There was an issue where the halobox outputs were not properly saved to file in Coeval objects
this was partly due to the naming of fields and variables in the boxes, which rely on the class names converted to snake case. There are a couple instances of classes having specific kwarg strings that I don't think is great long-term so hopefully we can change that in the near future
If there were no halos (for example above Z_HEAT_MAX), the mean turnover mass for minihalos was set to zero. Since this global was used for the optical depths in spin temperature, the impact of this numerically was minimal, but it's possible that it caused some integration errors in certain circumstances
This PR fixes two small bugs: