Open marxide opened 2 years ago
I will take a closer look, but I suspect the issue is with the code that does the extraction (in this case of the background region around an island for its parameterisation). It probably defaults to extracting the Stokes I, and needs to be told about the polarisation of the input image.
And to answer your question - yes, setting flagNegative to true would search for negative sources (by inverting the image and running the same detection algorithm)
I am trying to run Selavy on a Stokes V ASKAP image. It fails with the following error:
The error messages implies I may need to set the Stokes parameter in the configuration, but the only place I see to this from the documentation is for extracting spectra which is turned off in my parset. Furthermore, it appears the Stokes is set correctly in the image header (
CRVAL4=4
).The input image is a 1000 x 1000 px Stokes V image cutout from the VAST survey in which I inserted a couple of sources (one positive, the other negative). The positive source model has an amplitude of 0.007 with centre (500, 500) px. Selavy finds this source as expected. From the log:
I assume the negative source would be found if re-run using
flagNegative = true
?Invocation
Using the Yandasoft docker image in serial mode.
The same error occurs when I run this in distributed mode on Magnus using the ASKAPsoft module.
Probably not relevant but I noticed the versions reported in the log don't match the docker image tag:
Should one or both of these be 1.6.0?
Image header
Parset
Full log
Test data
test-image-data.tar.gz