Closed gostevehoward closed 7 years ago
The new script is documented at the bottom of this page
@kbarbary I'm seeing this error from WCS.jl on travis, but that curl command works fine for me on my local machine. Are you getting this error too? It just started recently.
INFO: Attempting to Create directory /home/travis/.julia/v0.7/WCS/deps/downloads
INFO: Downloading file https://cache.julialang.org/ftp://ftp.atnf.csiro.au/pub/software/wcslib/wcslib-5.13.tar.bz2
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 455 100 455 0 0 4739 0 --:--:-- --:--:-- --:--:-- 35000
curl: (35) error:1407742E:SSL routines:SSL23_GET_SERVER_HELLO:tlsv1 alert protocol version
=================================[ ERROR: WCS ]=================================
LoadError: failed process: Process(`curl -f -o /home/travis/.julia/v0.7/WCS/deps/downloads/wcslib-5.13.tar.bz2 -L https://cache.julialang.org/ftp://ftp.atnf.csiro.au/pub/software/wcslib/wcslib-5.13.tar.bz2`, ProcessExited(35)) [35]
while loading /home/travis/.julia/v0.7/WCS/deps/build.jl, in expression starting on line 39
Looks great!
This started as an investigation of #655, but I ended up finally understanding how image calibration and gain works, which was documented here, and I've updated the code in various places to more clearly reflect what I learned. I've updated the parameters used in synthetic field generation to more closely match SDSS. (It may eventually make sense for synthetic fields generation to automatically match the image parameters for the corresponding SDSS field, but that would requires various changes to the pipeline and doesn't seem worthwhile right now.) I also added a script,
extract_sdss_image.jl
, which writes a FITS file with raw SDSS data in "sky-added" nelec units, which is the form used internally by Celeste, and the form written for synthetic fields. This should aid debugging in the future and helps compare synthetic imagery to real imagery. As an example, here's a side-by-side of real (left) and Synthetic.jl (right) images for RCF 3325/5/110, "r" band:Here I've ensured (I think?) that the color scalings are identical in both images. You can see that now they look very similar. (Note this is not the same synthetic imagery as before, because I didn't have quite the correct background/calibration/gain parameters before.)