Simulation.Aivika.Trans.Stream

Tested with: GHC 7.10.1

The infinite stream of data in time.

Stream Type

newtype Stream m a

Represents an infinite stream of data in time, some kind of never-ending cons cell.

Constructors

Cons

Fields

runStream :: Process m (a, Stream m a)

Run the stream.

Instances

MonadDES m => Functor (Stream m)

MonadDES m => Applicative (Stream m)

MonadDES m => Alternative (Stream m)

MonadDES m => Monoid (Stream m a)

Merging and Splitting Stream

emptyStream :: MonadDES m => Stream m a

An empty stream that never returns data.

mergeStreams :: MonadDES m => Stream m a -> Stream m a -> Stream m a

Merge two streams applying the FCFS strategy for enqueuing the input data.

mergeQueuedStreams

Arguments

:: (MonadDES m, EnqueueStrategy m s)

=> s

the strategy applied for enqueuing the input data

-> Stream m a

the fist input stream

-> Stream m a

the second input stream

-> Stream m a

the output combined stream

Merge two streams.

If you don't know what the strategy to apply, then you probably need the FCFS strategy, or function mergeStreams that does namely this.

mergePriorityStreams

Arguments

:: (MonadDES m, PriorityQueueStrategy m s p)

=> s

the strategy applied for enqueuing the input data

-> Stream m (p, a)

the fist input stream

-> Stream m (p, a)

the second input stream

-> Stream m a

the output combined stream

Merge two priority streams.

concatStreams :: MonadDES m => [Stream m a] -> Stream m a

Concatenate the input streams applying the FCFS strategy and producing one output stream.

concatQueuedStreams

Arguments

:: (MonadDES m, EnqueueStrategy m s)

=> s

the strategy applied for enqueuing the input data

-> [Stream m a]

the input stream

-> Stream m a

the combined output stream

Concatenate the input streams producing one output stream.

If you don't know what the strategy to apply, then you probably need the FCFS strategy, or function concatStreams that does namely this.

concatPriorityStreams

Arguments

:: (MonadDES m, PriorityQueueStrategy m s p)

=> s

the strategy applied for enqueuing the input data

-> [Stream m (p, a)]

the input stream

-> Stream m a

the combined output stream

Concatenate the input priority streams producing one output stream.

splitStream :: MonadDES m => Int -> Stream m a -> Simulation m [Stream m a]

Split the input stream into the specified number of output streams after applying the FCFS strategy for enqueuing the output requests.

splitStreamQueueing

Arguments

:: (MonadDES m, EnqueueStrategy m s)

=> s

the strategy applied for enqueuing the output requests

-> Int

the number of output streams

-> Stream m a

the input stream

-> Simulation m [Stream m a]

the splitted output streams

Split the input stream into the specified number of output streams.

If you don't know what the strategy to apply, then you probably need the FCFS strategy, or function splitStream that does namely this.

splitStreamPrioritising

Arguments

:: (MonadDES m, PriorityQueueStrategy m s p)

=> s

the strategy applied for enqueuing the output requests

-> [Stream m p]

the streams of priorities

-> Stream m a

the input stream

-> Simulation m [Stream m a]

the splitted output streams

Split the input stream into a list of output streams using the specified priorities.

splitStreamFiltering :: MonadDES m => [a -> Event m Bool] -> Stream m a -> Simulation m [Stream m a]

Split the input stream into the specified number of output streams after filtering and applying the FCFS strategy for enqueuing the output requests.

splitStreamFilteringQueueing

Arguments

:: (MonadDES m, EnqueueStrategy m s)

=> s

the strategy applied for enqueuing the output requests

-> [a -> Event m Bool]

the filters for output streams

-> Stream m a

the input stream

-> Simulation m [Stream m a]

the splitted output streams

Split the input stream into the specified number of output streams after filtering.

If you don't know what the strategy to apply, then you probably need the FCFS strategy, or function splitStreamFiltering that does namely this.

