[PIPE2D-793] Update sky subtraction to work across all fiber configurations Created: 23/Mar/21  Updated: 18/Jun/21  Resolved: 11/Jun/21

Status: Done
Project: DRP 2-D Pipeline
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Normal
Reporter: ncaplar Assignee: price
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Blocks
blocks PIPE2D-857 Apply sky subtraction, deduced from D... Done
Story Points: 3
Sprint: 2DDRP-2021 A5, 2DDRP-2021 A 6
Reviewers: hassan

 Description   

Updated sky subtraction code so that it can handle SuNNs data. At the moment, the code works by comparing position of the spot on the detector and provided locations in the 2d psf array and searching for nearest given the `xMaxDistance` parameter (see https://github.com/Subaru-PFS/drp_stella/blob/91432e704b39c1842f109b24e808a34e8332f796/python/pfs/drp/stella/NevenPsfContinued.py#L53). 

We should change the code that it works via a proper fiber identification mechanism, requesting the solutions from the provided array of 2d point spread function via fiber id. This will probably involve some work and planning by both me and price .



 Comments   
Comment by ncaplar [ 17/Apr/21 ]

After consultations with Paul, I have uploaded a new code (/tigress/ncaplar/PIPE2D-793) which takes as input wavelength and fiber, as well as test cases. This is now being implemented in C and pipeline by Paul.

Comment by hassan [ 27/Apr/21 ]

Re-assigned to Paul as he will take care of the pipeline updates.

Comment by price [ 10/Jun/21 ]

What is the desired behaviour when the user requests the PSF for a fiber without a PSF solution? The reference implementation returns the PSF for the closest fiber and warns "this is probably not what you want". In order to protect the user from getting strange results, I currently have the pipeline code throwing an exception.

Comment by ncaplar [ 10/Jun/21 ]

I think that is ok.

 

Comment by price [ 10/Jun/21 ]

Thanks. Then I think this is ready to be reviewed.

Comment by hassan [ 11/Jun/21 ]

Changes look fine. You mention in one of the commits that you "managed to improve the interpolation precision from 1e-8 to 1e-9" . Do you have code/plots that demonstrate that?

Comment by price [ 11/Jun/21 ]

The test of fidelity to Neven’s reference implementation has had the precision tightened from 1e-8 to 1e-9.

Comment by price [ 11/Jun/21 ]

Merged.

Generated at Sat Feb 10 15:58:09 JST 2024 using Jira 8.3.4#803005-sha1:1f96e09b3c60279a408a2ae47be3c745f571388b.