exclude PS1 filters not covered by PFS spectra in the focal plane fitting

The flux calibration in run15/20240328 is very bad. The focal plane variation correction is not working very well and I suspect at least part of the problem is due to the missing arms; we were missing n1, b3, and n4 in the March 2024 run (run15). The plot attached here illustrates that we cannot compute synthetic magnitudes in the g and r-bands when we are missing the b-arm, and y-band when we are missing the n-arm. These filters have to be excluded from the focal plane fitting. We do not have a record of missing arms in pfsConfig, and we may have to infer missing arms from the data.