Refactored reduce_across_dimension to get_extreme_coord_along_dimension

[GoThrones]

Overview: What does this pull request change?

This PR improves the code of the erstwhile method reduce_across_dimension, and renames it to get_extreme_coord_along_dimenstion.

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

The erstwhile method's name: "reduce_across_dimension" was not intuitively named. This method is only called by one other method named length_over_dim() which is also in mobject.py file, in the entire source code of manimce, and the method's name was not intuitive enough as to what it did.

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

[behackl]

Hi! IIRC length_over_dim is one of the most-called methods in a standard rendering flow. Did you benchmark in how far your changes affect performance / rendering time?

[GoThrones]

Here' the benchmarking code that i used:

from manim import *
import timeit

class BenchmarkScenePR(Scene):
    def construct(self):
        mob = Square()
        self.add(mob)

        results = timeit.repeat(
            lambda: mob.get_extreme_coord_along_dimension(np.max,0),
            number= 100000,
            repeat = 10
        )
        best = min(results[1:])
        print(f"100k calls: {best:.4f}s  ({best/100_000*1e6:.2f} µs/call)")


class BenchmarkScenemain(Scene):
    def construct(self):
        mob = Square()
        self.add(mob)

        results = timeit.repeat(
            lambda: mob.reduce_across_dimension(np.max,0),
            number=100000,
            repeat = 10
        )
        best = min(results[1:])
        print(f"100k calls: {best:.4f}s  ({best/100_000*1e6:.2f} µs/call)")
        
        

For main: time was 4.73 µs/call and for PR branch it was 4.68 µs/call.