Store, version, edit and execute notebooks in sandboxes and integrate them directly via REST interfaces.
- Ability to write machine learning logic and expose them to systems as rest api
- Write Jupyter nb locally and run them in a centralised powerful machine to reduce cost
- Create framework to directly connect Jupyter notebook to other systems
- docker
- redis
- Redis
- FastAPI
- Papermill
- Jupyter
- Poetry
- Docker
- Clone the repo
- Run
poetry install
- Run
run.py
orscripts\launch.sh
orcd docker;docker-compose up -d
- Clone the repo and in
docker
folder, rundocker-compose build
. The docker image will be build - Push to registry or use your custom publishing method to publish the image
- Start the application
- Go to
localhost:8000/docs
for swagger andlocalhost:8000/redoc
for redoc
define
for defining projects and its dependenciesstore
for storing notebook and associated filesrun
for executing a jupyter filehtml
to get a rendered page of executed notebookoutput
to get the output of Jupyter execution in json formatplain_text
to get the plain text output
edit
for editing notebook in a sandboxview
for viewing notebook in sandbox and can run it, but not save the changes
update
for storing the next version of notebook fromedit
endpoint
- Provide live environment for editing and running jupyter
- Custom transformations for jupyter output
- Scheduled cleanup of created jupyter docker containers
- Change container implementation to podman or other rootless systems