Change conventions for host/subhalo weighting in weighted_lc function

[aphearin]

This PR changes our conventions for the weights that get attached to subhalo lightcones by the weighted_lc function. There are two new field names:

• halopop.cen_weight equals the HMF weight for centrals, and for satellites equals the HMF weight of the associated host
• halopop.sat_weight equals 1 for centrals, and equals <Nsat(Msub) | Mhost>

As a result of this PR, there is a new way to compute weighted histogram results with twndhist and signdhist. The weights that should be used for histograms are now:

gal_weight = halopop.cen_weight * halopop.sat_weight

Note that this is different from #37. This PR supersedes #37.

The diffsky PR ArgonneCPAC/diffsky#416 establishes successful CI testing against the code in this PR. So we can proceed to merge this in, do a new minor release of diffhalos, and then merge ArgonneCPAC/diffsky#416 into diffsky, and finally do a new release of diffsky.