dummy-link

Signals

Multi-Paradigm Dynamic Fast Functional Reactive Programming in Julia

Readme

Signals

Build Status[codecov.io

Signals provides a multi-paradigm fast functional reactive programing for julia. It supports both pull and push operations and async(default) and non-async modes.

Signal Creation

S = Signal(val;strict_push = false)

Create a source Signal with initial value val, setting strict_push to true guarantees that every push to this Signal will be carried out independently. otherwise if updates occur faster than what the eventloop can process, then only the last value before the eventloop kicks in will be used(default)

S = Signal(f,args...)

Creates a derived Signal who's value is f(args...) , args can be of any type, Signal args get replaced by their value before calling f(args...). reads best with with do notation(see example).

Syntax

S[] = val

sets the value of S to val without propagating the change to the rest of the signal graph, useful in pull based paradigm

S()

pull! Signal, pulling any changes in the Signal graph that affects S

S(val)

sets the value of S to val and pushes the changes along the Signal graph

S[]

gets the current value stored in S without pulling changes from the graph

Example

julia> A = Signal(1)   # source Signal
Signal
value: 1

julia> B = 2           # non-Signal input
2

julia>  C = Signal(A,B) do a,b   # derived Signal
            a+b
        end

Signal
value: 3

julia> A[] = 10        # set value without propagation
10
julia> C[]             # read current value
3
julia> C()             # pull latest changes from the Signal graph
12
julia> A(100)          # set value to a signal and propagate this change
100
julia> C[]
102

Operators

Signals supports several reactive operators

  • droprepeats
  • when
  • filter
  • sampleon
  • foldp
  • count
  • prev
  • merge
  • async_signal
  • remote_signal
  • bind!
  • unbind!

individual documentation files are available from within julia

Time operators

Signals supports several operators that takes time into consideration, for example debounce which executes only after a set amount of time has passed since the inputs were updated or throttle which creates a Signal that is guaranteed not to be executed more than set amount of times per second.

  • debounce
  • throttle
  • for_signal
  • fps
  • every
  • buffer

Async mode

By default Signals run asynchronically in a dedicated eventloop, this behavior can be changed using

Signals.async_mode(false)

or by individual non-async pushes to the signal graph using:

push!(s,val,false)

Dynamic

Signals is dynamic , one can push values of any type to a source signal

julia> using Signals
julia> A = Signal(1)
Signal
value: 1

julia> B = Signal(A,A) do a,b
       a*b
       end
Signal
value: 1

julia> A(rand(3,3));
julia> B()
3×3 Array{Float64,2}:
 0.753217  0.796031  0.265298
 0.28489   0.209641  0.249161
 0.980177  0.810512  0.571937

Fast

Signals package was rigorously optimized for speed of execution and minimal recalculation of signal graph updates.

First Commit

09/27/2017

Last Touched

about 2 months ago

Commits

38 commits

Used By: