Problems with keyword modifying

[nikorose87]

Hi Codie

I was following your, tutorial established in the 15th international conference of LS-DYNA, and I want to modify the material on the keyword in order to run many simulations, so let's say that we have:

>>> kf = KeyFile("sample_snapping.k",  read_keywords=True,    parse_mesh=True,            load_includes=True)

>>> kp0 = kf['*PART'][0]
>>>print(kp0)

*PART
$#                                                                         title
CTSD                                                                  
$#     pid     secid       mid     eosid      hgid      grav    adpopt      tmid
2                  1         3         0         0         0         0         0

>>> kp0['mid'] = 2

Here the error come up:

ValueError: Can not find field: mid in comments.

If I set the command kp0.get_lines(), only the first part is shown:

['*PART',
'$# title']

Am I doing something wrong?

Thanks

[codie3611]

Hey buddy, I tried to reproduce using the fragment above (put it into yay.key) but failed ... Do you have an isolated example or file you can share with me?

Python 3.6.4 |Anaconda, Inc.| (default, Jan 16 2018, 18:10:19) [GCC 7.2.0] on linux Type "help", "copyright", "credits" or "license" for more information. 
>>> from qd.cae.dyna import * 
>>> kf = KeyFile('yay.key', read_keywords=True, parse_mesh=True, load_includes=True) 
>>> kp0 = kf['*PART'][0] 
>>> print(kp0)  
*PART 
$#                                                                         title 
CTSD 
$#     pid     secid       mid     eosid      hgid      grav    adpopt      tmid 
2                  1         3         0         0         0         0         0   
>>> kp0['mid'] = 2 
>>> print(kp0)  
*PART 
$#                                                                         title 
CTSD 
$#     pid     secid       mid     eosid      hgid      grav    adpopt      tmid
2                  12                  0         0         0         0         0   
>>> kp0.get_lines() 
['', '*PART', 
'$#                                                                         title', 
'CTSD                                                                  ', 
'$#     pid     secid       mid     eosid      hgid      grav    adpopt      tmid', 
'2                  12                  0         0         0         0         0', 
'']

[nikorose87]

There you go!
codie.zip

[codie3611]

Thanks, will try it.

[nikorose87]

However, after aligning the part wanted to modify, I could modify the text value.

[codie3611]

However, after aligning the part wanted to modify, I could modify the text value.

Such bugs never let me sleep well ... I can reproduce now thanks. If I can't find the issue right now it might take a few days since I'm at the german dyna forum next week.

[codie3611]

Analysis

Ok the issue is quite simple: since the mesh is parsed, the part is not stored as a string but as a real part in the mesh database. The code therefore uses the pid field and stores the rest including mid in another buffer, which is not searched/accessed when the bracket operator [] is called from python. This also explains why the part is still correct when written to the file, but you don't see the rest of the data in kw.get_lines(), since it is stored in another buffer.

Circumvention

As a circumvention you can disable mesh parsing by parse_mesh=False in the constructor. It saves time anyways if you really don't modify the mesh.

Fix

I need to think about how to fix it internally with given datastructure, I don't intend though to leave it as it is.

[nikorose87]

Thanks so much Codie!

Effectively, I set the parse_mesh=False and It worked. I will modify the data in mesh, but I will do it in separate files and later merge them.

Thanks a lot, have a nice forum.

[codie3611]

Sry for the inconvience ... fixing it may seem simple, but the part keyword has some internal conditionals which are annoying to handle ...

[codie3611]

Will be officially fixed in version 0.8.4. If you require a fixed version, simply clone the dev branch. Tests run through so it seems fine.