VSCode integration

Visual Studio Code is currenlty the most popular integrated development environment (IDE).
Purdue AF + VSCode use cases
Running local notebooks with remote Conda environments (Jupyter kernels), which ensures that you use exactly the same versions of packages without a need to reinstall the environment.
Scaling out from you computer to 100s cores using Dask Gateway instances at Purdue AF.

Installation instructions
Install VSCode and JupyterHub extension
Open the “Extensions” panel in VSCode sidebar.
Search for
JupyterHub
extension and install it.
Create or open a Jupyter notebook
Open a local Jupyter notebook that you want to use with Purdue AF, or create a new notebook.
You may need to install
Jupyter
extension for VSCode for better experience.
Obtain authentication token for your AF session
In a web browser, login to Purdue AF and start a session.
Go to
File -> Hub Control Panel
.Click
Token
tab in top left of the page.Click
Request new API token
to obtain the token string - you will need it in the next step.
Connect your notebook to AF session
Switch back to the notebook opened in VSCode.
In the top right corner of the notebook, click
Select kernel
button, which will open Command Palette.In the Command Palette, select
Existing JupyterHub Server
.When prompted for URL of the server, paste
https://cms.geddes.rcac.purdue.edu
.When prompted for username:
If you are using Purdue account, type your Purdue username.
If you are using CERN account, type your CERN username followed by
-cern
.If you are using Fermilab account, type your FNAL account followed by
-fnal
.
When prompted for token or password, paste the token obtained in step 3.
Type any name (e.g. “Purdue AF”) to save the JupyterHub server setup for future.
Select kernel
Once setup is complete, you will be able to choose from the Purdue AF kernels, including default Python kernels, as well as any custom Conda environments that you normally have access to.
To change kernel in the notebook, simply click on
Select Kernel
in top right corner, and choose from recently used kernels or clickSelect Another Kernel
and thenExisting JupyterHub Server
. You will not need to repeat steps 4.4 - 4.7.
(optional) Start a Dask Gateway cluster
Follow instructions to start Dask Gateway cluster from a Jupyter notebook - they will work in your local notebook in VSCode too, but interactive widgets will not be displayed.