This version differs from the simple one in providing an extra argument to the sampling action that will be globally distributed to every node and can be used to update the state. For instance, it can hold the time step between the two samplings, but it could also encode all the external input to the system.
The interface of this module differs from the old Elerea in the following ways:
- the delta time argument is generalised to an arbitrary type, so it
is possible to do without
externalaltogether in case someone wants to do so;
- there is no
samplerany more, it is substituted by
join, as signals are monads;
generatorhas been conceptually simplified, so it's a more basic primitive now;
- there is no automatic delay in order to preserve semantic soundness (e.g. the monad laws for signals);
- all signals are aged regardless of whether they are sampled (i.e. their behaviour doesn't depend on the context any more);
- the user needs to cache the results of applicative operations to be
reused in multiple places explicitly using the
- data Signal p a
- data SignalGen p a
- start :: SignalGen p (Signal p a) -> IO (p -> IO a)
- external :: a -> IO (Signal p a, a -> IO ())
- externalMulti :: IO (SignalGen p (Signal p [a]), a -> IO ())
- delay :: a -> Signal p a -> SignalGen p (Signal p a)
- stateful :: a -> (p -> a -> a) -> SignalGen p (Signal p a)
- transfer :: a -> (p -> t -> a -> a) -> Signal p t -> SignalGen p (Signal p a)
- memo :: Signal p a -> SignalGen p (Signal p a)
- generator :: Signal p (SignalGen p a) -> SignalGen p (Signal p a)
- noise :: MTRandom a => SignalGen p (Signal p a)
- getRandom :: MTRandom a => SignalGen p a
- debug :: String -> SignalGen p ()
A signal can be thought of as a function of type
Nat -> a, and
Monad instance agrees with that intuition. Internally, is
represented by a sampling computation.
|Monad (Signal p)|
|Functor (Signal p)|
|Applicative (Signal p)|
|Bounded t => Bounded (Signal p t)|
|Enum t => Enum (Signal p t)|
|Eq (Signal p a)|
Equality test is impossible.
|Floating t => Floating (Signal p t)|
|Fractional t => Fractional (Signal p t)|
|Integral t => Integral (Signal p t)|
|Num t => Num (Signal p t)|
|Ord t => Ord (Signal p t)|
|Real t => Real (Signal p t)|
|Show (Signal p a)|
A signal generator is the only source of stateful signals. Internally, computes a signal structure and adds the new variables to an existing update pool.
|:: SignalGen p (Signal p a)|
the generator of the top-level signal
|-> IO (p -> IO a)|
the computation to sample the signal
Embedding a signal into an
IO environment. Repeated calls to
the computation returned cause the whole network to be updated, and
the current sample of the top-level signal is produced as a
result. The computation accepts a global parameter that will be
distributed to all signals. For instance, this can be the time
step, if we want to model continuous-time signals.
A signal that can be directly fed through the sink function returned. This can be used to attach the network to the outer world. Note that this is optional, as all the input of the network can be fed in through the global parameter, although that is not really convenient for many signals.
|:: IO (SignalGen p (Signal p [a]), a -> IO ())|
a generator for the event signal and the associated sink
An event-like signal that can be fed through the sink function
returned. The signal carries a list of values fed in since the
last sampling, i.e. it is constantly  if the sink is never
invoked. The order of elements is reversed, so the last value
passed to the sink is the head of the list. Note that unlike
external this function only returns a generator to be used within
the expression constructing the top-level stream, and this
generator can only be used once.
delay transfer function emits the value of a signal from
the previous superstep, starting with the filler value given in the
A pure stateful signal. The initial state is the first output, and every following output is calculated from the previous one and the value of the global parameter.
A stateful transfer function. The current input affects the current output, i.e. the initial state given in the first argument is considered to appear before the first output, and can never be observed. Every output is derived from the current value of the input signal, the global parameter and the previous output.
Memoising combinator. It can be used to cache results of
applicative combinators in case they are used in several
places. Other than that, it is equivalent to
A reactive signal that takes the value to output from a monad carried by its input. It is possible to create new signals in the monad.