Coding a Real-World App Using Purely Functional Techniques in 

Ernesto Chero

WHAT IS FUNCTIONAL PROGRAMMING?

Functional programming is based on a simple premise with far-reaching implications: functions that have no side effects.
  • Modifying a variable
  • Modifying a data structure in place
  • Setting a field on an object
  • Throwing an exception or halting with an error
  • Printing to the console or reading user input
  • Reading from or writing to a file
  • Drawing on the screen

What are the side effects? 

 What is a pure function? 

A pure function is side-effect free, plus the result does not depend on anything other than its inputs.

More about pure functions

For a given input, the only effect of a pure function is the output it produces—there are no other effects.

Why pure functions?

Benefits of pure funtions:
  • They're easier to reason about
  • They're easier to combine
  • They're easier to test
  • They're easier to debug
  • They're easier to parallelize
val x = f(a)
val y = g(b)
val z = h(c)
val result = x + y + z

EVERYTHING IS A TYPE

Algebraic Data Types(ADTs)
  • They are a way to structuring data and work well with the pattern matching
  • An algebraic data type is any data that uses the product and sum pattern

Sum Types

  • define a type which can assume different values
  • A sum type is a type that is composed of different possible values and value shapes
  // sum Type [Either, sealed trait, CoProduct]
  sealed trait VehicleType
  case object Car extends VehicleType
  case object Moto extends VehicleType
  // 2 VehicleTypes

  sealed trait Colour
  case object Red extends Colour
  case object Yellow extends Colour
  case object Blue extends Colour
  // 3 Colours

Product Types

  • A product type is essentially a way of sticking multiple values inside of one - a Tuple, or something that’s very similar to one. Case classes are the prototypical product type:
case class Vehicle(vehicleType: VehicleType, colour: Colour, isUsed: Boolean)
  • how many vehicles? if 2 VehicleTypes and 3 Colours and 2 Boolean types
  • result => 12 vehicles

Function Types

  • a function is a type in Scala, it represents an exponential operation
type Like = Colour => Boolean

How many implementations?

2*2*2 = (2)^3 = 8
Red -> {true, false}
Yellow -> {true, false}
Blue -> {true, false}

WHERE CAN I START

  • Category Theory - Basic concepts 
    • Semigroup
    • Monoid
    • Functor
    • Monad
    • Applicative

Cats FP Hierarchy

Semigroup
Monoid
Foldable
Semigroupal
Traverse
Functor
Apply
MonadError
Monad
Applicative
FlapMap
ApplicativeError

CODING A REAL-WORLD APP

Use Case : 

 - Build a Spark Job to process data from different Json files and transform them so it could be sent to an external resource. [for example API, GraphQL, Store on DB, etc]
INPUTS

file: bands.json

{"id":1,"name":"AC/DC","hometown":"Sydney","year":1973}
{"id":0,"name":"Led Zeppelin","hometown":"London","year":1968}
{"id":3,"name":"Metallica","hometown":"Los Angeles","year":1981}
{"id":4,"name":"The Beatles","hometown":"Liverpool","year":1960}

file: guitars.json

{"id":0,"model":"EDS-1275","make":"Gibson","guitarType":"Electric double-necked"}
{"id":5,"model":"Stratocaster","make":"Fender","guitarType":"Electric"}
{"id":1,"model":"SG","make":"Gibson","guitarType":"Electric"}
{"id":2,"model":"914","make":"Taylor","guitarType":"Electric"}
{"id":3,"model":"M-II","make":"ESP","guitarType":"Electric"}


file: guitarsPlayers.json

{"id":0,"name":"Jimmy Page","guitars":[0],"band":0}
{"id":1,"name":"Angus Young","guitars":[1,2],"band":1}
{"id":2,"name":"Eric Clapton","guitars":[1,5],"band":2}
{"id":3,"name":"Kirk Hammett","guitars":[1,3],"band":3}
EXPECTED OUTPUT

file: output.json

{
   "id":1,
   "name":"Angus Young",
   "guitars":[
      {
         "id":1,
         "model":"SG",
         "make":"Gibson",
         "guitarType":"Electric"
      },
      {
         "id":2,
         "model":"914",
         "make":"Taylor",
         "guitarType":"Electric"
      }
   ],
   "band":{
      "id":1,
      "name":"AC/DC",
      "homeTown":"Sydney",
      "year":1973
   }
}

// more records below
...
GRAPHIC REPRESENTATION
 read
 SOURCE
 send
 transform
 Success
 Failure

LET'S GO TO CODE

Defining types
Defining case classes [product type]
Spark Context Provider
A common trait Repository
 Repository Implementation 
Defining API
Transform Data [ filters ]
Transform Data [ joins ]
Transform Data [ transformations ]
Put all together
Thank You!
 
  • https://blog.knoldus.com/scala-best-practices-pure-functions/
  • http://alvinalexander.com/scala/fp-book/benefits-of-pure-functions/
  • https://livebook.manning.com/book/functional-programming-in-scala/chapter-1/
  • https://kubuszok.com/2018/kinds-of-types-in-scala-part-1/
  • https://alvinalexander.com/scala/fp-book/algebraic-data-types-adts-in-scala/

References :

GitHub Repository

https://github.com/ernestochero/sparkRealApp

FUNCTIONAL PROGRAMING

By Ernesto Chero

Made with Slides.com

FUNCTIONAL PROGRAMING

  • 66

More from Ernesto Chero