TODO list and bugs

• Examples demonstrating modules that one doesn't necessarily need to understand to model a lens via a phase / pipeline (e.g. grids, ray_tracing, pixelization, etc.)

• Spherical models do not inherite from SphericalProfile - add unit tests and correct this.

• Numerical integration is slow, because when scipy.quad is numba-fied it can only be called in a for loop. The large number of function calls leads to significant slow down. Solution: Need to come up with a new approach to numerical integration.

• Prior over-rides in the model_mapper change the ordering of parameters, which change the ordering of MultiNest output. This is fine, but information / visualization output labels are reordered in a different way, meaning that output is hard to follow / use. Solution: Build the prior id mappings into model_mapper.

• Parameter labels are drawn from the 'subscript' property in each class. This does not account for prior over-rides above and causes the code to crash as it infers there are too many parameters. Solution: Link the model-mapper to classes via a parameter-label config file, so that model mapper can account for this.

• Use an iterative grid for the LensSourceTracer when evaluating the image-plane lens galaxy light profile(s).