Working our way backwards from solution to problem, we define an applicative functor, then use it to apply a function of multiple arguments.
For example we have this line of code:
const res = Box(x => x +1).ap(Box(2))// Box(3);
We want to use a funciton ‘ap‘ (apply) on Box. And x will be 2.
To define ‘ap‘ function.
const Box = x => ({ chain: f => f(x), ap: other => other.map(x), map: f => Box(f(x)), fold: f => f(x), inspect: () => `Box(${x})` })
So ‘
Box(x => x +1).ap(Box(2))
‘
Can be translated to:
Box(2) => Box(2).map(x => x + 1);
This can be useful when apply curry function:
const res = Box(x => y => x + y).ap(Box(1)).ap(Box(2)); console.log(res.inspect()); //Box(3)
after apply .ap(Box(1)), it becomes to:
Box(y => 1 +y).ap(Box(2))
after apply .ap(Box(2)), it becomes to:
Box(1 +2 )
It ends up, we have a function and continue to using ‘ap‘:
const add = x => y => x + y; const res = Box(add).ap(Box(1)).ap(Box(2));
This partten is called click-functor!
The rule is:
F(val).map(fn) === F(fn).ap(F(val))
For example now we have:
const liftA2 = (fn, Fx, Fy) => F(fn).ap(Fx).ap(Fy);
The problem is we don‘t know what ‘F‘ it is here...
So what we can do is transform accorind to the rule we have:
const liftA2 = (fn, Fx, Fy) => Fx.map(fn).ap(Fy)
Therefore we don‘t need to memtion any Functor.
Example:
const res2 = liftA2(add, Box(1), Box(2)); console.log(res2.inspect()); //Box(3)
Applicate Functor is really good to work with Async functor, because async by natural, data arrives different time:
const add = x => y => z=> x + y + z; const addAsyncNumbers = liftA3(add); const res = addAsyncNumbers( Async.of(5), Async((_, res) => { setTimeout(() => { console.log(‘resolve 2‘); res(2) }, 500) }), Async((_, res) => { setTimeout(() => { console.log(‘resolve 3‘); res(3) }, 600) })); res.fork(e => console.error(e), x => console.log(‘async‘, x)) // 10