Demo workflow data not copying into cache directory

[gcapes]

Hi @aplowman,

For the two demo workflows I'm trying to put together, when I try to use demo data from a zip file, I get an error. Both workflows work with a normal full path to the same data.

Traceback (most recent call last):
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/cooling_DAMASK_Zr_2026-07-02_135725/artifacts/submissions/0/scripts/generate_rve_from_damask_input_files_act_0.py", line 68, in <module>
    outputs = generate_volume_element_from_damask_input_files(**func_kwargs)
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/cooling_DAMASK_Zr_2026-07-02_135725/artifacts/submissions/0/scripts/generate_rve_from_damask_input_files_act_0.py", line 8, in generate_volume_element_from_damask_input_files
    geom_dat = read_geom(geom_path)
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/.venv-3ddemoworkflow/lib/python3.13/site-packages/damask_parse/readers.py", line 325, in read_geom
    ve_grid = grid_cls.load(geom_path)
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/.venv-3ddemoworkflow/lib/python3.13/site-packages/damask/_geomgrid.py", line 262, in load
    return GeomGrid._load(fname,'material')
           ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/.venv-3ddemoworkflow/lib/python3.13/site-packages/damask/_geomgrid.py", line 233, in _load
    v = VTK.load(fname if str(fname).endswith('.vti') else str(fname)+'.vti')
  File "/home/mbexegc2/projects/lightform/3D_EBSD_workflow/.venv-3ddemoworkflow/lib/python3.13/site-packages/damask/_vtk.py", line 367, in load
    raise FileNotFoundError(f'file "{fname}" not found')
FileNotFoundError: file "/home/mbexegc2/.cache/matflow/data/Zr_cooling_data/geom_example.vti" not found

The cache directory is created /home/mbexegc2/.cache/matflow/data/Zr_cooling_data/ but it is empty. I've deleted the cache when I first got this error so I know the empty directories are created by the workflow.

The workflows themselves can be found at https://github.com/hpcflow/matflow/blob/zr_cooling_damask/matflow/data/workflows/cooling_DAMASK_Zr.yaml and https://github.com/hpcflow/matflow/blob/3d-mechanical-loading-demo/matflow/data/workflows/tension_DAMASK_Zr_3D.yaml

The tension_DAMASK_Zr_3D uses demo data which is already in the matflow-data repo, but I get the same cache issue whether or not my matflow config file points to my local copy of the matflow-data repo, or uses the default settings.

Are you able to reproduce this (for tension_DAMASK_Zr_3D) or have any ideas what the problem is?

Thanks!