seq.test_report() calculates wrong "Spatial resolution" if spoilers are used.

[h3lg3]

Describe the bug Output of seq.test_report() for spatial resolution is wrong if kspace is not centered. In ext_test_report.py, line 72, kspace extent is calculated

k_extent = np.max(np.abs(k_traj_adc), axis=1)

However, if the k-space is not centered around 0, this calculation leads to wrong estimation of the resolution. A fix would be: k_extent = 0.5 * (np.max(k_traj_adc, axis=1) - np.min(k_traj_adc, axis=1))

This has the disadvantage, that for a centered kspace ranging, e.g., from -32 to +31 with a total of 64 samples. The above calculation slightly misestimates the resolution since k_extent is calculated to be 31.5 instead of 32 (current implementation)

Screenshots Here you can see that the maximum kspace extent (of ADC) underestimates the resolution, if it is not accounted for the lower kspace extent (of ADC)

Desktop (please complete the following information):

• OS: Windows 10
  • pypulseq version: dev-branch 2f09cf6

[h3lg3]

I figured out, that the kspace is shifted only after I imported the sequence again. Plotting the kspace trajectory directly after building did not show the shift. Setting the optional parameter "detect_rf_use = True" for seq.red() resolved the issue.

# write sequence to disk
seq.write('test.seq')

# import sequence
seq_import = pp.Sequence(system=system)

seq_import.read('test.seq') # kspace is shifted, because 180° refocussing pulses are not recognized

seq_import.read('test.seq', detect_rf_use = True) # solves problem of shifted kspace