feature/dspl positions

[Jammy2211]

Updates and refactors image-plane multiple image based positions likelihood penalty to support multiple lens planes (e.g. double einstein ring systems).

Multiple Source Plane Systems

A double source plane system is a lens system where there are multiple source-planes at different redshifts, meaning that
incuding the image-plane there are at least 3 planes.

The PositionsLH class can have a plane_redshift input, which specifies the redshift of the source-plane
the positions are ray-traced to.

Multiple PositionsLH objects can be passed to the Analysis object, which then applies the penalty term to
both source-planes independently such that a double source-plane system can be fitted with the penalty based likelihood
functionality.

    positions_likelihood_source_plane_0 = al.PositionsLH(positions=positions, threshold=0.3, plane_redshift=1.0)
    positions_likelihood_source_plane_1 = al.PositionsLH(positions=positions, threshold=0.3, plane_redshift=2.0)

    analysis = al.AnalysisImaging(
        dataset=dataset, positions_likelihood_list=
            [
                positions_likelihood_source_plane_0,
                positions_likelihood_source_plane_1
            ]
    )

To set up an Analysis object with multiple PositionsLH objects from a result, each positions likelihood is
computed from the result for each plane_redshift:

    analysis = al.AnalysisImaging(
        dataset=dataset,
        positions_likelihood_list=[
            source_lp_result.positions_likelihood_from(
                factor=3.0, minimum_threshold=0.2, plane_redshift=source_lp_result.instance.galaxies.source_1.redshift,
            ),
            source_lp_result.positions_likelihood_from(
                factor=3.0, minimum_threshold=0.2, plane_redshift=source_lp_result.instance.galaxies.source_2.redshift,
            )
        ],
    )