WCS same within deepcoadded tracts

Support
,
#1

Hi all,

Not sure if I am doing something wrong, but when I try to get the sky position of a pixel in my deep_coadd patches they are all the same within the same tract. I made a minimal code to show here…

target_ra = 53.2
target_dec = -28.1
radius = 0.52
region = sphgeom.Region.from_ivoa_pos(f"CIRCLE {target_ra} {target_dec} {radius}")

coadd_datasetrefs_i = butler.query_datasets("deep_coadd",
                                          where="patch.region OVERLAPS region AND band='i'",
                                          bind={"region": region},
                                          with_dimension_records=True,
                                          order_by=["patch"])

num = len(coadd_datasetrefs_i)

for inum in range(num):
    coadd = butler.get(coadd_datasetrefs_i[inum])
    patch = coadd.getMetadata()['LSST BUTLER DATAID PATCH']
    tract = coadd.getMetadata()['LSST BUTLER DATAID TRACT']
    wcs = coadd.getWcs()
    p1 = wcs.pixelToSky(0.0,0.0)   
    print(patch, tract, p1)

This produces…

1 5063 (54.0342443848, -28.3506114684)
2 5063 (54.0342443848, -28.3506114684)
3 5063 (54.0342443848, -28.3506114684)
4 5063 (54.0342443848, -28.3506114684)
5 5063 (54.0342443848, -28.3506114684)
#2

Ah I think I found the solution in a tutorial video: https://www.youtube.com/watch?v=OsjxN7PfiYk

The pixel has to be translated using the bounding box information, such as…

    i = 0.0
    j = 0.0
    bbox = coadd.getBBox()
    y = i + bbox.y.min
    x = j + bbox.x.min
    p1 = wcs.pixelToSky(x,y)  
    print(patch, tract, p1)

which now produces…

1 5063 (53.8576198474, -28.3628255733)
2 5063 (53.6682598394, -28.3637587078)
3 5063 (53.4788871444, -28.3644302488)
4 5063 (53.2895058991, -28.3648401400)

So probably working now I guess.

#3

Hi @csabiu , thanks for your question. The origin of a deep_coadd is not (1,1); it depends on the patch. When you provide pixel (0,0), it will return the bottom left of the tract. Yes, your follow up test works.