search

spacetelescope / jwst

Python library for science observations from the James Webb Space Telescope
https://jwst-pipeline.readthedocs.io/en/latest/
Other
560 stars 167 forks source link

Discontinue use of drizpars reference file in favor of resample pars reference file #7080

Closed stscijgbot-jp closed 3 months ago

stscijgbot-jp commented 1 year ago

Issue JP-2682 was created on JIRA by Alicia Canipe:

Based on discussion at the DMSWG meeting today about JP-2580 (notes), we would like to investigate the possibility of discontinuing the use of the drizpars reference file now that we have resample parameter files. Previously, the benefit of the drizpars file was that it allows multiple entries for a given observing mode, e.g., it can change the values of pixfrac depending on how many images are being combined. We can't do that with a parameter reference file; however, the CalWG made a decision to only use pixfrac=1, meaning we shouldn't need that functionality any more (right Howard Bushouse ? Is that the only parameter with that feature being used?) This would simplify things for developers and the pipeline users.

Anton Koekemoer Swara Ravindranath Misty Cracraft  FYI 

stscijgbot-jp commented 3 months ago

Comment by Howard Bushouse on JIRA:

Code updates to remove the use the drizpars ref file in resample/resample_spec are in #8546

stscijgbot-jp commented 3 months ago

Comment by Howard Bushouse on JIRA:

Note for INS folks regarding use of jwst/master and the upcoming B11.0 (jwst 1.15.x): We should NOT have a problem with synchronization of code and pars ref file releases, despite the fact that as of right now jwst/master (and eventually jwst 1.15.x) don't use the drizpars ref file anymore. You will still get the same results as the current release, due to the fact that the current resample/resample_spec steps were effectively ignoring non-default param values in the drizpars ref file anyway. So once we remove them from use, there's no change to the results (other than not having a R_DRZPAR header keyword in the output). This is of course is the whole reason we started down this road of switching over to pars ref files to begin with.

So you can run the latest code in either of two ways, with two different results:

1) keep running everything as is, using the latest ref files in CRDS-OPS, in which case you'll get the same results as before (i.e. results that use weight_type=ivm).

2) run the latest code, but do so using the pars-resamplestep/pars-resamplespecstep ref files that're on CRDS-TEST, in which case you'll get different resampling results, due to the use of exptime weighting (at least for most modes).

stscijgbot-jp commented 3 months ago

Comment by David Law on JIRA:

I think it's now been reasonably well tested that the code is successfully using the resample pars reference files instead, so closing this ticket until/unless we run into any edge failure cases.