How to work in Jupyter Lab
What is Jupyter Lab?
Jupyter Lab is an interface for interactive programming and data analysis that runs in your browser. If you’re coming from R think about it as a browser-based R-Studio for all kinds of languages, but mostly used for Python.
Documentation and User Guide: https://jupyterlab.readthedocs.io/en/latest/
Install Jupyter Lab
Installation should be straightforward in most cases. Always follow the latest instructions by the developers:
We recommend simply using the pip installation for (most) purposes.
Install/uninstall kernels for virtual environments created with pipenv
Jupyter uses ‘kernels’ to accommodate working in different programming languages and environments
While it is possible to simply install and run
jupyter-lab within a Python virtualenv or conda env, it is maybe desirable to have certain environments offered to you by the launcher also within your system-level Jupyter installation. To get this, you have to specify a so-called ‘kernelspec’ that provides your system-level Jupyter with information about where your virtualenv lives, respectively.
To do so, assuming that Jupyter-Lab is installed on your system (outside a virtual env) already:
- enter virtual env with
pipenv install ipykernel or
pipenv install --dev ipykernel if end users don’t need to use Jupyter.
python -m ipykernel install --user --name=YOUR_RECOGNIZABLE_NAME_FOR_THIS_ENV
Now you should be able to start
jupyter-lab on your system (outside the virtual env) and the launcher will offer you to start a notebook using this virtualenv.
If you delete the project virtualenv for whatever reason and want to keep your kernel-list in Jupyter clean, you can list all kernels Jupyter knows of
jupyter kernelspec list
and then remove the respective kernel specifications with
jupyter kernelspec uninstall YOUR_RECOGNIZABLE_NAME_FOR_THIS_ENV
Sources/More to read:
Using Virtual Environments in Jupyter Notebook and Python: https://janakiev.com/blog/jupyter-virtual-envs/