Array shape mismatch between different exposures

[themiyan]

When extracting spectra to perform background subtraction at the slit level NirspecPipeline.extract_spectrum for some sources the dimensions between slits seems to be off by 1 which causes extractions to fail.

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Input In [17], in <cell line: 5>()
      1 ### check in nrs2
      3 key_nrs2 = '2565_32130' 
----> 5 _data = gratings['clear-prism'][1].extract_spectrum(key_nrs2, skip=[], yoffset=0, prof_sigma=0.6, trace_sign=-1, 
      6                               fit_profile_params=None)

File ~/opt/anaconda3/envs/grizli39/lib/python3.9/site-packages/msaexp/pipeline.py:1356, in NirspecPipeline.extract_spectrum(self, key, slit_key, prof_sigma, fit_profile_params, pstep, show_sn, flux_unit, vmax, yoffset, skip, bad_dq_bits, clip_sigma, ntrim, get_slit_data, verbose, center2d, trace_sign, min_dyoffset, **kwargs)
   1354 _dq = (_sbg.dq & bad_dq_bits) == 0
   1355 try:                    
-> 1356     _bkg[_dq] += _sbg.data[_dq]
   1357     _bkgn[_dq] += 1
   1359 except ValueError:
   1360     #print('background failed', j)

IndexError: boolean index did not match indexed array along dimension 1; dimension is 142 but corresponding boolean dimension is 143

ipdb>  np.shape(_dq)
(30, 143)
ipdb>  np.shape(_sbg.data)
(30, 143)
ipdb>  np.shape(_bkg)
(30, 142)

This is coming because

ipdb>  pipe[0].slits[0].data.shape
(35, 425)
ipdb>  pipe[1].slits[1].data.shape
(35, 425)
ipdb>  pipe[2].slits[2].data.shape
(35, 426)

using
grizli version = 1.8.4.dev4+ga5d2ece
msaexp version = 0.6.8.dev8+gf5c84b2.d20230403
jwst version = 1.10.0