Adding Holoviews/Datashader to Streamlit

Hello, I have funding support to add holoviews objects to Streamlit if that would be something doable in 40ish hours. More specifically: I love everything about Streamlit, but when I convert an xarray.hvplot to bokeh, it no longer re-renders decimate/rasters produced by holoviews’ datashader. ie, If I decimate to 1,000 data points and zoom in, it does not re-render to 1,000 data points. If I do this in Panel, it will re-render to 1,000 points when zooming in. Is adding datashader/holoviews interactively something doable?

HoloViews provides extensive dynamic interactivity built in Python that lets it support dynamically rerendering Datashader plots, dynamically decimating data on a zoom, dynamically linking selections across plots, etc. The HoloViz team would be very happy for these features to work in Streamlit, but the last we looked into it, Streamlit did not support transmitting Bokeh events back from JavaScript into Python to invoke Python code, which is necessary for any of this functionality. That seems like something difficult for an outsider to add to Streamlit, but I hope that it’s on Streamlit’s roadmap?