Feature/single image positions

[Jammy2211]

For computing the multiple image of a lens model between fits, this accounts for the case where the model is technically in the single image regime at the source centre:

        If the standard point solver only locates one multiple image, finds one or more additional images, which are
        not technically multiple image in the point source regime, but are close enough to it they can be used
        in a position threshold likelihood.

        This is performed by incrementally moving the source-plane centre's coordinates towards the centre of the
        source-plane at (0.0", 0.0"). This ensures that the centre will eventually go inside the caustic, where
        multiple images are formed.

        To move the source-plane centre, the original source-plane centre is multiplied by a factor that decreases
        from 1.0 to 0.0 in increments of 1/increments. For example, if the source-plane centre is (1.0", -0.5") and
        the `factor` is 0.5, the input source-plane centre is (0.5", -0.25").

        The multiple images are always computed for the same mass model, thus they will always be valid multiple images
        for the model being fitted, but as the factor decrease the multiple images may move furhter from their observed
        positions.