Proton event cannot be properly identified and masked

Part of a proton hit was found unmasked in visit 116099, r3, around (fiberId, wavelength) ~ (1409, 720). This particular event was masked along its fainter tail, but part of the very strong peak (with INTRP+SAT mask in green) due to a slow proton hit was left unmasked.

Currently, proton events in general are difficult to identify and mask due to the very large charge deposition and the resulting fatter profile than PSF. We may need a proton task to handle the proton events in addition to the cosmicray2 task to handle regular CR events.