IFC test files and systematic testing

[bernd]

me and hlg are highly interested.

Since I use ifc import in my daily work any regression ends up in work.

In the regard of wb, IMHO it depends. If you would like to test single methods of the importer a unit test would be much more helpful. If you would like to go for a whole file import and do not know much about all the FreeCAD, it might be better to start an own wb.

Years ago I wanted to have some coded examples in FEM. At that time I did not maintain the FEM. I did not know much about FEM. Thus I made my own femexamples workbench. It might not have been the smartest way but at that time it worked for me. Furthermore ATM we do move some of the work into FEM.

[bernd]

IMHO most important is to use fixed import parameter. This has been implemented a few weeks ago. They can be parsed in latest importIFC methods.

[bernd]

If we would have something up and running I would for sure provide examples to make import more robust.

[lambda]

IMHO most important is to use fixed import parameter. This has been implemented a few weeks ago. They can be parsed in latest importIFC methods.

Yes, my code uses that. Actually the main point of the WB at the moment is: You can create reference files, which include the importer settings used to create them, so you can test the importer with the same settings later.

It's a bit hacky though: The stored settings aren't shown in the GUI anywhere, so you can only inspect them at the python console.

If we would have something up and running I would for sure provide examples to make import more robust.

I'd be happy, if we moved the repo somewhere where you and others can contribute easily.

[bernd]

It's a bit hacky though: The stored settings aren't shown in the GUI anywhere, so you can only inspect them at the python console.

Should not be a problem, from time to time I do some Python.

I'd be happy, if we moved the repo somewhere where you and others can contribute easily.

Show what you have ... Best to put it on your github (Doyou have one?) As a start we could make PRs. If needed you could give others write access to the repo.

[lambda]

Show what you have ... Best to put it on your github (Doyou have one?) As a start we could make PRs. If needed you could give others write access to the repo.

Ah, this is the misunderstanding! I linked it in an earlier post, so I wasn't aware you missed it: https://github.com/haraldg/Test_IFC_Importer

[bzzn87]

Hi! This is already a very old topic but for sharing the IFC test files, you could consider creating a separate repository or folder where interested contributors can access and download them. This way, you can keep the file size manageable while still making them available to everyone.
Including the test framework and a small collection of interesting IFC files in FreeCAD sounds like a good approach. As for the larger collection, having it in an external workbench/module would be a convenient option. It allows flexibility and keeps the main FreeCAD installation lightweight.

[yorik]

Sure, there are many of those repos around already! Many members of osarch.org have one (there is one on OpeningDesign's github - https://github.com/OpeningDesign/FreeMV ... Test_Files ) you could contribute with some files there...

IFC is all about collaboration. So it makes sense to have test files not only aimed at FreeCAD, but to verify the interaction and feature support with as many as possible other BIM applications