[BUG] Not able to plot using cuxfilter after activating cudf.pandas kernel

[patelmiteshn]

Describe the bug
using cuxfilter for plotting issues the following error: ValueError: ClassSelector parameter 'InteractiveDatashader.source_df' value must be an instance of (DataFrame, DataFrame).

The notebook is modified from this version.

The modifications were made to incorporate zero code change with cudf.pandas kernel

Steps/Code to reproduce bug
To reproduce the bug follow this notebook

Expected behavior
The script should be plotting a line graph

Environment details (please complete the following information):

• Environment location: [Bare-metal, Docker, Cloud(specify cloud provider)]
  Ubuntu 22.02 on local desktop with RTX A6000
• Method of cuxfilter install: [conda, Docker, or from source]
  pip install (https://docs.rapids.ai/install)

Additional context
Add any other context about the problem here.

[exactlyallan]

We are waiting on an upstream fix that enables datashader to work cudf.pandas - rapidsai/cudf#15694
I suspect the above is related.

[Matt711]

The proxy DataFrame returned from InteractiveDatashader.source_df when cudf.pandas is turned on is not an instance of cudf.DataFrame but is of pd.DataFrame.