vir smear correction saturation indexing

Line 586 in 24193ba

vis_tmp[i_sat] = 0

Could you please check if this is definitely working as intended? I think potentially when i_sat is used to index vis_tmp it is actually zeroing out the wrong bands (since i_sat is found on Line 585 using vis_tmp[85:] not vis_tmp). Also, why 85 when comment on Line 592 says 90:95? Great package - thanks.

Thanks for pointing that!
Indeed it was zeroing the wrong wavelengths, it has been fixed with commit 4e2720d

omegapy/omegapy/omega_data.py

Lines 582 to 590 in dbb973c

    
           # Check for saturation in the 85:95 channels 
        
           vis_tmp = np.swapaxes(vis, 0, 1)    # Inversions axes pour calcul (lam, npix, nscan) 
        
           means = np.mean(vis_tmp[0:85], axis=0) 
        
           i_sat = np.where(vis_tmp[85:] > means) 
        
           i_sat1, i_sat2, i_sat3 = i_sat 
        
           i_sat1 += 85 
        
           i_sat = (i_sat1, i_sat2, i_sat3) 
        
           vis_tmp[i_sat] = 0 
        
           vis = np.swapaxes(vis_tmp, 0, 1)        # On remet dans l'ordre

And regarding the 90 vs 85 in the comment, it is a typo from the IDL file that has not been fixed when porting to Python, now corrected.

Thanks Aurélien. I have made some efficiency improvements in my local version of your code (e.g., getting rid of all loops in _readomega) - I will share these with you soon but maybe it's no longer a faithful translation of the IDL code.

	# Check for saturation in the 85:95 channels
	vis_tmp = np.swapaxes(vis, 0, 1) # Inversions axes pour calcul (lam, npix, nscan)
	means = np.mean(vis_tmp[0:85], axis=0)
	i_sat = np.where(vis_tmp[85:] > means)
	i_sat1, i_sat2, i_sat3 = i_sat
	i_sat1 += 85
	i_sat = (i_sat1, i_sat2, i_sat3)
	vis_tmp[i_sat] = 0
	vis = np.swapaxes(vis_tmp, 0, 1) # On remet dans l'ordre