Continue Bokeh 3.0 update

[ndmlny-qs]

PR #51 began the work necessary for updating to Bokeh 3.0. @hoxbro points out in PR #77 that we did not fully complete the updates required for Bokeh 3.0.

• bump Bokeh versions
• fix breaking changes from Bokeh 3.0 update

[theeldermillenial]

Is there an ETA on this? Is there any way I can contribute? It seems like basically all the major interactive graphing libraries are broken in ipywidgets 8 (altair, bokeh, plotly, etc) when exporting to html. This seems to be related to the way nbconvert is exporting things as json objects and the accompanying npm packages used to render those json objects. As far as I can tell, it's related to the general way these graphing libraries use divs and scripts. The scripts seem to not embed into json objects in a way that the jupyter-widgets npm package can handle them.

If you have this handled, excellent. I am going to figure out a way to do it for at least one graphing library. If I can contribute to an existing project rather than create my own hacky approach, I'm more than happy to.

[theeldermillenial]

Actually, I think I may want the jupyter_bokeh project rather than this one. I am trying to wrap bokeh in an ipywidget.

[ndmlny-qs]

@theeldermillenial just to update, I have a version of ipywidgets_bokeh with Bokeh 3.0 correctly serving a Python module as a webpage using bokeh serve .... There are a few visual bugs that still need to be ironed out, but when they are I'll update this ticket.

[theeldermillenial]

Related, I see jupyter_bokeh just had an update to fix some broken things. I've been testing it. I suspect some of the things I've dug through are probably related quirks.

There's a wall of text, but I think I got to the bottom of things.

bokeh/jupyter_bokeh#178