Azure Databricks Compatability

[ayjayt]

Not sure how to circumvent this new requirement in a serverless compute on Azure Databricks...

It seems chromium cannot be accessed in those conditions and will always end up in error if trying to save plotly plots as static images.

Originally posted by @pdmkdz in #224 (comment)

Hi @pdmkdz,

I will investigate this and report back. For now, if its really impossible to pull in chrome, you can probably downgrade Kaleido to a v0.1.xx or maybe a v0.2.xx- but I will have a better answer for you this week.

[pdmkdz]

Thanks for opening this, I was about to open an issue, perfect timing.
I am pinning to %pip install kaleido==0.2.1 for the time being, and works fine.
I contacted the Databricks reps to see if we can circumvent these issues and have the headless browser, but my attempts have been pointless when running from a notebook browser using a serverless compute.
So to be able to run it in Databricks gotta be self contained like in the past.

[dgibbon-plotly]

@pdmkdz Can you explain the issue to me in layperson terms? Also, I am adding @sachds as he has been working recently on some serverless databricks notebook stuff

[pdmkdz]

@dgibbon-plotly , @sachds It seems that while using a serverless compute I cannot use shell commands.
I asked even the Databricks reps and said that there is no support for it at this time.
So for an automated Job run in serverless I won't be able to download, install and set variables for chromium with the new version of Kaleido on each new job compute.
I do not even know the technicality of it on why shell commands are not allowed, so cannot go to explain further than this.
Let me know if you need anything else.

[pdmkdz]

My Databricks rep will pull in the plotly rep they work with to try to look into this as well (no clue on who exactly) and will report this internally to be looked into their side as well.

[dgibbon-plotly]

Hey @pdmkdz My email is dave@plot.ly ... let's connect directly. I know who covers your org for each of Databricks and Plotly, and we're also doing some other things w Databricks notebooks you might be intrigued about, and for which we'd like to get your feedback.