fix: Fixed to add_to_canvas() returns ax instead of plots

[i-aki-y]

I want to fix the error which is discussed in #920.

The plotting.show() function is a alias of the show_canvas() method and it expect for an ax object in matplotlib, while the GP.plot()(==add_to_canvas()) function returns plots dict.

See also:

GPy/GPy/plotting/matplot_dep/plot_definitions.py

Lines 95 to 97 in f63ed48

	def show_canvas(self, ax, **kwargs):
	ax.figure.canvas.draw()
	return ax.figure

To fix this, I just replace the return value of add_to_canvas() from plots to ax.

Note:
The return value of the add_to_canvas() function was changed at the commit 4f3047e from ax to plots.
Unfortunately, I could not find the reason why this fix was needed.

I would appreciate it if you reviewed this.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 54.51%. Comparing base (f63ed48) to head (2b3e705).
⚠️ Report is 306 commits behind head on devel.

Additional details and impacted files

@@            Coverage Diff             @@
##            devel     #989      +/-   ##
==========================================
+ Coverage   54.49%   54.51%   +0.01%     
==========================================
  Files         210      210              
  Lines       21560    21560              
  Branches     3169     3169              
==========================================
+ Hits        11750    11753       +3     
+ Misses       9255     9252       -3     
  Partials      555      555              

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.