sciencx Thursday December 8, 2022 » Functional Programming: Functors., viewed ,<https://www.scien.cx/2022/12/08/functional-programming-functors/>

VANCOUVER

Ahmad M. Hawwash | Sciencx - » Functional Programming: Functors. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2022/12/08/functional-programming-functors/

CHICAGO

" » Functional Programming: Functors." Ahmad M. Hawwash | Sciencx - Accessed . https://www.scien.cx/2022/12/08/functional-programming-functors/

IEEE

" » Functional Programming: Functors." Ahmad M. Hawwash | Sciencx [Online]. Available: https://www.scien.cx/2022/12/08/functional-programming-functors/. [Accessed: ]

rf:citation

» Functional Programming: Functors | Ahmad M. Hawwash | Sciencx | https://www.scien.cx/2022/12/08/functional-programming-functors/ |

Please log in to upload a file.

There are no updates yet.
Click the Upload button above to add an update.

You must be logged in to translate posts. Please log in or register.