Safe Haskell | None |
---|---|
Language | Haskell98 |
Sodium Reactive Programming (FRP) system.
See the examples directory for test cases and examples.
Some functions are pure, and others need to run under the Reactive
monad via
sync
. An Event
(Reactive
a) can be flattened
to an Event
a using the execute
primitive.
In addition to the explicit functions in the language, note that you can use
- Functor on
Event
andBehavior
- Applicative on
behaviour
, e.g.let bsum = (+) <$> ba <*> bb
- Applicative
pure
is used to give a constantBehavior
. - A Monoid instance on
Event
wheremempty
=never
andmappend
=merge
- Recursive do (using the DoRec language extension) to make state loops with the
rec
keyword. - Data.Traversable.
sequenceA
is useful to convert [Behavior a] into Behavior [a].
Here's an example of recursive do to write state-keeping loops. Note that
all hold
s are delayed, so snapshot
will capture the old value of the state s.
{-# LANGUAGE DoRec #-} -- | Accumulate state changes given in the input event. accum :: Context r => a -> Event r (a -> a) -> Reactive r (Behavior r a) accum z efa = do rec s <- hold z $ snapshot ($) efa s return s
- data Plain
- type Reactive = Reactive Plain
- sync :: Reactive a -> IO a
- newEvent :: Reactive (Event a, a -> Reactive ())
- newBehavior :: a -> Reactive (Behavior a, a -> Reactive ())
- newBehaviour :: a -> Reactive (Behavior a, a -> Reactive ())
- listen :: Event a -> (a -> IO ()) -> Reactive (IO ())
- type Event = Event Plain
- type Behavior = Behavior Plain
- type Behaviour = Behavior Plain
- never :: Event a
- merge :: Event a -> Event a -> Event a
- filterJust :: Event (Maybe a) -> Event a
- hold :: a -> Event a -> Reactive (Behavior a)
- updates :: Behavior a -> Event a
- value :: Behavior a -> Event a
- snapshot :: (a -> b -> c) -> Event a -> Behavior b -> Event c
- switchE :: Behavior (Event a) -> Event a
- switch :: Behavior (Behavior a) -> Reactive (Behavior a)
- execute :: Event (Reactive a) -> Event a
- sample :: Behavior a -> Reactive a
- coalesce :: (a -> a -> a) -> Event a -> Event a
- once :: Event a -> Event a
- split :: Event [a] -> Event a
- mergeWith :: (a -> a -> a) -> Event a -> Event a -> Event a
- filterE :: (a -> Bool) -> Event a -> Event a
- gate :: Event a -> Behavior Bool -> Event a
- collectE :: (a -> s -> (b, s)) -> s -> Event a -> Reactive (Event b)
- collect :: (a -> s -> (b, s)) -> s -> Behavior a -> Reactive (Behavior b)
- accum :: a -> Event (a -> a) -> Reactive (Behavior a)
- changes :: Behavior a -> Event a
- values :: Behavior a -> Event a
- snapshotWith :: (a -> b -> c) -> Event a -> Behavior b -> Event c
- count :: Event a -> Reactive (Behavior Int)
Documentation
Phantom type for use with Context
type class.
Context Plain | |
Monad (Reactive Plain) | |
Functor (Reactive Plain) | |
Functor (Event Plain) | |
Functor (Behavior Plain) | |
MonadFix (Reactive Plain) | |
Applicative (Reactive Plain) | |
Applicative (Behavior Plain) | |
data Reactive Plain = Reactive (StateT ReactiveState IO a) | |
data Event Plain = Event {
| |
data Behavior Plain = Behavior {
|
Running FRP code
sync :: Reactive a -> IO a Source
Execute the specified Reactive
within a new transaction, blocking the caller
until all resulting processing is complete and all callbacks have been called.
This operation is thread-safe, so it may be called from any thread.
State changes to hold
values occur after processing of the transaction is complete.
newEvent :: Reactive (Event a, a -> Reactive ()) Source
Returns an event, and a push action for pushing a value into the event.
Create a new Behavior
along with an action to push changes into it.
American spelling.
Create a new Behavior
along with an action to push changes into it.
