Jupyter Saagie plugin
This plugin allows you to easily create a Saagie Python job from a local Jupyter notebook.
In your Jupyter Python environment, run:
pip install jupyter-saagie-plugin
jupyter notebook as always, you will see a new button
in the toolbar. Use it to log in to your account and deploy a job
using the code in your notebook. You can update an existing job as well.
For deploying to a Python job, Python 3.5 & 2.7 are supported.
If you want to pre-fill your Saagie root URL (used in the plugin's login form),
you can setup a
SAAGIE_ROOT_URL environment variable before you run the notebook.
pip uninstall jupyter-saagie-plugin jupyter nbextension uninstall saagie --user
Modifying this extension
In dev mode, you can use a Docker of Jupyter notebook and install this extension in it. First, run a Jupyter Docker in bash mode:
docker run -it --rm -p 8888:8888 --user root -e GRANT_SUDO=yes -v [your_project_path]/jupyter-saagie-plugin/:/jupyter-saagie-plugin/ [--env SAAGIE_ROOT_URL="https://xxx-manager.prod.saagie.io/"] saagie/jupyter-python-nbk:latest /bin/bash
(Seting up the
SAAGIE_ROOT_URL environment variable is optional)
Then install the Jupyter extension in your Docker container:
cd /jupyter-saagie-plugin/ python3 setup.py install cd /notebooks-dir /usr/local/bin/start-notebook.sh --NotebookApp.token='' --NotebookApp.password=''
Build and share
In order to build this plugin, run the following command:
python setup.py sdist
This will generate a tar.gz archive in the
If you want to share it on Pypi, just install twine (with
pip install twine) and run:
twine upload dist/jupyter-saagie-plugin-x.x.x.tar.gz
But first, You can test your Pypi deployment on TestPypi test environment. If so, just run:
twine upload --repository-url https://test.pypi.org/legacy/ dist/jupyter-saagie-plugin-x.x.x.tar.gz