Streaming feature

[femtotrader]

Hello,

I wonder if Temporal.jl have streaming feature ie a kind of circular buffer datastructure which is able to receive for example tick events and store them or to resample to candlestick (OHLCV).

Kind regards

PS: see StreamTimeArray https://github.com/femtotrader/TimeSeriesIO.jl/blob/master/src/stream_timearray.jl StreamTimeArrayOHLCV https://github.com/femtotrader/TimeSeriesIO.jl/blob/master/src/stream_timearray_ohlcv.jl

[dysonance]

@femtotrader This is something that I would definitely like to add on later. Something I am always considering is the amount of dependencies required to use my packages. As the package becomes more fully developed and feature-rich, additional dependencies will surely follow, but for now I'm trying to keep the pre-requisites in moderation. For now my primary goal is to be able to analyze and manipulate time series data as efficiently as possible, so that my planned efforts to develop a systematic trading & backtesting platform will go smoothly. Getting live streaming quotes strikes me as something that would come after much of the historical analytical side of things has been fully implemented.

[femtotrader]

You might be interested by https://github.com/femtotrader/TALib.jl see https://github.com/femtotrader/TALib.jl/issues/19

[femtotrader]

Some inspiration from Python ecosystem with streamz https://github.com/JuliaQuant/MarketTechnicals.jl/issues/93#issuecomment-341361834

[femtotrader]

A streaming example (probably not very efficient)

using Temporal
import Base: circshift!, push!

function circshift!(ts::TS, n)
    ts.values = circshift(ts.values, n)
    ts.index = circshift(ts.index, n)
    ts
end

function push!(ts::TS, values::Array, idx::Vector)
    n = length(idx)
    circshift!(ts, -n)
    ts.values[end-n+1:end,:] = values
    ts.index[end-n+1:end] = idx
    ts
end

srand(1234)
idx=DateTime(2010,1,1):Dates.Hour(1):DateTime(2017,1,1)-Dates.Hour(1)
n=length(idx)
price=100+cumsum(2*(rand(n)-0.5))
volume=rand(n)*1000
ts = TS([price volume], collect(idx), [:price, :volume])
println(ts)

#ts = circshift(ts, -1)
#println(ts)

new_values = [204.0 100.0]
new_idx = [DateTime(2017,1,1,0,0,0)]

push!(ts, new_values, new_idx)

new_values = [205.0 100.1 ; 206.0 100.2]
new_idx = [DateTime(2017,1,1,1,0,0),DateTime(2017,1,1,2,0,0)]
push!(ts, new_values, new_idx)

println(ts)