JupyterLab, prodigy and extension: mixed content issue

[Sandor (Sandor Szalma)]

Hi,

I try to setup prodigy on a GCP AI platform Notebook (JupyterLab) running together with the JupyterLab extension for prodigy.

Installation is done, prodigy is working.

Prodigy uses protocol http.

JupyterLab uses protocol https and blocks mixed content, so the extension is blocked, too.

Do you have an idea, who this problem can be resolved?

Thanks and Regards,
Sándor

[ines (Ines Montani)]

Hi! One option would be to serve Prodigy via HTTPS – for example, by putting a reverse proxy in front of it:

Prodigy with https

I would recommend adding a reverse proxy in front that is restricted to only https connections. For instance, you would advertise the service on localhost, and have something like Nginx, Traefik or Apache communicating externally, via https.

You could also edit the app.py included with Prodigy and add your certificate:

jupyter-notebook: how to ssl the prodigy output?

yes Ines thats right. Adding some lines in the in the prodigy/app.py at uvicorn.run let it work. ssl_keyfile="/address/.Certs/key.pem", ssl_certfile="/address/.Certs/cert.pem", ssl_ca_certs=None, ssl_ciphers="TLSv1", ssl_version=2, ssl_cert_reqs=0, But it doesn't solve the prob with the tab in the jupyter lab.

[Sandor (Sandor Szalma)]

Hi,

thanks for your reply!

At the end we found this solution/workaround (without the extension):

• define port forwarding with the compute engine, docker image with JupyterLab (port 6006)

• configure prodigy to use this port

• start tunneling from my pc (cloud compute --project ppp ssh --zone europe-westx "optimised-container-xxx" --internal-ip -- -L 6006:localhost:6006)

• open browser with http://localhost:6006