British spelling.
listen :: Event a -> (a -> IO ()) -> Reactive (IO ()) Source
Listen for firings of this event. The returned IO ()
is an IO action
that unregisters the listener. This is the observer pattern.
To listen to a Behavior
use listen (value b) handler
or
listen (updates b) handler
NOTE: The callback is called with the transaction held, so you cannot
use sync
inside a listener. You can delegate to another thread and have
that start the new transaction. If you want to do more processing in
the same transction, then you can use listenTrans
but this is discouraged unless you really need to write a new primitive.
FRP core language
type Event = Event Plain Source
A stream of events. The individual firings of events are called 'event occurrences'.
merge :: Event a -> Event a -> Event a Source
Merge two streams of events of the same type.
In the case where two event occurrences are simultaneous (i.e. both within the same transaction), both will be delivered in the same transaction. If the event firings are ordered for some reason, then their ordering is retained. In many common cases the ordering will be undefined.
filterJust :: Event (Maybe a) -> Event a Source
Unwrap Just values, and discard event occurrences with Nothing values.
hold :: a -> Event a -> Reactive (Behavior a) Source
Create a behavior with the specified initial value, that gets updated by the values coming through the event. The 'current value' of the behavior is notionally the value as it was 'at the start of the transaction'. That is, state updates caused by event firings get processed at the end of the transaction.
value :: Behavior a -> Event a Source
An event that is guaranteed to fire once when you listen to it, giving
the current value of the behavior, and thereafter behaves like updates
,
firing for each update to the behavior's value.
snapshot :: (a -> b -> c) -> Event a -> Behavior b -> Event c Source
Sample the behavior at the time of the event firing. Note that the 'current value'
of the behavior that's sampled is the value as at the start of the transaction
before any state changes of the current transaction are applied through hold
s.
switchE :: Behavior (Event a) -> Event a Source
Unwrap an event inside a behavior to give a time-varying event implementation.
switch :: Behavior (Behavior a) -> Reactive (Behavior a) Source
Unwrap a behavior inside another behavior to give a time-varying behavior implementation.
execute :: Event (Reactive a) -> Event a Source
Execute the specified Reactive
action inside an event.
coalesce :: (a -> a -> a) -> Event a -> Event a Source
If there's more than one firing in a single transaction, combine them into one using the specified combining function.
If the event firings are ordered, then the first will appear at the left input of the combining function. In most common cases it's best not to make any assumptions about the ordering, and the combining function would ideally be commutative.
split :: Event [a] -> Event a Source
Take each list item and put it into a new transaction of its own.
An example use case of this might be a situation where we are splitting a block of input data into frames. We obviously want each frame to have its own transaction so that state is updated separately each frame.
Derived FRP functions
mergeWith :: (a -> a -> a) -> Event a -> Event a -> Event a Source
Merge two streams of events of the same type, combining simultaneous event occurrences.
In the case where multiple event occurrences are simultaneous (i.e. all within the same transaction), they are combined using the supplied function. The output event is guaranteed not to have more than one event occurrence per transaction.
The combine function should be commutative, because simultaneous events should be considered to be order-agnostic.
filterE :: (a -> Bool) -> Event a -> Event a Source
Only keep event occurrences for which the predicate is true.
gate :: Event a -> Behavior Bool -> Event a Source
Let event occurrences through only when the behavior's value is True. Note that the behavior's value is as it was at the start of the transaction, that is, no state changes from the current transaction are taken into account.
collectE :: (a -> s -> (b, s)) -> s -> Event a -> Reactive (Event b) Source
Transform an event with a generalized state loop (a mealy machine). The function is passed the input and the old state and returns the new state and output value.
collect :: (a -> s -> (b, s)) -> s -> Behavior a -> Reactive (Behavior b) Source
Transform a behavior with a generalized state loop (a mealy machine). The function is passed the input and the old state and returns the new state and output value.
accum :: a -> Event (a -> a) -> Reactive (Behavior a) Source
Accumulate state changes given in the input event.
Deprecated
snapshotWith :: (a -> b -> c) -> Event a -> Behavior b -> Event c Source