Improve performance for :meth:.Mobject.become

[Darylgolden]

Overview: What does this pull request change?

• Added a boolean parameter, skip_point_alignment to the align_data method. When True, the expensive point alignment method is not called.
• Added documentation for the align_data method.
• Added three tests testing the become method.

Motivation and Explanation: Why and how do your changes improve the library?

While profiling the following scene:

class Test(Scene):
    def construct(self):
        nplane = NumberPlane()
        b = ValueTracker(1)
        a = always_redraw(
            lambda: nplane.plot_implicit_curve(
                lambda x, y: x ** 3 - y ** 2 - x - b.get_value()
            )
        )
        self.add(nplane, a)
        self.wait()
        self.play(b.animate.set_value(-1))
        self.wait()

I noticed that Manim was spending an inordinate amount of time on the become method of Mobject. In particular, when using the become method for VMobjects, a lot of time is wasted aligning the points of the VMobject to its target. This is despite the fact that the become method simply changes a VMobject's point data to match its target, with no need to interpolate where its initial points should lie (that will immediately get overwritten). Skipping point alignment caused the render time for the above scene on medium quality on my PC to drop from 80 seconds to 15 seconds with Cairo. For OpenGL, the time spent on the become method was already negligible, so I left it untouched.

Links to added or changed documentation pages

Further Information and Comments

Reviewer Checklist

• The PR title is descriptive enough for the changelog, and the PR is labeled correctly
• If applicable: newly added non-private functions and classes have a docstring including a short summary and a PARAMETERS section
• If applicable: newly added functions and classes are tested

[Darylgolden]

Docs build has dropped from ~980 seconds to ~930 seconds 💯

[behackl]

LGTM! Left one suggestion for documentation, but other than that this looks good to me.