DM-36768: Add psf image to alert cutouts #221
Conversation
| @@ -389,11 +389,12 @@ def createCcdDataCutout(self, image, skyCenter, extent, photoCalib, srcId): | |||
| CCDData object storing the calibrate information from the input | |||
| difference or template image. | |||
| """ | |||
| point = image.getWcs().skyToPixel(skyCenter) | |||
There was a problem hiding this comment.
Wouldn't it be better to pass in the pixel centroid to this method from the diaSource record? No need to do WCS math at all, since we know where the source centroid is.
There was a problem hiding this comment.
Can we trust that the (x,y) of the DIASource are the same on the calexp and matched template?
There was a problem hiding this comment.
Yes, those are made to be 1-1 pixel aligned (with the template having a buffer, but the xy0 set such that they are aligned).
tests/test_packageAlerts.py
Outdated
| @@ -265,6 +265,9 @@ def testCreateCcdDataCutout(self): | |||
| self.cutoutWcs.wcs.cd) | |||
| self.assertFloatsAlmostEqual(ccdData.data, | |||
| calibExposure.getImage().array) | |||
| self.assertFloatsAlmostEqual(ccdData.psf, | |||
| self.exposure.psf.computeKernelImage(self.center).array) | |||
| self.assertFloatsAlmostEqual(np.sum(ccdData.psf), 1.0) | |||
There was a problem hiding this comment.
Not sure this last one is necessary: you don't need to check that computeKernelImage is normalized here, since it's already well tested within afw.
| @@ -410,6 +411,8 @@ def createCcdDataCutout(self, image, skyCenter, extent, photoCalib, srcId): | |||
| % srcId) | |||
| return None | |||
|
|
|||
| cutoutPsf = image.psf.computeKernelImage(point).array | |||
There was a problem hiding this comment.
Maybe a comment here as to why this is computeKernelImage instead of computeImage? I'm still not convinced as to which is correct for this use case.
There was a problem hiding this comment.
I'm convinced by previous discussion that we want the PSF centered in the array.
No description provided.