Source code for the AI on OpenShift site, using Material for MkDocs.
Rendered at https://ai-on-openshift.io/
The AI on OpenShift site aims at being a one-stop shop for installation recipes, patterns, demos for various AI/ML tools and applications used in Data Science and Data Engineering projects running on OpenShift.
-
Prerequisites: Python >=3.7
-
Install the
mkdocs-material
package from PyPI, either in your main Python environment or in a virtual environment:pip install mkdocs-material
. -
A Pipfile is also provided. To make use of it, you can execute:
pip install pipenv pipenv install pipenv shell
-
From the root of the repo, launch
mkdocs serve
-
The documentation will be accessible at http://127.0.0.1:8000
-
All saved modifications are watched and rendered real time
-
Once your environment is fully set up, you can also start the development environment by launching the
start-dev.sh
file.
Alternatively, you can also try serving the content using a container image.
-
For example, using podman and an image from DockerHub
podman run --rm -it \ -p 8000:8000 \ -v ${PWD}:/docs \ squidfunk/mkdocs-material
- Create a branch (maintainers team) or fork the repo (other contributors)
- Develop locally as described above
- Submit a Pull Request to the
main
branch of the repo