Do not allow partial matches when updating OverlayPlot

[philippjfr]

For the longest time OverlayPlots tried to be smart and remappped dynamically created elements to subplots as long as they partially matched the plot spec. A plot spec consists of (<element-type>, <group>, <label>, <ndoverlay-dims>). Remapping plots like this can be more efficient but may also cause significant issues because a plot created for one element may have a bunch of state that is not correct for the new element. In theory all this state could be updated but we never did a good job at this, resulting in a variety of issues, e.g. subplots would end up assigning the wrong colors, legend order or a variety of other incorrect properties to a glyph/renderer.

Fixes:

import holoviews as hv
import pandas as pd
import panel as pn

hv.extension("bokeh")

color_dim = hv.dim("category").categorize(
    categories={"B": "red", "A": "blue", "C": "green"}, default="grey"
)
data = pd.DataFrame(
    {"x": range(5), "y": range(5), "category": ["A", "B", "A", "C", "B"]}
)


def inner(value):
    el = hv.Dataset(data).to(hv.Points, kdims=["x", "y"], groupby="category")
    if value:
        el = el.get(value)
    el.opts(color=color_dim, show_legend=False, xlim=(-1, 6), ylim=(-1, 6))
    return el.overlay()


w1 = pn.widgets.MultiSelect(value=["A"], options=list("ABC"))
pn.Row(w1, hv.DynamicMap(inner, streams=[w1.param.value])).servable()

[codecov-commenter]

Codecov Report

Merging #5962 (bc7f4f0) into main (345dd00) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #5962      +/-   ##
==========================================
- Coverage   88.58%   88.58%   -0.01%     
==========================================
  Files         315      315              
  Lines       65523    65507      -16     
==========================================
- Hits        58043    58027      -16     
  Misses       7480     7480              

Flag	Coverage Δ
ui-tests	23.14% <0.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
holoviews/plotting/bokeh/element.py	88.36% <100.00%> (-0.02%)	⬇️
holoviews/tests/plotting/bokeh/test_overlayplot.py	100.00% <100.00%> (ø)

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