CosmoStat/shapepipe#886 makes ShapePipe write one boolean column per UNIONS external mask map (11 columns, MASK_1_Faint_star_halos … MASK_2048_z2) and cut on none of them. Which columns make up the default cut is sp_validation's choice; this issue is where to make it.
Open question for Martin: are the star halos (MASK_1, MASK_2) in the default? mask_r_nside131072.hsp equals the OR of the 1, 2, 4, 8, 64 and 1024 columns (pixel-exact on three interior granules of the Feb-2025 r ladder). On those three granules the halos add 14.5 / 6.3 / 18.5% of area, against 11.2 / 12.8 / 9.6% for the 4 + 8 + 64 + 1024 cut alone, so they make up about half of mask_r. The 64_r column has to be in any default (otherwise positions with no r data pass). An OR of all 11 columns is wrong, because the u/g/i/z/2048_z2 footprint columns mask whole regions that lack coverage in those filters.
Claude Opus 5.5 on behalf of Cail
CosmoStat/shapepipe#886 makes ShapePipe write one boolean column per UNIONS external mask map (11 columns,
MASK_1_Faint_star_halos…MASK_2048_z2) and cut on none of them. Which columns make up the default cut is sp_validation's choice; this issue is where to make it.Open question for Martin: are the star halos (
MASK_1,MASK_2) in the default?mask_r_nside131072.hspequals the OR of the 1, 2, 4, 8, 64 and 1024 columns (pixel-exact on three interior granules of the Feb-2025 r ladder). On those three granules the halos add 14.5 / 6.3 / 18.5% of area, against 11.2 / 12.8 / 9.6% for the 4 + 8 + 64 + 1024 cut alone, so they make up about half ofmask_r. The 64_r column has to be in any default (otherwise positions with no r data pass). An OR of all 11 columns is wrong, because the u/g/i/z/2048_z2 footprint columns mask whole regions that lack coverage in those filters.Claude Opus 5.5 on behalf of Cail