Demonstrates a NodeJS client using jupyter-js-services sending Python or Scala code to a kernel gateway, through an nginx proxy, with basic auth configured. The code examples use the Spark API.
Note that nginx only exists here to show that the kernel gateway can run behind a common proxy. (You can connect the client directly to the kernel gateway of course!)
git clone https://github.com/jupyter/kernel_gateway_demos
cd kernel_gateway_demos/node_client_example
docker-compose build
docker-compose up
When you run docker-compose up the first time, it will launch an nginx proxy, a kernel gateway, and the client with all logs aggregated. The client will send the src/example.py code to the kernel gateway for execution. Once the client completes, it will exit. If you want to run the client again without restarting the other two services, do this in another terminal:
docker-compose run -e DEMO_LANG=<scala|python|r> client
You can also run the src/client.js on your host without Docker if you npm install its dependencies and configure it with the proper environment variables. See what the docker-compose.yml and Dockerfile.client do as a reference.