issue with g band - Githubissues

esheldon commented 7 years ago

i band worked but I got a failure in g band

producer_config={'fake_seg_radius':5}

filter="HSC-G"
tract=8766
patch="4,7"

butler = dafPersist.Butler("/datasets/hsc/repo/rerun/RC/w_2017_30/DM-11185")
producer = medsdm.producer.LSSTProducer(
    butler,
    tract,
    patch,
    filter,
    config=producer_config,
)

maker = medsdm.maker.DMMedsMaker(producer)

fname="test-medsdm-g-tract008766-patch47.fits"
maker.write(fname)

esheldon commented 7 years ago

ping @TallJimbo

TallJimbo commented 7 years ago

Fixed (on the reject-no-psf-objects branch). The problem was that we were checking if the PSF could be evaluated at the g-band centroid, and then using the reference centroid (i-band, I think, in this case) to evaluate the PSF when making the MEDS files. This object was right on the edge of where we had data in g, so even that sub-pixel difference made it impossible to evaluate the PSF. I've switched the checking to use the reference centroid, so the problematic object here is now removed from the catalog before we start making the MEDS files.

esheldon commented 7 years ago

Thanks, I'll check after the NCSA maintenance

esheldon / medsdm

issue with g band #13