This GitHub template allows you to create deployments of JupyterLite with a custom set of conda packages.
Creating a new deployment can be done in three easy steps:
Step 1: Apply the GitHub template
Step 2: Enable building the GitHub pages from GitHub actions.
Step 3: Customize your conda environment
environment.yml
file to include the required packages.For a step-by-step visual guide, check out the screencast below:
You can install specific kernels and extra packages by adding them to the environment.yml
file.
To create a JupyterLite deployment with NumPy and Matplotlib pre-installed, edit the environment.yml
file as follows:
name: xeus-kernel
channels:
- https://repo.mamba.pm/emscripten-forge
- conda-forge
dependencies:
- xeus-python
- numpy
- matplotlib
To use the R kernel and the coursekata package, edit the environment.yml file as follows:
name: xeus-kernel
channels:
- https://repo.mamba.pm/emscripten-forge
- conda-forge
dependencies:
- xeus-r
- r-coursekata
Only no-arch
packages from conda-forge
and packages from emscripten-forge
can be installed.
How do I know if a package is no-arch
on conda-forge
? no-arch
means that the package is OS-independent, usually pure-python packages are no-arch
. To check if your package is no-arch
on conda-forge
, check if the “Platform” entry is “no-arch” in the https://beta.mamba.pm/channels/conda-forge?tab=packages page. If your package is not no-arch
but is a pure Python package, then you should probably update the feedstock to turn your package into a no-arch
one.
How do I know if my package is on emscripten-forge
? You can see the list of packages pubished on emscripten-forge
here. In case your package is missing, or it’s not up-to-date, feel free to open an issue or a PR on https://github.com/emscripten-forge/recipes.