Work with non-gridded xarray dataset

[hoxbro]

Attempt at supporting request in holoviz/hvplot#1080.

This is very much still work-in-progress.

[codecov-commenter]

Codecov Report

Merging #5744 (925a7db) into main (14d7d9d) will increase coverage by 0.00%.
The diff coverage is 96.42%.

@@           Coverage Diff           @@
##             main    #5744   +/-   ##
=======================================
  Coverage   88.18%   88.18%           
=======================================
  Files         306      306           
  Lines       62739    62762   +23     
=======================================
+ Hits        55326    55348   +22     
- Misses       7413     7414    +1     

Impacted Files	Coverage Δ
holoviews/core/data/xarray.py	88.96% <96.42%> (+0.09%)	⬆️

... and 1 file with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[hoxbro]

My basic thought process is to treat data with empty vdims as tabular data, which makes it possible to do the following:

import holoviews as hv
import numpy as np
import xarray as xr

hv.extension("bokeh")

ds = xr.Dataset(
    data_vars={
        "lat": ("trajectory", np.linspace(30, 35, 15)),
        "lon": ("trajectory", np.linspace(16, 20, 15)),
        "example": ("trajectory", np.linspace(16, 20, 15)),
    },
    coords={"trajectory": np.arange(15)},
)

hv.Dataset(ds, kdims=["lat", "lon"], vdims=[]).to(hv.Points)  

And when not setting vdims to an empty list we get this (hv.Dataset(ds, kdims=["lat", "lon"]).to(hv.Points)):