Performing FFTs of collected data no longer working

[jtenbarg]

To see this problem, select a single point in a dataset and collect over several frames to produce, e.g., Bz(x=0,t).

pgkyl 'pkpm_alf_wave_p1-field_[0-9]*.gkyl' interp -b ms -p 1 sel --comp 5 --z0 0. collect fft -p pl

File "/home/tenbarge/postgkyl/postgkyl/commands/fft.py", line 36, in fft
diag.fft(dat, kwargs['psd'], kwargs['iso'], overwrite=True)
File "/home/tenbarge/postgkyl/postgkyl/diagnostics/fft.py", line 22, in fft
grid = np.delete(grid, idx)
File "<array_function internals>", line 200, in delete
File "/home/tenbarge/miniconda3/lib/python3.8/site-packages/numpy/lib/function_base.py", line 5126, in delete
arr = asarray(arr)
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (2,) + inhomogeneous part.

[jtenbarg]

This problem arises because numpy no longer accepts ragged arrays. Older versions work, they just give the below warning.

/home/tenbarge/miniconda3/lib/python3.8/site-packages/numpy/lib/function_base.py:5030: 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 must specify 'dtype=object' when creating the ndarray.
arr = asarray(arr)