This PR adds read_mtl_ledger support for the mtl/1.0.0 format, which doesn't have Z_QN but does columns ZS and ZINFO that are now gone.
Motivation: fiberassign needs latest desitarget to get the write_targets(..., subpriority=False) functionality. But, current desitarget doesn't support reading the mtl/1.0.0 format, which breaks the ability to test current fiberassign on mtl/1.0.0 files for comparison with fiberassign/4.0.0 on previously designed tiles. Thus this PR.
The solution here is admittedly fragile if the mtl format continues to evolve. I certainly hope it doesn't, but if it does change again and this becomes problematic, I suggest tackling a deeper refactor about the optimizations for speed-reading the ecsv file without requiring a specific single set of expected columns (via mtldatamodel).
Coverage decreased (-0.009%) to 58.91% when pulling 4fee94632f7b08e65e7f2d35395ea2968af9f11a on mtl-compatibility into 8e6894c7270c4f19fb8bac2c0b8921a0bd4d4950 on master.
This PR adds
read_mtl_ledger
support for the mtl/1.0.0 format, which doesn't have Z_QN but does columns ZS and ZINFO that are now gone.Motivation: fiberassign needs latest desitarget to get the
write_targets(..., subpriority=False)
functionality. But, current desitarget doesn't support reading the mtl/1.0.0 format, which breaks the ability to test current fiberassign on mtl/1.0.0 files for comparison with fiberassign/4.0.0 on previously designed tiles. Thus this PR.The solution here is admittedly fragile if the mtl format continues to evolve. I certainly hope it doesn't, but if it does change again and this becomes problematic, I suggest tackling a deeper refactor about the optimizations for speed-reading the ecsv file without requiring a specific single set of expected columns (via mtldatamodel).