search

spacetelescope / jwst

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

Error in matching the NIRSpec Imprint image to the Science image for Mosaic IFU data #8215

Closed stscijgbot-jp closed 7 months ago

stscijgbot-jp commented 10 months ago

Issue JP-3518 was created on JIRA by Jane Morrison:

For NIRSpec IFU data taken in a Mosaic and with  dither pattern the correct imprint image is not being matched to the science image.  The pool file contains information on the mosaic in the  MOSTILNO column, but that information is not written to the file or is even part of the datamodel. This information should be written to the file and included in the FITS header and the datamodel. Using the mosaic information, dither pattern and observation number the correct imprint image can be selected. We can select the correct imprint either in the association generating code or the imprint step. There might be benefits to it on over location over the other.  

stscijgbot-jp commented 9 months ago

Comment by Howard Bushouse on JIRA:

Jane Morrison When you say that the mosaic tile number info "is not written to the file", which file are you referring to? If it's in the pool file, the ASN generator rules can make use it. It doesn't need to be in a FITS header keyword in order for the generator to use it. If MOSTLNO is not populated (e.g. set to NULL) in the pool files for NRS IFU observations, that's an issue that needs fixing and would likely need changes upstream somewhere in SDP, in order to have it populated values in the pool for NRS IFU exposures.

stscijgbot-jp commented 9 months ago

Comment by Howard Bushouse on JIRA:

I've grabbed the pool file for jw01192 and put it in ██████████████████████████████████████████ Looking at all the entries for the NIRSpec IFU exposures in obs-008 they all have populated MOSTILNO entries that run from 1 to 6 through the mosaic pattern. There are 6 files listed for each tile position, which is 2 detectors * 3 dithers. So it looks like the MOSTILNO entries make sense and are available to the ASN generator rules to make use of. So this suggests that the ASN rules are simply not doing the proper thing with those values.

There are 3 duplicate entries for each exposure, because each one of them has been assigned to 3 different ASN candidate ID's: "o008: OBSERVATION", "c1001: GROUP", and "C1006: BACKGROUND". So the duplicate entries may also be leading to the problems the ASN rules are having to correctly key off of MOSTILNO to match up science and imprint exposures.

stscijgbot-jp commented 7 months ago

Comment by Howard Bushouse on JIRA:

Fixed by #8410

stscijgbot-jp commented 6 months ago

Comment by Melanie Clarke on JIRA:

While working on JP-3629, I noticed that the fix for this ticket changed the association constraints for standard NIRSpec IFU data (Asn_Lv2Spec rules), but not for nodded IFU data (Asn_Lv2NRSIFUNod rules).

James Muzerolle Bethan James Christian Hayes - do you know if nodded IFU data also needs the same support for matching imprint images?  Or can you point me to some test data for this mode so I can investigate myself?

stscijgbot-jp commented 6 months ago

Comment by Melanie Clarke on JIRA:

At Christian Hayes's recommendation, I looked at p1257 obs4, which is a nodded IFU observation with leakcals.  The association files from MAST all look like they have the right imprints, matched with the science and the nod pair for background subtraction.  So, I think the Asn_Lv2NRSIFUNod rules are okay as is.

stscijgbot-jp commented 1 month ago

Comment by David Law on JIRA:

Melanie Clarke Christian Hayes Can this ticket be closed now?

stscijgbot-jp commented 1 month ago

Comment by Christian Hayes on JIRA:

Spot checking the PID 1192 obs 8 on MAST (reduced 1.15.1, which includes these changes), the correct imprints are matched with the correct dither/tile number science exposure in the association candidates.  This appears to be addressed now, closing.