profile-448-b-339.png
profile-448-r-339.png
[PIPE2D-472] Strategy for determining traces from non-odd/even high-density fiber data Created: 02/Nov/19 Updated: 05/Jan/21 Resolved: 27/Jun/20 |
|
| Status: | Done |
| Project: | DRP 2-D Pipeline |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Story | Priority: | Normal |
| Reporter: | hassan | Assignee: | price |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Attachments: |
profile-448-b-339.png
profile-448-r-339.png
|
| Story Points: | 4 |
| Sprint: | 2DDRP-2019 I, 2DDRP-2019 J, 2DDRP-2021 A |
| Reviewers: | hassan |
| Description |
|
In the ideal case, in order to determine the traces for high-density fiber data, two quartz exposures need to be prepared, one with the even-numbered fibers illuminated, and the other with the alternate odd-numbered fibers illuminated. However, this has not been possible for recent high-density fiber data from LAM. Please develop a rough strategy of determining the traces from such data in order to reduce those data. |
| Comments |
| Comment by price [ 05/Mar/20 ] |
|
My plan is to get a rough analytical fit to the spatial profile from low-density fiber data, and use that whenever high-density fiber data restricts the spatial range. hassan recommends looking at "visits 436..441 . Under /projects/HSC/PFS/Subaru/raw/pfs/2019-12-17/ on the Princeton cluster." |
| Comment by price [ 16/Jun/20 ] |
detrend.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price --rerun price/pipe2d-472/fiberTraces --id visit=448 -c isr.doFlat=False
measureFiberProfile.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price/ --rerun price/pipe2d-472/fiberTraces --id visit=448 arm=r -c buildFiberTraces.profileRadius=7 --log measureFiberProfile.buildFiberTraces=DEBUG
constructFiberTrace.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price --rerun price/pipe2d-472/fiberTraces --id visit=455 arm=r -C constructFiberTrace-config.py --batch-type=none
ingestCalibs.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price /projects/HSC/PFS/Subaru/rerun/price/pipe2d-472/fiberTraces/FIBERTRACE/pfsFiberTrace-2019-12-17-000455-r1.fits --validity 1000
reduceExposure.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price --rerun price/pipe2d-472/fiberTraces --id visit=453..457 arm=r
plotArm.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-price --rerun price/pipe2d-472/fiberTraces --id visit=453..457 arm=r --filename 'spectra-%(visit)d-%(arm)s%(spectrograph)d.png'
All the extracted spectra from 455 are unity +/- epsilon (with epsilon ~ 2e-8), as expected. >>> butler = Butler("/projects/HSC/PFS/Subaru/rerun/price/pipe2d-472/fiberTraces") >>> ss = butler.get("pfsArm", visit=457, arm="r") >>> ss.flux[ss.mask == 0].mean() # exposure time is twice that of 455, so expect 2 2.005703 >>> ss.flux[ss.mask == 0].std() 0.08082811 >>> np.percentile(ss.flux[ss.mask == 0], (25.0, 75.0)) array([1.99344251, 2.01826459]) >>> 0.741*(2.01826459-1.99344251) 0.018393161279999893 >>> np.sqrt(ss.variance[ss.mask == 0].mean()) 0.0118158533126398 So noise is not too crazy. Outliers (std Unable to render embedded object: File (= robust std) are likely due to CRs (1200 sec exposure) not found.). constructFiberTrace-config.py is: from pfs.drp.stella.extendFiberProfile import BuildExtendedFiberTracesTask from pfs.drp.stella.buildFiberTraces import BuildFiberTracesTask config.trace.retarget(BuildExtendedFiberTracesTask) config.trace.filename = "profile-448-r1.pkl" config.trace.profileRadius = 2 config.trace.centroidRadius = 2 |
| Comment by price [ 16/Jun/20 ] |
|
Did some optimisation of the new BuildFiberTraceTask so it wouldn't take forever on dense fiber data, and added new code to be able to extract spectra on dense fiber data. |
| Comment by price [ 27/Jun/20 ] |
|
Merged to master. |