Add method to convert DAGMC volume to STEP #203

Edgar-21 · 2025-03-19T21:32:33Z

Closes #202

Two methods are added to parastell.utils, one which reads STL files defining the surfaces of the volume to be converted and returns a CadQuery Solid defined by those surfaces, and another than streamlines the process by taking a PyDAGMC DAGModel object and a volume ID and writing the corresponding CadQuery Solid to STEP.

gonuke

This all looks pretty straightforward, but just one question out of ignorance that might simplify writing the STLs.

gonuke · 2025-03-20T01:47:32Z

parastell/utils.py

+    """
+    volume = dagmc_model.volumes_by_id[volume_id]
+
+    num_tris = len(volume.triangle_handles)


Did you think about just writing these triangles (volume.triangle_handles) into a single STL? I have no idea whether it would work, just curious?

I'm not sure what it would look like to use the triangle handles directly - I know that both mbconvert and something like vol.model.mb.write_file(stl_path, output_sets=[vol.handle]) fail to write volumes to stl, while writing surfaces to stl works just fine. I also have had better luck making these volumes in CadQuery starting from surfaces, though that might be more related to user error when I was first figuring things out.

It might be necessary to make a new set that contains the volume triangles - the write method may want a "set" rather than a "range" (which is what volumes.triangle_handles is)

It may not be worth it - or we could make it an issue for an optional simplification in the future.

Since I can't write the triangles in the volume directly to STL with the same method that works with surfaces, (seems to be some problem with MOAB) I think anything we come up with will be about the same level of complexity as writing each surface in a volume to STL. I'm inclined to leave it as is.

Right, I think the issue is that the volume entity set itself doesn't contain any triangles. See the image on the data model below from one of @gonuke's slide decks.

The VTK writer in MOAB seems to traverse any child sets of the output_sets argument to collect elements -- probably because the VTK format is more generic, accommodating many different element types/dimensions so it does more work. In contrast, STL files (generally) contain single groups of triangles (this is how an STL will be read into MOAB, for instance, so I think MOAB is expecting to find the elements in the set(s) passed to the write_file call. This doesn't mean the STL writer in MOAB shouldn't be traversing the child sets like the VTK reader, just an explanation as to why it may not based on the allowed contents of these two different file types.

So, rather than passing the volume handle, collecting the handles of the volume's surface sets (which do contain triangle elements) would do the trick. It would look something like:

model.mb.write_file('vol.stl', output_sets=[s.handle for s in volume.surfaces])

I gave this a try using one of the .h5m files in the OpenMC DAGMC regression tests and it seemed to work alright.

Here's a simple example:

import dagmc dm = dagmc.DAGModel('dagmc.h5m') vol = dm.volumes_by_id[2] print('writing to vtk...') vol.model.mb.write_file('vol2.vtk',output_sets=[vol.handle]) print('writing to stl...') vol.model.mb.write_file('vol2.stl',output_sets=[vol.handle])

and the output

writing to vtk... writing to stl... Writer with name STL for file vol2.stl using extension stl (file type "(null)") was unsuccessful MOAB ERROR: --------------------- Error Message ------------------------------------ MOAB ERROR: No triangles to write! MOAB ERROR: write_file() line 96 in src/io/WriteSTL.cpp Using default writer WriteHDF5 for file vol2.stl

The vtk file is valid, I can open it up in paraview just fine. It does write a file for the stl, but writes to HDF5 with an STL suffix.

Attempting to write the triangle handles with:

vol.model.mb.write_file('vol2.stl', output_sets=vol.triangle_handles)

give this error

Cell In[2], line 1 ----> 1 vol.model.mb.write_file('vol2.stl', output_sets=vol.triangle_handles) File pymoab/core.pyx:153, in pymoab.core.Core.write_file() OSError: Only EntitySets should be passed to write file.

Which is strange since I believe vol.triangle_handles is a range and the docstring for write_file seems to indicate ranges are valid for output_sets.

Whoops, didn't see your comment here @gonuke. I think output_sets does need to be either an EntitySet or Iterable of EntitySet's as opposed to element handles, but it's been a while and I could be wrong here.

Ah I got it vol.model.mb.write_file('vol2.stl', output_sets=vol._get_triangle_sets()) works. output_sets needs a range of EntitySet

I need to refresh PR pages more often I missed all the comments in the past half hour and was just blissfully writing my own :)

gonuke

I think the simplification was worth it - now a few final touch ups.

gonuke · 2025-03-20T23:54:38Z

parastell/utils.py

+    with tempfile.NamedTemporaryFile(delete=True, suffix=".stl") as temp_file:
+        stl_path = temp_file.name
+        volume.model.mb.write_file(
+            stl_path, output_sets=volume._get_triangle_sets()


I think I like this suggestion from @pshriwise better since it doesn't rely on the nominally private interface

