When comparing IDL and Python results using the mgunit framework in IDL, this bit of code is failing if the IDL result has too many dimensions:
for var_idx=n_elements(vars)-1, 0, -1 do begin
data = (var_table[vars[var_idx]])['data']
for n=points_to_check-1, 0, -1 do begin
idx = (var_table[vars[var_idx]])['ntimes']*n/points_to_check
pyresult = spd_run_py_str_to_arr(pyoutput[-1-count])
idlresult = data.Y[idx, *]
matches = spd_run_py_compare(idlresult, pyresult, tolerance=tolerance)
if matches eq 0 then begin
dprint, dlevel=0, 'Error, python validation test failed for: ' + vars[var_idx] + ' (' + strcompress(string(idx), /rem) + ')'
passed = 0b
endif
count += 1
endfor
The failure occurs at the line:
idlresult = data.Y[idx,*]
We may also need some adjustments in the lines of code that get added to the Python script, to address similar issues on the Python side. Options might be to reform the arrays being tested to ensure a fixed number of dimensions, or adding logic to do the correct indexing based on the number of dimensions in the test output.
When comparing IDL and Python results using the mgunit framework in IDL, this bit of code is failing if the IDL result has too many dimensions:
The failure occurs at the line:
idlresult = data.Y[idx,*]We may also need some adjustments in the lines of code that get added to the Python script, to address similar issues on the Python side. Options might be to reform the arrays being tested to ensure a fixed number of dimensions, or adding logic to do the correct indexing based on the number of dimensions in the test output.