Implement mask generation into pipeline / phase module

[Jammy2211]

Previously, the mask was included as part of the CIData class, and it was extracted for parallel / serial calibration alongside the image, noise-map, etc.

I have decided to remove the mask from the CIData, as like AutoLens it is something we will want to update through the pipeline. Thus, we should have it set using a mask_function in phase.py, where the default_mask_function corresponds to a completely unmasked image (e.g. all False) with the same shape as the image. This mask is then extracted to the same cut-out as the image, noise-map etc and included in the 'ci_data_anallysis'.

To make it so that we can cleanly extract the mask we need to include it as in input variable in the CI_Data '_calibration_data' routines, and I guess we need these routines to output a CIDataAnalysis class which includes the (extracted) mask.

[Jammy2211]

I have added mask_function to all phases in an analogous fashion to autolens. I have also made it so that a list of masks can be supplied using the phase / pipeline run functions, which are used if a mask_function is not supplied to the pipeline.

The mask is attatched to the ci_data_fit class, but this is done ad-hoc and needs to be implemented in the CIData class. I have put a todo in phase.py here. I'm hoping that 'hack' will allow me to get an integration test running soon.