main :: IO ()
main = runJSorWarp 8080 $ (1)
simple (2)
runParDiff (3)
() (4)
(const "hello world") (5)
getBody (6)
Calculator
Here we will build a simple calculator, following pioneers of the space Luite Stegemann, Ryan Trinkle, and Ivan Perez.
Hello World
Let’s start with a "Hello World" app.
1 | This function either runs a server or generates a JavaScript-enabled web
page. If built with GHC, it will run a server on port 8080 . If built with
GHCjs, it will just start the application as a normal JavaScript file. |
2 | This is a wrapper around the shpadoinkle function. |
3 | No matter how simple the app, you must still choose a backend explicitly. This function chooses the ParDiff backend. |
4 | The initial state. Because this is a "hello world" example, it’s just () . |
5 | This is the view. Because we don’t care about the state (() ) we’re using
const here. "hello world" is making use of the OverloadedStrings
pragma. |
6 | This the DOM node on the page that will hold our application. |
Addition
Now that we have our hello world, let’s add some interactivity and perform addition on the part of the user:
To start, we will need a simple single number input:
num :: Int -> Html m Int
num x = input'
[ value . pack $ show x (1)
, onInput (const . fromMaybe 0 . readMay . unpack) (2)
]
1 | Set the value of the input to the current state of the application. |
2 | When an "input" event occurs, update the state with the provided
function. |
Now we can use our single number input in our view to add two numbers.
view :: (Int, Int) -> Html m (Int, Int)
view (l,r) = (1)
div_
[ liftC (,r) fst $ num l (2)
, " + "
, liftC (l,) snd $ num r (3)
, text $ " = " <> pack (show $ l + r) (4)
]
1 | For now we can just use a tuple to house our two numbers. |
2 | We use the num component, rendering it with l . |
3 | We use the num component, rendering it with r . |
4 | We display the result of l + r to the user. |
Heterogeneous Composition
We are using
To lift |
Selectable Operations
Now that we have some inputs that can perform addition, let the user select common operations. We start by making a new ADT:
data Operation
= Addition
| Subtraction
| Multiplication
| Division
deriving (Eq, Show, Read, Enum, Bounded)
We will need some functions to get human readable display, as well as mapping to functions:
opFunction :: Operation -> (Int -> Int -> Int)
opText :: Operation -> Text
We will use a traditional Html <select>
element to let the user pick the operation.
opSelect :: Html m Operation
opSelect = select [ onOption $ const . read . unpack ] (1)
$ opOption <$> [ minBound .. maxBound ]
where
opOption o = option
[ value . pack $ show o ] (2)
[ text $ opText o ]
1 | When the select element changes, we read the value and use that as the model. |
2 | Values are just the operation applied to show , so it matches when we read . |
Great! Now we have three members of our model, and so it’s time for an actual data type.
data Model = Model
{ operation :: Operation
, left :: Int
, right :: Int
} deriving (Eq, Show)
We will use the same technique with liftC
to compose these components together:
view :: Model -> Html m Model
view model = div_
[ liftC (\l m -> m { left = l }) left $ num (left model)
, liftC (\o m -> m { operation = o }) operation $ opSelect
, liftC (\r m -> m { right = r }) right $ num (right model)
, text $ " = " <> pack (show $ opFunction
(operation model) (left model) (right model))
]
That’s it! The user can select an operation, enter two numbers, and see a result rendered.
Now if you are looking at the above code and starting to twitch because you are an optics fiend, that is the right instinct.
You can see the final code here running below:
Next we emulate a real-world immediate execution calculator in part 3.