Details
-
Type:
Story
-
Status: In Review
-
Resolution: Unresolved
-
Fix Version/s: None
-
Component/s: meas_deblender
-
Labels:
-
Epic Link:
-
Team:Data Release Production
Description
When running tests/testTicket-2871.py, the unit test will pass, but will throw the following exception. The origin for this exception should be caught and handled properly, and the test should be investigated to see if it is properly testing behavior and arriving at correct conclusions.
sourceDeblend WARN: Unable to deblend source 2: Footprint bounding-box Box2I((-3, -3), (3, 3)) extends outside image bounding-box Box2I((0, 0), (127, 127)) |
Traceback (most recent call last):
|
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/deblend.py", line 320, in deblend |
maxTempDotProd=self.config.maxTempDotProd |
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/baseline.py", line 671, in deblend |
debResult = newDeblend(debPlugins, footprint, maskedImage, psf, psffwhm, filters, log, verbose, avgNoise) |
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/baseline.py", line 739, in newDeblend |
maxNumberOfPeaks=maxNumberOfPeaks, avgNoise=avgNoise) |
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/baseline.py", line 133, in __init__ |
psffwhms[n], avgNoise[n], maxNumberOfPeaks, self) |
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/baseline.py", line 198, in __init__ |
self.updateFootprintBbox() |
File "/home/nate2/repos_lsst/meas_deblender/python/lsst/meas/deblender/baseline.py", line 226, in updateFootprintBbox |
(str(self.bb), str(self.imbb))) |
ValueError: Footprint bounding-box Box2I((-3, -3), (3, 3)) extends outside image bounding-box Box2I((0, 0), (127, 127)) |
sourceDeblend INFO: Deblended: of 2 sources, 2 were deblended, creating 2 children, total 4 sources |
...
|
---------------------------------------------------------------------- |
Ran 3 tests in 0.145s |
|
|
OK
|
Attachments
Issue Links
- links to
Activity
Hey Sophie Reed — this has been “in review” for the better part of a year now, and it seems unlikely that it's ever going to be a priority for Bob Armstrong since he is no longer working for DM.
Can you do something to bring it to a conclusion please? Finding another reviewer might be a good option, but it seems like a bit more work is actually needed to do what the ticket asked for (“the test should be investigated to see if it is properly testing behavior and arriving at correct conclusions”).
I sort of remember this ticket. I think we concluded that everything was working as expected, but the trace was being printed whether or not it was correctly handled. I do remember looking through everything. It is likely that I was supposed to merge this, but forgot.
Princeton pair coding attempt 1:
We changed it from printing the trace to logging the trace in case anyone needed the information. The previous behaviour can be recovered by setting log.setLevel(log.TRACE).