Installation

Installation

BAT.jl is written in the Julia programming language. To use BAT.jl, you will need to install Julia itself, the BAT.jl Julia package, and some additional Julia packages. Depending on your use case, you may also want to need a Python installation with certain Python packages (optional).

Table of contents:

Installing Julia

Julia is available for Linux, OS-X and Windows, and easy to install:

Note

BAT.jl requires Julia >= v1.0.5. Julia v1.3 or higher is recommended for full functionality and performance.

Installing Jupyter and matplotlib/pyplot (optional)

If you plan to use Jupyter notebooks and/or the matplotlib/pyplot backend of the Julia Plots.jl package, you will also need a Python installation and certain Python packages (see below). BAT.jl is fully usable without Jupyter and matplotlib/pyplot, but both can come in very handy.

Julia can either use existing installations of Jupyter and pyplot , or install both internally by creating an internal Conda installation within $HOME/.julia/conda. We recommend the first approach (especially using Anaconda), since Julia will otherwise have to download over 1 GB of software, the $HOME/.julia directory will grow very large, and you will need to start Jupyter in an indirect fashion via Julia (only to have Jupyter then start additional Julia instances as Jupyter kernels in return).

For details, see the IJulia.jl, PyCall.jl and PyPlot.jl documentation (you should not need to if you follow the steps below).

On Linux, Julia (more specifically the Julia packages IJulia.jl, PyCall.jl, and PyPlot.jl) will by default try to use the matplotlib/pyplot installation associated with the python3 (resp. python) executable on your $PATH. Likewise, Julia will by default try to use the Jupyter installation associated with the jupyter executable on your $PATH.

However, on OS-X and Windows, both IJulia.jl and PyCall.jl by default always create a Julia-internal Conda installation (see above), even if Jupyter and matplotlib/pyplot are available (apparently broken Jupyter/Python installations on these platforms caused frequent support requests). In contrast to this default behavior, we recommend to use a standalone Jupyter and Python installation on all OS platforms. Set the environment variables $JUPYTER and $PYTHON to point to your Jupyter and Python executable to force Julia to use the existing installation.

We recommend that you install the Anaconda Python distribution, it includes both Jupyter and pyplot (it is of course possible to use non-Anaconda Jupyter and pyplot installations instead).

Installing Anaconda (optional)

To install Anaconda

Environment variables

You may want/need to set the following environment variables:

If you intend to use Jupyter, you will probably want to include the directory containing the jupyter binary to your PATH as well.

See the Julia manual for a description of other Julia-specific environment variables.

Installing BAT.jl and related Julia packages

BAT.jl is provided as a registered Julia package. To install it, simply run

julia> using Pkg
julia> pkg"add BAT"

However, you will likely need other Julia packages too. We recommend that you install certain statistics, plotting, I/O and array packages as well:

julia> using Pkg
julia> pkg"add BAT ArraysOfArrays Distributions ElasticArrays IntervalSets Parameters Plots ValueShapes StatsBase Tables TypedTables"

To install the latest development version of BAT (master branch), instead of the latest stable release, use

julia> pkg"add BAT#master"

If you'd like to precompile all installed packages right aways (otherwise they'll get precompiled when loaded for the first time), run

julia> pkg"precompile"

Installing additional Julia packages

HDF5 (File I/O)

To enable BAT's HDF5 file I/O capabilities, add the package "HDF5.jl":

pkg"add HDF5"

IJulia (Jupyter Julia kernel)

To use the the Julia Jupyter kernel, you need to add the package "IJulia.jl":

On Linux, simply use

pkg"add IJulia"

On OS-X, if you have an existing Jupyter installation (e.g. via Anaconda) and would like Julia to use it (instead of an internal Conda installation, see above), use (e.g.)

ENV["JUPYTER"] = "$(homedir())/opt/anaconda3/bin/jupyter"; pkg"add IJulia"

On Windows, if would like Julia to use an existing Jupyter installation (see above), use something like

ENV["JUPYTER"] = "DRIVE:/path/to/your/anaconda/.../jupyter.exe"; pkg"add IJulia"

Julia will remember the chosen Jupyter installation permanently, ENV["JUPYTER"] only needs to be set the first time you run pkg"add IJulia".

PyPlot (optinal backend for Plots.jl)

To use the Plots.jl matplotlib/pyplot backend (see above), add the package "PyPlot.jl". To call Python code from Julia yourself directly, you may want to add "PyCall.jl" as well:

On Linux, use:

pkg"add PyPlot PyCall"

On OS-X, if you have an existing Python/matplotlib/pyplot installation and would like Julia to use it (see above), use (e.g.):

ENV["PYTHON"] = "$(homedir())/opt/anaconda3/bin/python", pkg"add PyPlot PyCall"

On Windows, if would like Julia to use an existing Python installation (see above), use something like:

ENV["PYTHON"] = "DRIVE:/path/to/your/anaconda/.../python.exe"; pkg"add PyPlot PyCall"

Julia will remember the chosen Python installation permanently, ENV["PYTHON"] only needs to be set the first time you run pkg"add PyPlot PyCall".

Additional customization options

Note: If you want Julia to install packages in another location than $HOME/.julia, set the environment variables JULIA_DEPOT_PATH and JULIA_PKG_DEVDIR (see above).