Tng correctunits

[alexalar]

Correcting the units of mass of halo MAHs for TNG, so that the output MAHs are in Msun. For more details see diffmah PR#185

The main changes are:

• Updated TNG data loaders.
• Updated diffstarpop params coming from fits to 2) individual diffstar fits; and 2) from fitting DiffstarPop to processed target data.
• Some updates to diffstarpop fitting scripts to adapt to new signature of Diffstar parameters.

[aphearin]

Just checking: is it necessary to re-run our optimizations in light of #116? Or is this unnecessary? Your call, just double-checking before this gets merged.

[alexalar]

We don’t need to re-run our optimizations for the results in the released paper. In the DiffstarPop paper, we decided to neglect satellite-specific quenching that depends on merger histories, leaving that for future work (and we discuss that somewhere in section 5). But as a result, the released DEFAULT_DIFFSTARPOP_PARAMS are a mix of calibrated and uncalibrated parameters. If someone supplies halos with sensible values for the lgmu_infall,logmhost_infall,gyr_since_infall paremeters, then the resulting SFHs will be different from the ones in the paper. To reproduce the results from the paper, one should set e.g. gyr_since_infall = -99.0 to all halos, making sure the satquench model has no effect.

[aphearin]

Looks ready to merge to me! Once you merge this in, we can do a new minor release of diffstar.