Specifying Identifier

streamUsingId :: MonadDES m => ProcessId m -> Stream m a -> Stream m a

Create a stream that will use the specified process identifier. It can be useful to refer to the underlying Process computation which can be passivated, interrupted, canceled and so on. See also the processUsingId function for more details.

Prefetching and Delaying Stream

prefetchStream :: MonadDES m => Stream m a -> Stream m a

Prefetch the input stream requesting for one more data item in advance while the last received item is not yet fully processed in the chain of streams, usually by the processors.

You can think of this as the prefetched stream could place its latest data item in some temporary space for later use, which is very useful for modeling a sequence of separate and independent work places.

delayStream :: MonadDES m => a -> Stream m a -> Stream m a

Delay the stream by one step using the specified initial value.

Stream Arriving

arrivalStream :: MonadDES m => Stream m a -> Stream m (Arrival a)

Transform a stream so that the resulting stream returns a sequence of arrivals saving the information about the time points at which the original stream items were received by demand.

Memoizing, Zipping and Uzipping Stream

memoStream :: MonadDES m => Stream m a -> Simulation m (Stream m a)

Memoize the stream so that it would always return the same data within the simulation run.

zipStreamSeq :: MonadDES m => Stream m a -> Stream m b -> Stream m (a, b)

Zip two streams trying to get data sequentially.

zipStreamParallel :: MonadDES m => Stream m a -> Stream m b -> Stream m (a, b)

Zip two streams trying to get data as soon as possible, launching the sub-processes in parallel.

zip3StreamSeq :: MonadDES m => Stream m a -> Stream m b -> Stream m c -> Stream m (a, b, c)

Zip three streams trying to get data sequentially.

zip3StreamParallel :: MonadDES m => Stream m a -> Stream m b -> Stream m c -> Stream m (a, b, c)

Zip three streams trying to get data as soon as possible, launching the sub-processes in parallel.

unzipStream :: MonadDES m => Stream m (a, b) -> Simulation m (Stream m a, Stream m b)

Unzip the stream.

streamSeq :: MonadDES m => [Stream m a] -> Stream m [a]

To form each new portion of data for the output stream, read data sequentially from the input streams.

This is a generalization of zipStreamSeq.

streamParallel :: MonadDES m => [Stream m a] -> Stream m [a]

To form each new portion of data for the output stream, read data from the input streams in parallel.

This is a generalization of zipStreamParallel.

Consuming and Sinking Stream

consumeStream :: MonadDES m => (a -> Process m ()) -> Stream m a -> Process m ()

Consume the stream. It returns a process that infinitely reads data from the stream and then redirects them to the provided function. It is useful for modeling the process of enqueueing data in the queue from the input stream.

sinkStream :: MonadDES m => Stream m a -> Process m ()

Sink the stream. It returns a process that infinitely reads data from the stream. The resulting computation can be a moving force to simulate the whole system of the interconnected streams and processors.

Useful Combinators

repeatProcess :: MonadDES m => Process m a -> Stream m a

Return a stream of values generated by the specified process.

mapStream :: MonadDES m => (a -> b) -> Stream m a -> Stream m b

Map the stream according the specified function.

mapStreamM :: MonadDES m => (a -> Process m b) -> Stream m a -> Stream m b

Compose the stream.

accumStream :: MonadDES m => (acc -> a -> Process m (acc, b)) -> acc -> Stream m a -> Stream m b

Accumulator that outputs a value determined by the supplied function.

apStream :: MonadDES m => Stream m (a -> b) -> Stream m a -> Stream m b

Sequential application.

apStreamM :: MonadDES m => Stream m (a -> Process m b) -> Stream m a -> Stream m b

Sequential application.

filterStream :: MonadDES m => (a -> Bool) -> Stream m a -> Stream m a

Filter only those data values that satisfy to the specified predicate.

