======================

I implemented the generalized PCA algorithm that I had previously described only conceputally for our proposal. The code is at https://github.com/eford/RVSpectraKitLearn.jl

Some instructions below for trying it out...

Download and install a v0.5 release candidate of Julia from http://julialang.org/downloads/

Start julia

julia

Install the package

Pkg.clone("[email protected]:eford/RvSpectraKitLearn.jl.git")  # Install package

Edit the file path_to_spectra.jl to point to the directory with your spectra, perhaps like

path_to_spectra = "C:\\Users\\eford\\Box Sync\\SOAP simulations\\Aug2016_workshop\\SOAP_Spectra\\planet_10ms_150k"
path_to_spectra = joinpath(homedir(),"SOAP_Spectra/planet_10ms_150k")

include(joinpath(Pkg.dir("RvSpectraKitLearn"),"test","runtests.jl"))

You can see how to use as an example (very similar to tests), either from julia

include(joinpath(Pkg.dir("RvSpectraKitLearn"),"examples","gpca_ex1.jl"))

or from the shell

julia examples/gpca_ex1.jl

If you want to call Python from Julia, see https://github.com/stevengj/PyCall.jl . If you want to call Julia from python, see https://github.com/JuliaInterop/pyjulia (I haven't tested this).