Auxtel fit pt

[sylvielsstfr]

I have corrected two problems

• problem 1 in tools.py : rebin : there was no energy conservation, explaining why the final spectrum has not the same flux when CCD_REBIN =1 from CCD_REBIN != 1

return arr.reshape(shape).sum(-1).sum(1)
instead of
return arr.reshape(shape).mean(-1).mean(1)

I have checked the final spectrum for CCD_REBIN = 2 was at the same flux level as for CCD_REBIN =1

• problem 2
  After a call to find_target_init in image.py I often got a crash. Most of the cases, it happens with CCD_REBIN = 2, but I think it has nothing to do with this.

When making the sub_image_subtracted little image
sub_image_subtracted = sub_image - bkgd_2D(X, Y)

we must suppress negative values like this:

SDC : very important clipping negative signal, avoiding crash later sub_image_subtracted = np.where(sub_image_subtracted<0,0,sub_image_subtracted)

Because later these values are considered as weights (to find the target center).

Since that correction, I haven't anymore crashes.

Other things added, : few control plots in debug mode and make nicer plots by changing some sizes.

[jeremyneveu]

Major improvements for Spectractor new version 2.3

• decouple fits of diffraction order 1 and diffraction order 2
• remove dependency of pysynphot and use getCalspec
• gather the Spectractor outputs into one single FITS file
• fix lambda_ref to 550nm
• debug and adapt the software for AuxTel
• new order 2/1 ratio and throughput for AuxTel
• repair nosetests and doctests and coverage tests
• repair documentation