You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
final_cat.param requests IMAFLAGS_ISO that the unmasked tile chain never produces; sp_validation cuts on it #912
workflow/config/cfis/final_cat.param lists IMAFLAGS_ISO, but the committed tile detection runs with FLAG_IMAGE = False and default_noimaflags.param (config_tile_Sx.ini), so no tile catalogue carries the column and merge_final_cat reads one that is never there. Downstream, sp_validation's classification_galaxy_base cuts on IMAFLAGS_ISO == 0, so the v2 catalogue either fails that selection or the cut has to move to the healsparse MASK_<band> columns. This is the object-level half of the masking decision (the pixel-level half — epoch flag source — is settled). Decide: drop IMAFLAGS_ISO from final_cat.param and port sp_validation's cut to MASK_<band>, or re-enable tile flag images. Recorded as a [LINT] on detection.detection_source_mode in astra.yaml (#875).
Activity
docs(astra): retire lints resolved by #907, #909, #918
Record the pipeline's scientific decisions in astra.yaml, linked to t…
Merge origin/develop into feat/persist-exp-products