Removing astropy.constants.G from various calculations

[dlakaplan]

To improve consistency, removing astropy.constants.G from various derived quantities in favor of pint.GMsun or pint.Tsun. The differences are small, ~1e-10. And issues are still possible (e.g., if a mass is input in kg instead of Msun, it will still need to be converted). But this is better (I think).

[dlakaplan]

Note that there is still extensive use of c.G in the DDGR model. That should probably also be fixed.

[dlakaplan]

Note that there is still extensive use of c.G in the DDGR model. That should probably also be fixed.

Now this has been fixed too.

[dlakaplan]

As a related note, our value of G*Msun:

 Following are from here:
# http://ssd.jpl.nasa.gov/?constants (grabbed on 30 Dec 2013)
GMsun = 1.32712440018e20 * u.m**3 / u.s**2

has a newer value:

1.32712440041279419e20

from the website cited above. Fractional difference is ~1e-10. But we generally don't claim absolute precision at this level (for things like masses) so I'm not sure it's worth changing.

[abhisrkckl]

I will merge this if ready.

[dlakaplan]

I think it's ready if you don't have further issues (I agree some of it is not optimal, but not sure what else to do)