Closed minrk closed 3 months ago
@minrk for the pre-MVP the Docker Client Partner python application has a lightweight built in SQLite-based delegate that will house your credentials as well. However, for the post pre-MVP our plan is to build an additional SMART on FHIR authentication server option for authentication to the data. FYI - I just pushed up the architecture and technical documents for the TCP portion of the pre-MVP build.
I just pushed up the architecture and technical documents for the TCP portion of the pre-MVP build.
Great! Where is that?
I just pushed up the architecture and technical documents for the TCP portion of the pre-MVP build.
Great! Where is that?
@minrk issue #6 has the links https://github.com/jupyterhealth/jupyter-health-software/issues/6#issuecomment-2015793003
@minrk - is there a specific email address I need to send the credentials via 1Password? Also, if I remember correctly you wanted two sets of credentials setup. If you could provide the Name you would like me to use for each credential and if you have a logo file you would like to use for each send that along as well, otherwise will default to UCSF logo.
you can use benjaminrk at gmail.
For names, let's use:
Jupyter Health doesn't have a logo yet, so if that can wait for now, that would be great.
Thanks @minrk I just sent you an email with the credentials.
As I understand it, JupyterHub will be an OAuth client. These client credentials need to be issued by TCP and stored somewhere, ultimately in a (sops?) secret in the 2i2c deployment.
For the pre-MVP, JupyterHub users will not be distinguishable as discrete users with their own access to the data server, and instead be viewed as a single 'jupyterhub' client. Exactly how that's going to work in terms of secrets and keys will become clearer when we have the docs in #6.