Include light trap locations

[tskisner]

Eventually, after #105 is merged, we should also parse desi-0899 to get the locations of the light traps on each positioner. This should be included in the new focalplane file format so that it can be parsed by fiberassign.

[sbailey]

The traps are big enough that I don't think we need to worry about per-positioner differences, i.e. it could be a single number for distance along the phi arm and call it good.