Draft: Update interpolate.py

[adrabent]

Current runs with recent numpy cause the following error:

/usr/local/lib/python3.8/dist-packages/losoto/operations/interpolate.py:149: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you m 
ust specify 'dtype=object' when creating the ndarray. 
  new_vals[selection] = np.interp(new_axisvals, orig_axisvals[unflagged], 
Traceback (most recent call last): 
  File "/usr/local/bin/losoto", line 154, in <module> 
    returncode += ops[ op ]._run_parser( soltab, parser, step ) 
  File "/usr/local/lib/python3.8/dist-packages/losoto/operations/interpolate.py", line 18, in _run_parser 
    return run( soltab, outSoltab, axisToRegrid, newDelta, delta, maxFlaggedWidth, log) 
  File "/usr/local/lib/python3.8/dist-packages/losoto/operations/interpolate.py", line 149, in run 
    new_vals[selection] = np.interp(new_axisvals, orig_axisvals[unflagged], 
IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

This branch aims for fixing it.

[adrabent]

Please DO NOT merge the branch yet. Changes need testing.

[tikk3r]

Is there any update/progress on this? I'm running into the same issue now as well since we made our containers up-to-date.

I missed this pull request, so I made my own fix by casting selection to a tuple before slicing with it: tikk3r@d9c41de This seems to work for me, but then I noticed this open pull request. I can't find it clearly in the official NumPy release notes or documentation, but according to this post it does have to do something with tuples vs lists as indices.

It would be great if one of the fixes could make it into master soon.