Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: meas_mosaic
-
Labels:None
-
Story Points:2
-
Epic Link:
-
Sprint:DRP F18-4, DRP F18-5
-
Team:Data Release Production
Description
The implementation of RFC-322 in DM-10302 saw the renaming of the flux fields in the catalogs from *_flux to *_instFlux along with an increment in the afw table VERSION to 3. Aliases are added on read of catalogs that have VERSION < 3 for backwards compatibility. However, the getFluxKeys() function in meas_mosaic cannot take advantage of these aliases as written, so meas_mosaic is currently broken for VERSION < 3 catalogs. This needs to be fixed.
Attachments
Issue Links
- relates to
-
DM-10302 Rename "*_flux" fields to "*_instFlux" in SourceCatalogs
- Done
Activity
That's the plan...but it's currently slightly complicated by the alias issues of DM-15891 in that I end up picking up the likes of:
base_SdssShape_flux_xxinstFlux
|
base_SdssShape_flux_yyinstFlux
|
base_SdssShape_flux_xyinstFlux
|
if I search on "instFlux". These go away if I search on "_instFlux", but when I try tagging the "Err" on the end to get the flux error keys, I end up picking up two keys per aperture flux:
base_CircularApertureFlux_3_0_instFluxErr
|
base_CircularApertureFlux_3_instFluxErr
|
base_CircularApertureFlux_4_5_instFluxErr
|
base_CircularApertureFlux_4_instFluxErr
|
...
|
For now I plan to just search on "_instFlux" to get the flux keys and then manually tag the Err on the end to look up the associated error key (which I can update once DM-15891 gets resolved).
I can confirm that this runs on both old (RC/w_2018_32/DM-15184/, which is pre both calib_psfUsed->calib_psf_used & flux->instFlux) latest (RC/w_2018_41/DM-16011) reruns.
Specifically, I ran the following:
mosaic.py /datasets/hsc/repo/ --rerun RC/w_2018_41/DM-16011/:private/lauren/mosaic --numCoresForRead=4 --id tract=9813 ccd=0..8^10..103 filter=HSC-I visit=1228..1232:2 |
mosaic.py /datasets/hsc/repo/ --rerun RC/w_2018_32/DM-15184/:private/lauren/mosaicBack --numCoresForRead=4 --id tract=9813 ccd=0..8^10..103 filter=HSC-I visit=1228..1232:2 |
See comments on DM-15569.
Why does the meas_mosaic method have to be so complicated? Just have it return all keys that end in instFlux and instFluxErr and be done with it.