[INSTRM-596] Arrange to create a 2d PFSA file for each 3d PFSB file Created: 12/Jan/19  Updated: 21/Apr/23  Resolved: 18/Nov/21

Status: Won't Fix
Project: Instrument control development
Component/s: ics_hxActor
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Normal
Reporter: cloomis Assignee: cloomis
Resolution: Won't Fix Votes: 0
Labels: NIR, SPS
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Relates
relates to INSTRM-1061 Apply IRP subtraction to PFSA files. Won't Fix
relates to DAMD-7 Define format for up-the-ramp files Done
relates to INSTRM-1928 Create intermediate quicklook NIR files In Progress
Story Points: 1
Sprint: 2019 C, SM1-2019 E, SM1-2019 F, SM1-2019 G, SM1PD-2020 K, SM1PD-2020 M, SM1PD-2020 L, SM1PD-2020 N, SM1PD-2021 A 12, SM1PD-2021 A, SM1PD-2021 A 2, SM1PD-2021 A 3

 Description   

The hxActor currently writes PFSB files only: the full ramp. We have planned to also write PFSA files: single image versions reduced from the full stack.

Note that we never settled on exactly what to write. A reality is that the DAQ BEEs most certainly do not have enough CPU or RAM to compute more than lastRead-firstRead (or the only read), so that's what we start with.

One concern is that "A" files from Subaru could be seen as the "correct" raw instrument files, just from past convention. I'll state that any proper 3d-to-2d reduction is almost certain to evolve, so there might never be a dependably stable science-grade 2d product: it will always depend on a particular version of the 2D DRP pipeline. [ Besides, any good reduction will come with (at least) both variance and mask images. ]

I suppose we could call any such quick raw 2d reductions something other than "PFSA" files, and simply not write "PFSA" files for the NIR arms?

rhl, Hisanori Furusawa, eric



 Comments   
Comment by hassan [ 21/Mar/19 ]

Finish that this week.

Comment by hassan [ 18/Apr/19 ]

Simple case done; interleaved reference pixel situation needs to be implemented.cloomis will break this into separate tickets.

Comment by cloomis [ 10/Dec/20 ]

The main question still holds. PFSB files holds all the raw data, and we cannot generate anything other than a quick and preliminary CDS files for the entire ramp.

I think we will need a quick CDS image, but there are strong arguments for not creating official PFSA files from them. We currently do, but they are not documented in datamodel.txt. Should I just call them something else? What?

Comment by cloomis [ 04/Feb/21 ]

I think this must be an NAOJ question, and the original description still holds. Hisanori Furusawa? eric?

 

We do currently generate a preliminary (lastRead-firstRead) FITS file, currently named PFSA. These are very practical to have available, but are not science-quality. The instrument cannot improve on the content. 

Comment by eric [ 05/Feb/21 ]

For traditional Subaru instruments, "A" files are raw data, while "Q" files are quick reduced data.  Both kinds are distributed using the data traversal mechanism of Gen2.  For some instruments, the Q files are archived to STARS, and for some instruments they are not–this is generally dependent on whether the developers put in any work to make the FITS headers compliant enough to make the data archive people happy.

I think it would be useful to create such quick reduction files, available to the observers at least during the night. These could be displayed, or have special GUIs to show tables out of them, create graphs or what not.

Comment by cloomis [ 05/Feb/21 ]

OK, so we could name them PFSQ files.

In that case, should the (currently off-site only) PFSB ramp files be renamed to PFSA? They are wildly different from normal CCD images and in general pretty inconvenient (~3GB, with ~100 HDU, and no HDU looking like a normal image). The nature of PFSA* would surely surprise people.

Since we have not yet taken NIR data at Subaru I think we are still free to rename.

Comment by cloomis [ 18/Nov/21 ]

We long ago decided not to create single image PFSA files for the H4 ramps: if written by the ICS they can only be poor approximations, but by virtue of being named "PFSA" could easily be viewed as real data.

Generated at Sat Feb 10 16:26:40 JST 2024 using Jira 8.3.4#803005-sha1:1f96e09b3c60279a408a2ae47be3c745f571388b.