filterStreamM :: MonadDES m => (a -> Process m Bool) -> Stream m a -> Stream m a

Filter only those data values that satisfy to the specified predicate.

takeStream :: MonadDES m => Int -> Stream m a -> Stream m a

Return the prefix of the stream of the specified length.

takeStreamWhile :: MonadDES m => (a -> Bool) -> Stream m a -> Stream m a

Return the longest prefix of the stream of elements that satisfy the predicate.

takeStreamWhileM :: MonadDES m => (a -> Process m Bool) -> Stream m a -> Stream m a

Return the longest prefix of the stream of elements that satisfy the computation.

dropStream :: MonadDES m => Int -> Stream m a -> Stream m a

Return the suffix of the stream after the specified first elements.

dropStreamWhile :: MonadDES m => (a -> Bool) -> Stream m a -> Stream m a

Return the suffix of the stream of elements remaining after takeStreamWhile.

dropStreamWhileM :: MonadDES m => (a -> Process m Bool) -> Stream m a -> Stream m a

Return the suffix of the stream of elements remaining after takeStreamWhileM.

singletonStream :: MonadDES m => a -> Stream m a

Return a stream consisting of exactly one element and inifinite tail.

joinStream :: MonadDES m => Process m (Stream m a) -> Stream m a

Removes one level of the computation, projecting its bound stream into the outer level.

Failover

failoverStream :: MonadDES m => [Stream m a] -> Stream m a

Takes the next stream from the list after the current stream fails because of cancelling the underlying process.

Integrating with Signals

signalStream :: MonadDES m => Signal m a -> Process m (Stream m a)

Return a stream of values triggered by the specified signal.

Since the time at which the values of the stream are requested for may differ from the time at which the signal is triggered, it can be useful to apply the arrivalSignal function to add the information about the time points at which the signal was actually received.

The point is that the Stream is requested outside, while the Signal is triggered inside. They are different by nature. The former is passive, while the latter is active.

The resulting stream may be a root of space leak as it uses an internal queue to store the values received from the signal. The oldest value is dequeued each time we request the stream and it is returned within the computation.

Cancel the stream's process to unsubscribe from the specified signal.

streamSignal :: MonadDES m => Stream m a -> Process m (Signal m a)

Return a computation of the signal that triggers values from the specified stream, each time the next value of the stream is received within the underlying Process computation.

Cancel the returned process to stop reading from the specified stream.

Utilities

leftStream :: MonadDES m => Stream m (Either a b) -> Stream m a

The stream of Left values.

rightStream :: MonadDES m => Stream m (Either a b) -> Stream m b

The stream of Right values.

replaceLeftStream :: MonadDES m => Stream m (Either a b) -> Stream m c -> Stream m (Either c b)

Replace the Left values.

replaceRightStream :: MonadDES m => Stream m (Either a b) -> Stream m c -> Stream m (Either a c)

Replace the Right values.

partitionEitherStream :: MonadDES m => Stream m (Either a b) -> Simulation m (Stream m a, Stream m b)

Partition the stream of Either values into two streams.

Assemblying Streams

cloneStream :: MonadDES m => Int -> Stream m a -> Simulation m [Stream m a]

Create the specified number of equivalent clones of the input stream.

firstArrivalStream :: MonadDES m => Int -> Stream m a -> Stream m a

Return a stream of first arrivals after assembling the specified number of elements.

lastArrivalStream :: MonadDES m => Int -> Stream m a -> Stream m a

Return a stream of last arrivals after assembling the specified number of elements.

assembleAccumStream :: MonadDES m => (acc -> a -> Process m (acc, Maybe b)) -> acc -> Stream m a -> Stream m b

Assemble an accumulated stream using the supplied function.

Debugging

traceStream

Arguments

:: MonadDES m

=> Maybe String

the request message

-> Maybe String

the response message

-> Stream m a

a stream

-> Stream m a

Show the debug messages with the current simulation time.