Feature/magnification map

[Jammy2211]

Refactors the inversion module to make it simpler to compute magnifications for PyAutoLens.

See this docstring for a run throuhg of the main changes:

The pixelized source reconstruction output by an Inversion is often on an irregular grid (e.g. a
Voronoi triangulation or Voronoi mesh), making it difficult to manipulate and inspect after the lens modeling has
completed.

Internally, the inversion stores a Mapper object to perform these calculations, which effectively maps pixels
between the image-plane and source-plane.

After an inversion is complete, it has computed values which can be paired with the Mapper to perform calculations,
most notably the reconstruction, which is the reconstructed source pixel values.

By inputting the inversions's mapper and a set of values (e.g. the reconstruction) into a MapperValued object, we
are provided with all the functionality we need to perform calculations on the source reconstruction.

We set up the MapperValued object below, and illustrate how we can use it to interpolate the source reconstruction
to a uniform grid of values, perform magnification calculations and other tasks.