search

desihub / desitarget

DESI Targeting
BSD 3-Clause "New" or "Revised" License
18 stars 23 forks source link

some target selection cuts for MWS missing from cuts.py #139

Closed moustakas closed 7 years ago

moustakas commented 7 years ago

Several of the targeting bits for the MWS (including those for white dwarfs) are assigned in mock.selection but missing from cuts, the latter of which contains all our "official" cuts. We should only ever use cuts, so we know we're applying the same target selection criteria everywhere, including on the mocks (where possible).

Any volunteers to rectify this injustice? @forero @apcooper @geordie666?

geordie666 commented 7 years ago

I'll try to get to this over the weekend, if nobody has time today.

geordie666 commented 7 years ago

I started to look through the mocks code this morning. It seems like there "is no" set of selection criteria for the MWS MAIN or the MWS WD sample. In mock.selection information is drawn directly from the mock file as to whether something is a MWS or a WD. Presumably, in desitarget.cuts we will want the actual selection cuts to apply to GAIA input files, rather than having an input file that simply has a list of positions with MWS or WD turned on or off. I dug around to find the MWS cuts and they aren't, e.g., at https://desi.lbl.gov/trac/wiki/TargetSelectionWG/TargetSelection. So, maybe we don't have the knowledge yet to put the MWS cuts into desitarget.cuts? Guidance @moustakas? Are there MWS cuts lying around?

Note, also, that the FSTD cuts appear to be applied (identically) in both mock.selection and desitarget.cuts. So, nothing new seems to be needed in desitarget.cuts and I'd suggest that people working on the mocks side of things edit the mock.selection module. All it would seem to require would be a call to desitarget.cuts.isFSTD_colors instead of the line

select_color = (gmr - grcolor)**2 + (rmz - rzcolor)**2 < colortol**2

I don't want to break the formalism of @forero, though

moustakas commented 7 years ago

@geordie666 Thank you very much for reviewing the code. I agree that the (F) standard-star selection can be done using existing code, and I've already implemented that change.

As for MWS, I'm hoping @apcooper @crockosi @callendeprieto can help us out. Is it the case that for the full-survey data challenge we want the various samples to be selected perfectly from the mocks, or are there actual target selection cuts floating around I'm not aware of? This is definitely an agenda item we should talk about at the next Targeting telecon.

moustakas commented 7 years ago

There was significant traffic on this question on-list and for the full-survey Data Challenge we've essentially confirmed (with the relevant Working Groups) that we will be using a 'perfect' selection of MWS targets based on the mocks themselves. We should reopen an issue on this topic once we have Gaia/DR2 in hand.