#6 Constraints only written to file after reloading state file in Paraview

[Alex]

Dear VoxFE developers,
When setting up the constraints in paraview/VoxFE on the good old block example, it seems that the chosen constraints and their values are only written to the constraints.txt after saving in paraview as a state file and a reloading of the state. Attached are examples for this behaviour: In the constraints_BEFORE.... file, the chosen nodes are correctly written to file but the chosen constraints in xyz (lines 1-15) and the force in y (lines 16-19) are not written correctly after the node numbers.
When I save the setup as a state file in paraview/VoxFE and then reload this state file, the constraints and forces are correctly written to file -> see the constraints_AFTER.... file.

I am using Paraview 4.1.0 64bit on Windows but this behaviour is the same on a linux machine and for Paraview 5.1.0 on both operating systems.
Any idea how to work around this? With a few hundred insect models on the horizon, I would like to skip this step of closing and reloading if possible (especially with larger models).

Thank you in advance,
Alex Blanke

[swordfishtrombone]

Hi Alex,
I haven't been able to reproduce the problem, but there may well be an issue.... can I just clarify where you're saving the constraints file? (Saving a backup constraints file might always be a good idea.)

There's always been an issue between the state file and the constraints output because the constraints are not sufficiently well defined. In short, it's a bug in the original file format - the constraints should have a group number so that eg. forces can be correctly distributed over them. Hence the state file can't really represent the constraints correctly unless we fix the format.

That means having a new version which can cope with a group number or not and so it means a lot more work than it sounds (though it would probably be worth it eventually). And the solver would have to know how to cope too. An interim solution might be to write a separate group constraints file, which could work I think, though it could also easily cause confusion in other ways. (In this way, we could have more control over the process, so

Anyhow, the constraints file should only be written and trusted when you go through the select points/extract/write out script (target) operations. If loading the state file is re-running those steps (as it probably is), then that could be a source of error in the 'new' constraints file. (In effect, because of the format issue, loading from the state file only gives a visual representation of the constraints, or at least that's how I think of it. I hadn't noticed this was a causing an over-writing problem though.)

Hope that helps,
Richard

[Alex]

Hi Richard,
apologies for taking so long to respond. Teaching obligations and all that....

The constraints file is saved in the same folder lke the .model and .script files.

I think the issue here is not how the syntax within the constraints file looks like: After I save, close and reopen the model in Paraview the constraints.txt is written with the correct forces.

The issue is that I cannot write a correct constraints file unless I close and reopen Paraview (via the saved state file). When I set up the block model with some forces and try to directly export a constaints file without closing and reopening Paraview, the node numbers are written into the constraints.txt but the forces are all -1.#IND.

So what might cause the different behaviour of Paraview after reloading the model?

We can also have a short skype session so that I can sho you whats going on on my side.

Thanks for any help!

Alex