fix: support ELEMENT_SHELL_THICKNESS_OFFSET with conditional table cards

[RaphaelHeiniger]

• Add manifest entry for ELEMENT_SHELL_THICKNESS_OFFSET using table-card-group with a conditional card for the optional N5-N8 midside node thickness values.

• required a fix in table_card_group handler _dvidide_data_lines to correctly handle conditional cards.

[RaphaelHeiniger]

@koubaa the fix in ELEMENT_SHELL_THICKNESS_OFFSET required usage of table_card_group in combination with conditional cards. It required changes in table_card_group to make the conditional card work. Please review this and let me know if there is another way to get to the same result without changing anything in table_card_group or if there is a cleaner solution to this.

[koubaa]

@RaphaelHeiniger please also include the manifest change

[koubaa]

@koubaa the fix in ELEMENT_SHELL_THICKNESS_OFFSET required usage of table_card_group in combination with conditional cards. It required changes in table_card_group to make the conditional card work. Please review this and let me know if there is another way to get to the same result without changing anything in table_card_group or if there is a cleaner solution to this.

CardSet already handles this but it's less user-friendly than TableCardGroup. I'm not confident in the implementation proposed here yet, but I'll think about this some more. I'm concerned about the parameter handling (reading a deck with parameters in this keyword and writing it with and without substitution) since I believe the sub-card index when reading is encoded into the address of the parameter.

Moreover, the condition as understood in the manual seems to be applied per set, not for the entire n5 column. That makes the ['n5'].any() condition seem incorrect.

[RaphaelHeiniger]

@koubaa I pushed the manifest.json. I did not anticipate that a user mixes this format. Not sure if LS-DYNA would accept it but could be. So would you prefer to do it with CardSet?

[koubaa]

I think it is worth testing to see if ls-dyna accepts it. CardSet is preferable to the current implementation due to the doubts about correctness but I think table-card-set is generally preferable if it can be made to work because the api feels cleaner

[RaphaelHeiniger]

I can confirm that a mixed input like the one attached does work.
element_shell_thickness_offset.txt

[koubaa]

In that case I would prefer a card set since we know it will be correct. We can revisit it later once we have some more time to enhance the table card