Remove check for sizes=0 in legend_elements

[Illviljan]

PR Summary

This check removed legend labels that have sizes close to 0 in the legend. So now that I removed it, everything that's in the plot will always be shown in the legend labels as well which makes sense to me and is easy to test.

s=0 values are still in the plot being calculated with and slowing the down the plot. So using s=0 to remove "bad" points from plot and legend seems like a bad habit that doesn't need encouragement. If you want to filter it should've been done before creating the scatter like:

x = x[np.isclose(s, 0)]
y = y[np.isclose(s, 0)]
s = s[np.isclose(s, 0)]
ax.scatter(x, y, size=s)

One could also argue that we are just fixing a bug because:

• It's not documented in the docs.
• It's not tested for.
• Users can't change the np.isclose() parameters, so maybe there's some user out there working with small numbers and needs to change the parameters. So it's better to just let the user filter out points close to 0 before plotting the data.

PR Checklist

• Requires Allow the func argument in PathCollections.legend_elements to return strings #19556
• Has pytest style unit tests (and pytest passes).
• [N/A] Is Flake 8 compliant (run flake8 on changed files to check).
• [N/A] New features are documented, with examples if plot related.
• [N/A] Documentation is sphinx and numpydoc compliant (the docs should build without error).
• Conforms to Matplotlib style conventions (install flake8-docstrings and run flake8 --docstring-convention=all).
• [N/A] New features have an entry in doc/users/next_whats_new/ (follow instructions in README.rst there).
• API changes documented in doc/api/next_api_changes/ (follow instructions in README.rst there).