Allow Plot.label to control title(s)

[mwaskom]

Following on #2919 to support titles. It expands the API in a few ways:

1. The title= parameter allows setting directly a title for a single-axes plot

so.Plot().label(title="Here is a plot")

2. The title= parameter also can be passed a function that operates on the auto-titles in a faceting context

so.Plot().facet(["a", "b"]).label(title=str.capitalize)

3. The col= or row= parameters accept a string that adds a prefix to facet titles:

so.Plot().facet(["a", "b"], ["x", "y"]).label(col="Column", row="Row")

Note two relevant changes from the way titles work in FacetGrid:

• A label for the variable name is not added by default
• The order of the titles is {col} | {row}, corresponding to the change in the order of positional arguments (column faceting is more useful than row faceting so it should be prioritized)

I would like to add, and probably default to "margin titles" for the row variable, but that isn't implemented yet.

The combination of the faceting variables is not currently customizable with a template akin to what FacetGrid.set_titles accepts, but you can customize a single dimension:

so.Plot().facet([1, 2]).label(title="{}st column".format)

I see some potential for title vs col/row to be a little confusing so may revisit this during the experimental phase.

"Suptitles" and captions are likely in scope for this method too, but I don't want to commit to them yet because the interaction with matplotlib autolayouts is trickier.

[codecov]

Codecov Report

Merging #2934 (1562c2f) into master (762db89) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #2934   +/-   ##
=======================================
  Coverage   98.30%   98.30%           
=======================================
  Files          69       69           
  Lines       22954    22985   +31     
=======================================
+ Hits        22565    22596   +31     
  Misses        389      389           

Impacted Files	Coverage Δ
seaborn/_core/plot.py	97.33% <100.00%> (+0.03%)	⬆️
tests/_core/test_plot.py	98.56% <100.00%> (+0.02%)	⬆️