fix: starry light curves of multiple bodies occultation + API test from #148

This PR should be tested and merged only after #151. It includes:

Fixing the light curve of the central occulted by multiple bodies (same as starry where the phase curve of the central is subtracted for each body)
starry-like output for system light curve: (1+num_bodies, num_times) (the shape management is a bit shady, I'd appreciate a review in there)
@dfm, SurfaceMapSystem has a default Map of amplitude 1.0 here. Do we want to consider SurfaceMapSystem with a None map?
Major refactor of light_curve_test.py that now includes:
- None maps for both central and/or body
- Multiple bodies note that starry only allows multiple bodies with similar 'pytrees'

The tests against starry for multiple bodies pass! And the rest is a good excuse to think about some API choices (#148)

exoplanet-dev / jaxoplanet