[PIPE2D-492] Miscellaneous changes to the 2D DRP pipeline

Relates — Mon, 4 Jan 2021 20:23:10 +0000

Following preparations towards the PFS Collaboration Meeting Dec 2019, rhl encountered a number of areas of improvement. His initial rough notes are copied here, but separate tickets based on these notes, and additional related points post-meeting, will be filed as child tickets in due course.

 get("detectormap") – need to decide on good names for the butler
 detMap.fiberIds v. pfsArm.fiberId
 pfsArm.resample docs claim to return a pfsSpectra object not a pfsArm
 afwGeom -> geom change
 What happened to
 stella.utils.addFiberTraceSetToMask
 drpStella.markFiberTraceInMask(ft, mask, ftMask)
 ? Superceded by FiberTrace.getTrace():
 -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
 from pfs.drp.stella import fiberMaskPlane
 def addFiberTraceSetToMask(mask, fiberTraceSet, display=None):
 mask.addMaskPlane(fiberMaskPlane)

for ft in fiberTraceSet:
 ftMask = ft.trace.mask
 calexp.mask[ftMask.getBBox()] |= ftMask

if display:
 display.setMaskPlaneColor(fiberMaskPlane, "GREEN")
 -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
 - It looks as if we're not updating the DetectorMaps' fibre trace *columns* in calibrateWavelength. An old RHL problem, but it may be why the vector plots are weird (or why you built them off arcs!)
 - It looks as if you cannot run reduceArc on calexps. Is there a way to avoid rerunning ISR? The problem is actually in reduceExposure
 pfsSpectra.py and pfsSpectrum.py are confusing names
 The datamodel probably needs to define the values of the targetType enum. They also need to go in the header
 Now we're writing pfsSky the datamodel needs to be fleshed out. Looking at the code I don't understand the model; I'd expect it to be per-fiberId not position
 pfsConfig.objId[pfsConfig.targetType == TargetType.SKY] returns an array of 0s; we need real IDs for skies
 pfsArm files don't seem to fill out the sky HDU

PFS-JIRA

[PIPE2D-492] Miscellaneous changes to the 2D DRP pipeline