Removed the mag_rate from the planar array

[micheles]

For separation of concerns. It should help with #10627.
NB: this is changing the ps_grid_spacing approximation by replacing occurrence rates with average occurrence rates.
Here are some numbers coming from the newzealand calculation in the oq-risk-tests:

(Pdb) ctx.occurrence_rate  # before/after
array([3.62667e-05, 3.67693e-05, 3.56305e-05, 3.73712e-05, 4.02654e-05,
       4.13410e-05, 3.80362e-05, 4.27079e-05, 4.66359e-05, 3.77293e-05,
       4.40878e-05, 4.99884e-05])
array([4.05691333e-05, 4.05691333e-05, 4.05691333e-05, 4.05691333e-05,
       4.05691333e-05, 4.05691333e-05, 4.05691333e-05, 4.05691333e-05,
       4.05691333e-05, 4.05691333e-05, 4.05691333e-05, 4.05691333e-05])

[micheles]

NB: check the newzealand test in the oq-risk-tests