Copyright | (c) 2015 Schell Scivally |
---|---|
License | MIT |
Maintainer | Schell Scivally <schell.scivally@synapsegroup.com> |
Safe Haskell | None |
Language | Haskell2010 |
Using splines we can easily create continuous value streams from multiple piecewise event streams. A spline is a monadic layer on top of event streams which are only continuous over a certain domain. The idea is that we use do notation to "run an event stream" from which we will consume produced values. Once the event stream inhibits the computation completes and returns a result value. That result value is then used to determine the next spline in the sequence.
A spline can be converted back into a value stream using execSpline
or
execSplineT
. This allows us to build long, complex, sequential behaviors
using familiar notation.
- type Spline a b c = SplineT a b Identity c
- data SplineT a b m c
- runSplineT :: (Applicative m, Monad m) => SplineT a b m c -> VarT m a (Step (Event b) c)
- scanSpline :: (Applicative m, Monad m) => SplineT a b m c -> [a] -> m [(Event b, Event c)]
- fromEvents :: (Applicative m, Monad m) => b -> VarT m a (Event b) -> SplineT a b m b
- outputStream :: (Applicative m, Monad m) => b -> SplineT a b m c -> VarT m a b
- resultStream :: (Applicative m, Monad m) => SplineT a b m c -> VarT m a (Event c)
- step :: (Applicative m, Monad m) => b -> SplineT a b m ()
- untilEvent :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m (b, c)
- untilEvent_ :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m b
- _untilEvent :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m b
- pair :: Monad m => (b -> d -> f) -> SplineT a b m c -> SplineT a d m e -> SplineT a f m (c, e)
- race :: (Applicative m, Monad m) => (b -> d -> e) -> SplineT a b m c -> SplineT a d m c -> SplineT a e m c
- capture :: (Applicative m, Monad m, Eq b) => SplineT a b m c -> SplineT a b m (Maybe b, c)
- mapOutput :: (Applicative m, Monad m) => VarT m a (b -> t) -> SplineT a b m c -> SplineT a t m c
- adjustInput :: Monad m => VarT m a (a -> r) -> SplineT r b m c -> SplineT a b m c
- data Step b c = Step {
- stepOutput :: b
- stepResult :: Event c
Spline
type Spline a b c = SplineT a b Identity c Source
A SplineT monad parameterized with Identity that takes input of type a
,
output of type b
and a result value of type c
.
Spline Transformer
SplineT
shares a number of types with VarT
, specifically its monad,
input and output types (m
, a
and b
, respectively). A spline adds
a result type which represents the monadic computation's result
value.
Much like the State monad it has an "internal state" and an eventual
result value, where the internal state is the output value. The result
value is used only in determining the next spline to sequence.
MonadTrans (SplineT a b) Source | A spline is a transformer and other monadic computations can be lifted int a spline. |
(Applicative m, Monad m) => Monad (SplineT a b m) Source | A spline is monad if its output type is a monoid. A spline responds to bind by running until it produces an eventual value, then uses that value to run the next spline. |
(Applicative m, Monad m) => Functor (SplineT a b m) Source | A spline is a functor by applying the function to the result. |
(Applicative m, Monad m) => Applicative (SplineT a b m) Source | A spline is an applicative if its output type is a monoid. It
responds to |
(Functor m, Applicative m, MonadIO m) => MonadIO (SplineT a b m) Source | A spline can do IO if its underlying monad has a MonadIO instance. It
takes the result of the IO action as its immediate return value and
uses |
runSplineT :: (Applicative m, Monad m) => SplineT a b m c -> VarT m a (Step (Event b) c) Source
Unwrap a spline into a value stream.
scanSpline :: (Applicative m, Monad m) => SplineT a b m c -> [a] -> m [(Event b, Event c)] Source
Run the spline over the input values, gathering the output and result values in a list.
fromEvents :: (Applicative m, Monad m) => b -> VarT m a (Event b) -> SplineT a b m b Source
Create a spline using an event stream. The spline will run until the stream inhibits, using the stream's last produced value as the current output value. In the case the stream inhibits before producing a value the default value is used. The spline's result is the last output value.
outputStream :: (Applicative m, Monad m) => b -> SplineT a b m c -> VarT m a b Source
Evaluates a spline into a value stream of its output type.
resultStream :: (Applicative m, Monad m) => SplineT a b m c -> VarT m a (Event c) Source
Evaluates a spline to an event stream of its result. The resulting value stream inhibits until the spline's domain is complete and then it produces events of the result type.
step :: (Applicative m, Monad m) => b -> SplineT a b m () Source
Produce the argument as an output value exactly once.
Combinators
untilEvent :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m (b, c) Source
Create a spline from a value stream and an event stream. The spline uses the value stream as its output value. The spline will run until the event stream produces a value, at that point the last output value and the event value are tupled and returned as the spline's result value.
untilEvent_ :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m b Source
A variant of untilEvent
that only results in the left result,
discarding the right result.
_untilEvent :: (Applicative m, Monad m) => VarT m a b -> VarT m a (Event c) -> SplineT a b m b Source
A variant of untilEvent
that only results in the right result,
discarding the left result.
pair :: Monad m => (b -> d -> f) -> SplineT a b m c -> SplineT a d m e -> SplineT a f m (c, e) Source
Run two splines in parallel, combining their output. When both conclude, return their result values in a tuple.
race :: (Applicative m, Monad m) => (b -> d -> e) -> SplineT a b m c -> SplineT a d m c -> SplineT a e m c Source
Run two splines in parallel, combining their output. Return the result of the spline that concludes first. If they conclude at the same time the result is taken from the left spline.
capture :: (Applicative m, Monad m, Eq b) => SplineT a b m c -> SplineT a b m (Maybe b, c) Source
Capture the spline's last output value and tuple it with the spline's result. This is helpful when you want to sample the last output value in order to determine the next spline to sequence.
mapOutput :: (Applicative m, Monad m) => VarT m a (b -> t) -> SplineT a b m c -> SplineT a t m c Source
Map the output value of a spline.
adjustInput :: Monad m => VarT m a (a -> r) -> SplineT r b m c -> SplineT a b m c Source
Map the input value of a spline.
Step
A discrete step in a continuous function. This type discretely describes an eventual value on the right and an output value on the left.
Step | |
|
Functor (Step a) Source | A discrete step is a functor by applying a function to the contained event's value. |
Monoid f => Applicative (Step f) Source | A discrete spline is an applicative if its left datatype is a monoid. It
replies to |
(Monoid f, Monoid b) => Monoid (Step f b) Source | A discrete spline is a monoid if its left and right types are monoids. |