I've build PyPlot against python version 3.6.4 that I usually use, via ENV["PYTHON"] = /home/...
This python has matplotlib 3.0.0 installed and it works fine. Although building doesn't throw an error, using PyPlot throws
julia> using PyPlot
ERROR: InitError: PyCall.PyError("PyImport_ImportModule\n\nThe Python package matplotlib.pyplot couldnot be found by pyimport
if in the same REPL session I rebuild it, using PyPlot runs through only throwing the
WARNING: using PyPlot.plt in module Main conflicts with an existing identifier.
WARNING: using PyPlot.axes in module Main conflicts with an existing identifier.
warning. However, whenever I try to do something like, I get
julia> plot(randn(300))
ERROR: ArgumentError: haskey of NULL PyObject
Stacktrace:
[1] haskey(::PyObject, ::String) at /home/kloewer/.julia/packages/PyCall/0jMpb/src/PyCall.jl:310
[2] #plot#85(::Base.Iterators.Pairs{Union{},Union{},Tuple{},NamedTuple{(),Tuple{}}}, ::Function, ::Array{Float64,1}) at /home/kloewer/.julia/packages/PyPlot/fZuOQ/src/PyPlot.jl:176
[3] plot(::Array{Float64,1}) at /home/kloewer/.julia/packages/PyPlot/fZuOQ/src/PyPlot.jl:176
[4] top-level scope at none:0
Building PyCall with ENV["PYTHON"] = "" solves the problem, but I find it a bit redundant having two python environments that essentially fulfill the same purpose.
I've build PyPlot against python version 3.6.4 that I usually use, via
ENV["PYTHON"] = /home/...This python has matplotlib 3.0.0 installed and it works fine. Although building doesn't throw an error, using PyPlot throws
if in the same REPL session I rebuild it, using PyPlot runs through only throwing the
warning. However, whenever I try to do something like, I get
Building PyCall with
ENV["PYTHON"] = ""solves the problem, but I find it a bit redundant having two python environments that essentially fulfill the same purpose.