Reading input file 1 of 1: "3C_279V_000206853.FIT"...
Base: "./3C_279V_000206853", basefile "3C_279V_000206853.FIT", basedir ".", suffix "FIT"
Checking if file "3C_279V_000206853.FIT" ext 0 is xylist or image: image
(not xyls because: FITS file does not have any extensions)
Running: /usr/local/bin/image2pnm.py --sanitized-fits-outfile /tmp/tmp.sanitized.7eE4cV --fix-sdss --infile 3C_279V_000206853.FIT --uncompressed-outfile /tmp/tmp.uncompressed.lMRbIR --outfile /tmp/tmp.ppm.6uZyyt --ppm
sh: sysctl: command not found
Header has 47 cards
Traceback (most recent call last):
File "/usr/local/bin/image2pnm.py", line 323, in <module>
sys.exit(main())
File "/usr/local/bin/image2pnm.py", line 320, in main
mydir, fix_sdss=options.fix_sdss)
File "/usr/local/bin/image2pnm.py", line 248, in convert_image
(imgtype, errstr) = image2pnm(infile, outfile, sanitized, force_ppm, no_fits2fits, extension, mydir, fix_sdss)
File "/usr/local/bin/image2pnm.py", line 184, in image2pnm
errstr = fits2fits(infile, sanitized, fix_idr=fix_sdss)
File "/usr/local/Cellar/astrometry-net/HEAD/lib/python/astrometry/util/fits2fits.py", line 86, in fits2fits
hdu.scale('int16', '', bscale, bzero)
File "/usr/local/lib/python2.7/site-packages/pyfits/hdu/image.py", line 383, in scale
bzero=bzero, blank=None)
File "/usr/local/lib/python2.7/site-packages/pyfits/hdu/image.py", line 441, in _scale_internal
self.data += -_zero
TypeError: Cannot cast ufunc add output from dtype('float64') to dtype('uint16') with casting rule 'same_kind'
augment-xylist.c:588:backtick Failed to run command: /usr/local/bin/image2pnm.py --sanitized-fits-outfile /tmp/tmp.sanitized.7eE4cV --fix-sdss --infile 3C_279V_000206853.FIT --uncompressed-outfile /tmp/tmp.uncompressed.lMRbIR --outfile /tmp/tmp.ppm.6uZyyt --ppm
ioutils.c:567:run_command_get_outputs Command failed: return value 1
Any ideas as to what could be causing this?
Default casting for inplace operations has changed to 'same_kind'. For instance, if n is an array of integers, and f is an array of floats, then n += f will result in a TypeError, whereas in previous Numpy versions the floats would be silently cast to ints. In the unlikely case that the example code is not an actual bug, it can be updated in a backward compatible way by rewriting it as np.add(n, f, out=n, casting='unsafe'). The old 'unsafe'default has been deprecated since Numpy 1.7."