astrofrog
commented
2 years ago Here is an issue I encountered but not clear if this is an astropy or fitsio issue:
import os
import fitsio
import numpy as np
os.remove('test.fits')
data = np.random.random((15, 15))
fts1 = fitsio.FITS('test.fits', 'rw')
fts1.write(data, compress='HCOMPRESS_1', qlevel=20, qmethod=2, tile_dims=(5, 5))
from astropy.io import fits
print(fits.open('test.fits', disable_image_compression=True)[1].data)gives:
[([221, 153, 0, 0, 0, 5, 0, 0, 0, 5, 0, 0, 0, 0, 255, 255, 255, 248, 0, 0, 3, 224, 8, 8, 6, 246, 191, 173, 31, 166, 1, 43, 32, 18, 18, 60, 223, 253, 255, 123, 1, 8, 23, 128, 152, 30, 128, 72, 10, 159, 247, 252, 24, 94, 171, 220, 131, 144, 37, 2, 151, 253, 245, 243, 254, 251, 241, 250, 0, 64, 98, 22], 0.00956831, 20547788.20710866)
etc.The last value is ZZERO and is huge. Changing qmethod to 1 (which is identical except for its treatment of zeros which there are none of here) gives:
[([221, 153, 0, 0, 0, 5, 0, 0, 0, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 80, 8, 7, 6, 250, 81, 255, 127, 192, 161, 145, 225, 254, 155, 240, 255, 224, 20, 8, 128, 120, 12, 192, 208, 6, 207, 248, 16, 191, 224, 34, 5, 128, 88, 9, 15, 251, 235, 235, 199, 251, 215, 192, 0, 32, 1, 14], 0.01499177, 0.01499177)
In the test suite we currently skip some tests due to issues in fitsio - we should report these upstream if possible.