Software – Python Virtual Environments

Python is available across all systems, The version of Python will be different depending on the system’s OS.

This enables you to build your python environment without root access, as well as making it much more portable across systems, plus it’s really simple to do.

Python Versions

Create Environment

python3 -m virtualenv <environment_name>

python3 -m venv <environment_name>

source <environment_name>/bin/activate
or
source <environment_name>/bin/activate.csh
or
source <environment_name>/bin/activate.fish

then you can run pip/pip3 etc as you wish.

Upgrade the Environment

If the version of Python has updated on the system and is causing your envrionment to not load try this command in the shell

python3 -m venv --upgrade <environment_name>

Leave the Environment

deactivate

Known Issues with Virtual Environments

People with large 1GB+ environments especially when it’s lots of little files sometimes find that the activation can take a long time.

BYOP – Bring Your Own Python

Please note this compiles on the system you’re running so it may not be interoperable between operating systems.

If you want to use pyenv in its standard method:

curl -fsSL https://pyenv.run | bash

echo 'export PYENV_ROOT="$HOME/.pyenv"' >> ~/.bashrc
echo '[[ -d $PYENV_ROOT/bin ]] && export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.bashrc
echo 'eval "$(pyenv init - bash)"' >> ~/.bashrc

pyenv update
pyenv install <VERSION>
pyenv global <VERSION>

To get a list of all supported version:

pyenv install -l

If you wish to store your custom python in alternative storage do this first (change for your storage location)

export PYENV_ROOT=/mt/user-batch/$USER/.pyenv

echo 'export PYENV_ROOT="/mt/user-batch/$USER/.pyenv"' >> ~/.bashrc
echo '[[ -d $PYENV_ROOT/bin ]] && export PATH="$PYENV_ROOT/bin:$PATH"' >> ~/.bashrc
echo 'eval "$(pyenv init - bash)"' >> ~/.bashrc

To remove a version

pyenv uninstall <version>

If you wish to do some weird python magic, you can modify the python interpreter via the shebang on the first line to specify a pyenv compiled version for example:

#!/mt/home/<username>/.pyenv/versions/<version>/bin/python

See more at PyEnv on Github.

Jupyter Kernel

If you wish to add a Virtual Environment as a Jupyter Kernel then add you will need to install the ipykernel

source <envname>/bin/activate
pip install ipykernel

You might be able to then run

ipython kernel install --user --name=<envname>

otherwise you will also need to manually make a kernel.json file for this, to do so you must create a file in

~/.local/share/jupyter/kernels/<envname>/kernel.json

with something similar to the following

{
 "argv": [
  "/mt/home/myuser/myenv/bin/python",
  "-m",
  "ipykernel_launcher",
  "-f",
  "{connection_file}"
 ],
 "display_name": "myenv",
 "language": "python"
}

This is a modified excerpt from a Nikolai Janakiev’s Blog (External link not attributed to IPPP or Durham University).