Include old calibration files

[griffin-h]

I mentioned to @cmccully and @jfrostburke that I think we should not delete old calibration files (arc identifications, atmospheric corrections, sensitivity functions) because of two considerations: (1) reducing old data, where the old calibrations are the right ones to use, and (2) being able to reproduce old reductions, where the filenames of the calibration files are stored in the headers of the reduced data. I'm pretty sure this is how it used to work a few years ago, but at some point we started replacing old calibrations instead of just adding new ones.

I confirmed that the pipeline does choose the calibration taken closest in time to the data to be reduced, not just the first one alphabetically (e.g., see the function floyds.util.searchsens). This is mostly correct, with the only exception being when there is a drastic change like recoating the mirror or changing the CCD. This is an edge case that we already didn't account for, so I am going to ignore it for now.

I went through the git history and recovered as many calibration files as I could and added them under the new directory structure (accounting for en05 vs. en12). This commit simply adds all the files I found back into the repository.

[griffin-h]

I should note there was one file already in the ftn/en06 directory that had 2m0-02/en12 in the header, so I deleted that file. Not sure how it got there (or maybe the header is wrong).