INSTALLATION.md

Installation of bidsme

Since version 1.4.0 bidsme can be installed using pip. It is the best option if you don't plan to contribute to bidsme developmnent. For advanced usage, and development, you can install bidsme manually (with or without pip). I will suggest to install bidsme within it's own virtual environment (see below).

Automatic installation using pip (not editable)

To install bidsme using pip you just need to write in terminal:

pip install git+https://gitlab.uliege.be/CyclotronResearchCentre/Public/bidstools/bidsme/bidsme.git

It should install bidsme and nessesary dependencies in the local Python3 library.

Once installed, you should be able to run bidsme from terminal:

bidsme --help

or from Python console:

import bidsme
bidsme.init()

Manual installation (editable, for experts)

To install bidsme manually, you need first clone bidsme repository:

cd <installation dir>
git clone https://gitlab.uliege.be/CyclotronResearchCentre/Public/bidstools/bidsme/bidsme.git

Once repository is cloned, you can add nessesary paths to Python paths using pip -e:

pip install -e <installation dir>/bidsme

In this case, pip will automatically install needed dependancies, also it will be aviable to be imported in Python3 console or notebook:

import bidsme
bidsme.init()

Or just run bidsme directly from cloned project:

python3 <installation dir>/bidsme.py --help

In this case you need to manually install the dependencies from requirements.txt file.

Using virtual environments and kernels

bidsme will require the installation of some additional Python packages, some of them are very common, like pandas, and likely already present in your installation of Python, others are less common. In order to keep Python installation clean, usage of virtual environments and/or kernels are suggested.

If you are using virtual environment and/or (Ana)conda, then creating a new envoronment is straightforward, in terminal you just need to: venv:

python3 -m venv bidsme_env
source bidsme_env/bin/activate

Conda:

conda create --name bidsme_env
conda activate bidsme_env

In order to deactivate (return to your default) environment, you just need deactivate with venv or conda deactivate in conda.

Creating kernel for Jupyter-notebook/lab

If you intend to use bidsme in Jupyter-notebook, you need to install the kernel -- a library that will link iPython/jupyter interface with environment.

Still within the terminal, and active installed environment, do:

pip install ipykernel
python -m ipykernel install --user --name bidsme_env --display-name "bidsme_env (Python)"

The first line will install the kernel package, and second will create a new kernel with internal name bidsme_env and displayed name Python (bidsme_env). For more instructions and details, you can refer to the Kernel instructions.

Once kernel is installed, you can open a new jupyter(-lab) notebook, and check if the new kernel of name Python (bidsme_env) is available. This way all necessary packages will be installed in dedicated virtual environment and will not create conflicts with your other Python projects.

Now in the notebook, after choosing kernel Python (bidsme_env) bidsme should be aviable:

import bidsme
bidsme.init()