lsst-sims / legacy_sims_maf

LSST Simulations package for the metrics analysis framework (MAF)
13 stars 19 forks source link

U/lynnej/fbs wfd #175

Closed rhiannonlynne closed 4 years ago

rhiannonlynne commented 4 years ago

Add labelling for WFD, DD or "other" to the database output, using 'add_fbs_proposals'
(if you have a better suggestion than 'proposals' , happy to use instead .. but it fits one of the columns which is already there and didn't want to use 'survey').

Add healpix subset slicer, which basically lets you calculate WFD-only (or other healpix subsets) values. This is pretty similar to just setting the RA/Dec values for the UserPointsSlicer except that it also lets you calculate a power spectrum, as the entire healpix array is preserved.

rhiannonlynne commented 4 years ago

Before I added a WFD_bigsky, the Wfd_nogp was doing double duty, so they weren’t quite the same. But I agree that the defaults in the scheduler repo should be the defaults here now, and with the addition of that function I think you’re right to dump these and just use defaults.

It would be good to update the footprints stuff a bit more so that that code is using the defaults where possible too.

On Thu, Dec 19, 2019 at 9:32 AM Peter Yoachim notifications@github.com wrote:

@yoachim approved this pull request.

Looks good, maybe toss some of the get_wfd functions.

In bin.src/add_fbs_proposals.py https://github.com/lsst/sims_maf/pull/175#discussion_r359979990:

+ +def get_standard_wfd(nside=64):

  • wfd_footprint = schedUtils.WFD_no_gp_healpixels(nside, dec_min=-62.5, dec_max=3.6,
  • center_width=10., end_width=4.,
  • gal_long1=290., gal_long2=70.)
  • return wfd_footprint
  • +def get_extended_wfd(nside=64):

  • wfd_footprint = schedUtils.WFD_no_gp_healpixels(nside, dec_min=-72.25, dec_max=12.4,
  • center_width=14.9,
  • gal_long1=0, gal_long2=360)
  • return wfd_footprint
  • +def get_extended_wfd_dust(nside=64):

I'm a little worried that these get_*_wfd functions have a lot of magic numbers that could get out of sync with the scheduler repo. Do we really need to wrap one-line calls in utility functions?

It probably means it's finally time to make a footprint class rather than just using dicts. Then the footprint object can have extra info about where WFD is for convenience and the scheduler can just ignore those bits.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/lsst/sims_maf/pull/175?email_source=notifications&email_token=AAOFBX5AKCCXT6FHE4XF46TQZOV2XA5CNFSM4JUIVNWKYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCP2IBIA#pullrequestreview-334790816, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAOFBXZ6IZX6NG7J6HMPAEDQZOV2XANCNFSM4JUIVNWA .

-- ========================= Dr R. Lynne Jones Research Scientist, UW / LSST

she/hers =========================

rhiannonlynne commented 4 years ago

Deleting u/lynnej/fbs_wfd branch.