Move HaloBox Fixed integrals to Lagrangian space

[daviesje]

The CMF integrals which occur between the minimum sampled halo mass and the minimum source mass currently are performed on PerturbedField.density, similar to the default ionisation model. They should be calculated on the evolved initial conditions InitialConditions.lowres_density instead (see Trac+22 https://arxiv.org/pdf/2109.10375).

At first we can assume that these small sources don't move from the ICs, but if we want to implement them for the larger halos we should perturb their positions using the 2LPT functions.

The change itself will be easy to make, but we will need to test how it effects these small sources, and if we set FIXED_HALO_GRIDS, the fixed source boxes.

[daviesje]

A stopgap for this has been implemented in #433. Where only the sources below the sampler are treated as non-moving and calculated from the evolved IC density, and FIXED_HALO_GRID uses the Eulerian density.