This project is designed to initialize the Python Environment for IFTPipeline.jl. It should be used once before IFTPipeline.jl is instantiated. It will:
- Build PyCall using the
Conda.jlpackage.
Since the base environment is shared between all PyCall instances by default, it only needs to be run if the Conda environment is:
- missing – as in a new installation, or
- broken – new incompatible dependencies have been added.
The goal is to remove the need for this subpackage at some point in future.
To use, run:
julia --project="PythonSetupForIFTPipeline.jl" PythonSetupForIFTPipeline.jl/setup.jlIf this fails, try removing the existing Conda installation for Julia:
rm -r ~/.julia/conda/... and then retry.
If you're working on a system where the home directory might not be reliable, e.g. you're building a Docker container but you want to use it with Apptainer which mounts things in different ways, you may want to change the default location for the Conda installation.
Export a different path in CONDA_JL_HOME before running setup:
export CONDA_JL_HOME="/opt/conda-env/"
julia --project="PythonSetupForIFTPipeline.jl" PythonSetupForIFTPipeline.jl/setup.jl