bm32esrf / lauenn Goto Github PK

I believe that exctinction conditions not predefined in CrystalParameters.py are not dealt with correctly. Note that I added my own material with space group 191 (its PrNi5 for full reference) and crashed at generating trainig data via GUI.

Since group 191 has not predefined data, in CrystalParameters.py -> ApplyExtinctionrules (line 62 onwards) it satisfies Extinc.isdecimal() and so calls wyckpos_lauetools.py -> testhklcond_generalrules_array (line 428 onwards).

1st issue is that there gp[2] (defined by sorting appropriate wp entry at around 448) is None. But in the condition at 452 it is not 'n/a', so the cond gives true. But None does not seem to be valid argument for hklcond_group.findall (reported as #2).

Assuming there should have been and instead of or the return True statement seems weird, as the expected output of the function is smth like 2dim array. I tried to bypass this by outputting np.array(((True,),)), but it than crashes at utils_lauenn.py -> get_material_data -> dstar_square_2 = np.diag(np.inner(np.inner(H2, metrics), H2)) (line 912), since H2 is only 1x1 array (which I believe is connected to the fact that ApplyExtinctionrules got only one True instead of expected hkl triplet as described in this paragraph).

Hope I did not misinterpreted the code completely, I am mostly guessing. But all in all a material with space group 191 is not processed correctly and I tried to suggest the reason why it might be so.

In lauetoolsnn/lauetools/wyckpos_lauetools.py on lines 451 to 455 there is

# load reflection conditions if needed
if gp[2] is not None or gp[2] != 'n/a':
    hklcond = hklcond_group.findall(gp[2])
else:
    return True

IIUIC if gp[2]==None then it is not 'n/a' thus the whole condition is true and the code crashes as None seems to be not valid input for hklcond_group.findall

Similar case is at lines 96 to 100.

Note: I found that when trying to implement my own material into lauetoolsnn.