Cluster pgen Cleanup

[forrestglines]

We identified a few aspects of the cluster setup in #2 that could be cleaned up, although they are low priority

• Unify the calculation mean molecular masses from helium mass fraction
• Simplify/avoid the passing around of hydro_pkg for the different cluster functions
• k_0 and k_100 have bad orders of magnitude, <-100. This could be partially rectified by dividing these quantities by mean particle mass
• Make determining the radial points where the 1D HSE profile is integrated more robust without the r_sampling parameter

[pgrete]

• check source term handling of (updated) prims and cons
• add safety check for floors and ceilings so that a negative values translate to deactivate
• document (cluster) floors/ceilings
• use periodic bounds for default cluster pgen
• (optional) extend init to MeshData (@forrestglines : "the HSE profile piece would be the trickiest to extend to multiple meshblocks instead of just one meshblock. But not that hard, you would just need to integrate a profile that extends all SMR levels")
• add option to specify length scale instead of wavenumber for initial perturbation
• add safety check so that jet thickness matches resolution