Self-consistent mass and radius for interior modules

[nichollsh]

The current iteration of the code requires the user to specify interior mass and radius in the configuration file. However, interior mass is calculated as part of the Aragog/SPIDER/Dummy step anyway. This means that allowing the user to input the interior mass leads to two physically-inconsistent values. This is currently only physically consistent with using planetary masses and radii near Earth values, since then the SPIDER solution will happen to be identical to that provided by the user.

The simplest solution to this is to use the interior model to calculate the interior radius based on the inputted interior mass. This PR achieves this and closes #253.

Because Aragog currently uses constant physical properties, it does not make a very good structure model, but it's still more physically consistent this way. SPIDER (and even the Dummy module) does a good job in this regard.

Minor changes:

• Provide aragog with initial temperature specified in config file
• Enable aragog core cooling boundary condition
• Configuration file for K2-18 b
• All SPIDER options are passed to SPIDER via spider.py, not through the .opts file. This is to ensure consistency with Aragog (in that all parameters are set via the Python code).
• Reworked plot_stacked into plot_structure which includes the core temperature, R_xuv, and R_obs. This also (importantly) plots the entire profile on the same radius scale, so it's clearer what the structure of the planet looks like.
• Does not run atmosphere module when atmosphere has already escaped, and goes straight to checking the break criteria.
• Do not allow all gases to be set as condensible when running AGNI, as this can lead to sum(VMR)=0 in some regions, which causes the model to crash.

Requires https://github.com/ExPlanetology/aragog/pull/29 to be merged first.

Next steps:

• Consider more realistic structure model
• Allow user to input radius instead of mass, which avoids this structure calculation entirely.

[nichollsh]

@timlichtenberg this is now ready for review I think. There are lots of changed files, but this is mostly a result of the changed tests. It makes the changes we discussed in the meeting earlier, plus a few bug fixes that I discovered in the last week or so.

[nichollsh]

Thanks for the review Tim - this is really useful. I'll make these changes tonight/tomorrow, and if you could have a look again tomorrow then that would be great.

[nichollsh]

@timlichtenberg please can you have another look at this again today? I have made the changes that you suggested and responded to the other comments above. I also had to update the test data.

I will then address this issue: https://github.com/FormingWorlds/PROTEUS/issues/267