Allow NumPy-like multi-indexing and slicing of Mobjects

[chopan050]

Closes #2916

This allows using multi-indexing, multi-slicing and fancy-indexing of Mobjects, in a similar way to how NumPy arrays can be indexed.

For example:

mobs = [VMobject(name=f"Mob{i}") for i in range(10)]
vgroups = [VGroup(*mobs[1:3]), VGroup(*mobs[3:6]), VGroup(mobs[7])]

base_mob = mobs[0]
base_mob.add(vgroups[0], vgroups[1], mobs[6])
mobs[3].add(vgroups[2], mobs[8])
mobs[8].add(mobs[9])

"""
Structure:

Mob0
 ├──VGroup [0]
 │   ├──Mob1 [0, 0]
 │   └──Mob2 [0, 1]
 ├──VGroup [1]
 │   ├──Mob3 [1, 0]
 │   │   ├──VGroup [1, 0, 0]
 │   │   │   └──Mob7 [1, 0, 0, 0]
 │   │   └──Mob8 [1, 0, 1]
 │   │       └──Mob9 [1, 0, 1, 0]
 │   ├──Mob4 [1, 1]
 │   └──Mob5 [1, 2]
 └──Mob6 [2]
"""

• base_mob[0, 0] returns Mob1
• base_mob[1, 0, 0] returns VGroup(Mob7)
• base_mob[:2, 0] returns VGroup(Mob1, Mob3)
• base_mob[:2, 1] returns VGroup(Mob2, Mob4)
• base_mob[1, 0, :, 0] returns VGroup(Mob7, Mob9)
• base_mob[[2, 0]] returns VGroup(Mob6, VGroup(Mob1, Mob2))
• base_mob[[True, False, True]] returns VGroup(VGroup(Mob1, Mob2), Mob6)
• etc.

See the added test for more examples of this.

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