Suggested change

stl_path, output_sets=volume._get_triangle_sets()

stl_path, output_sets=[s.handle for s in volume.surfaces]

gonuke · 2025-03-21T00:02:40Z

parastell/utils.py

+            stl_path, output_sets=volume._get_triangle_sets()
+        )
+        cq_solid = stl_to_cq_solid(stl_path, tolerance)
+    num_faces = len(cq_solid.Faces())


Since it is only used here, can we move the definition of num_tris to here?

gonuke · 2025-03-21T00:03:39Z

tests/test_utils.py

+
+    with tempfile.NamedTemporaryFile(delete=True, suffix=".stl") as temp_file:
+        stl_path = temp_file.name
+        vol.model.mb.write_file(stl_path, output_sets=vol._get_triangle_sets())


as above

Suggested change

vol.model.mb.write_file(stl_path, output_sets=vol._get_triangle_sets())

dagmc_model.mb.write_file(stl_path, output_sets=[s.handle for s in vol.surfaces])

gonuke · 2025-03-21T00:05:48Z

parastell/utils.py

+
+    with tempfile.NamedTemporaryFile(delete=True, suffix=".stl") as temp_file:
+        stl_path = temp_file.name
+        volume.model.mb.write_file(


Since you already have the model, this may be better

Suggested change

volume.model.mb.write_file(

dagmc_model.mb.write_file(

gonuke · 2025-03-21T00:07:24Z

tests/test_utils.py

+    num_faces = len(cq_solid.Faces())
+
+    assert num_faces == num_tris
+    assert math.isclose(dagmc_volume_volume, cq_solid_volume)


either pytest or numpy have a way to do this instead of math?

gonuke · 2025-03-21T00:07:53Z

tests/test_utils.py

+    dag_model = dagmc.DAGModel("files_for_tests/one_cube.h5m")
+    vol = dag_model.volumes_by_id[vol_id]
+    dagmc_volume_volume = vol.volume
+    num_tris = len(vol.triangle_handles)


Move this close to the definition of num_faces since they are used together

gonuke

Thanks for this addition @Edgar-21

Edgar-21 added 3 commits March 19, 2025 16:07

correct docstring in cubit utils as we no longer use the plugin

e947f87

add function to utils for writing DAGMC volumes to step

697c2b2

improve docstrings

b6bfe86

Edgar-21 added the enhancement New feature or request label Mar 19, 2025

gonuke reviewed Mar 20, 2025

View reviewed changes

write the entire volume to a single stl file

d8ecec6

gonuke requested changes Mar 21, 2025

View reviewed changes

Edgar-21 added 2 commits March 21, 2025 08:35

dont use private method

f2c4da1

tidying up per pr comments

233cf1f

gonuke approved these changes Mar 21, 2025

View reviewed changes

gonuke merged commit ca82364 into main Mar 21, 2025
3 checks passed

Edgar-21 deleted the pydagmc_ivb_step_files branch March 21, 2025 21:20

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add method to convert DAGMC volume to STEP #203

Add method to convert DAGMC volume to STEP #203

Edgar-21 commented Mar 19, 2025

gonuke left a comment

gonuke Mar 20, 2025

Edgar-21 Mar 20, 2025 •

edited

Loading

gonuke Mar 20, 2025

gonuke Mar 20, 2025

Edgar-21 Mar 20, 2025

pshriwise Mar 20, 2025 •

edited

Loading

Edgar-21 Mar 20, 2025 •

edited

Loading

pshriwise Mar 20, 2025 •

edited

Loading

Edgar-21 Mar 20, 2025

Edgar-21 Mar 20, 2025

gonuke left a comment

gonuke Mar 20, 2025

gonuke Mar 21, 2025

gonuke Mar 21, 2025

gonuke Mar 21, 2025

gonuke Mar 21, 2025

gonuke Mar 21, 2025

gonuke left a comment

	stl_path, output_sets=volume._get_triangle_sets()
	stl_path, output_sets=[s.handle for s in volume.surfaces]

	vol.model.mb.write_file(stl_path, output_sets=vol._get_triangle_sets())
	dagmc_model.mb.write_file(stl_path, output_sets=[s.handle for s in vol.surfaces])

Add method to convert DAGMC volume to STEP #203

Add method to convert DAGMC volume to STEP #203

Conversation

Edgar-21 commented Mar 19, 2025

gonuke left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Edgar-21 Mar 20, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pshriwise Mar 20, 2025 • edited Loading

Choose a reason for hiding this comment

Edgar-21 Mar 20, 2025 • edited Loading

Choose a reason for hiding this comment

pshriwise Mar 20, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gonuke left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gonuke left a comment

Choose a reason for hiding this comment

Edgar-21 Mar 20, 2025 •

edited

Loading

pshriwise Mar 20, 2025 •

edited

Loading

Edgar-21 Mar 20, 2025 •

edited

Loading

pshriwise Mar 20, 2025 •

edited

Loading