Fix Field Indexer and get_element_indices for layered solids and add example of a thermal analysis

[roosre]

While adding an example for a thermal analysis (examples/013_thermal_example.py), several issues were found in the result extraction for layered solid elements where the field indexers are used. The main issue is that fields which have one entry per entity behave slightly different if compared with one that has multiple entries per entity. This became visible because a solid model with "analysis ply-wise" extrusion was used. So there is only one entry per entity for the layered fields such as angles, materials etc.

Here are some details:

• FieldIndexer has to handle cases where the data pointers of a field are set in DPF or not. This is a performance optimization in DPF. The data pointers are only available if multiple entries are set per entity. This is addressed by implementing helper functions get_property_field_indexer and get_field_indexer which automatically returns the correct indexer.
• The Protocol of the (Property)FieldIndexer was enriched by by_id_as_array to clearly distinguish between a single value and an array of values because the return type was dependent on if data pointers are available or not.
• An issue with the DPF container/server 2024 R1 pre0 on extracting laminate offsets is fixed. This did not show up before because the FieldIndexer which returns a single value was used explicitly. get_field_indexer returns now the correct indexer but the bug in the backend becomes visible now.
• By accident, the number of corner nodes was used instead of number of nodes per spot when computing the indices. This makes a difference if layered solid elements are processed. See get_selected_indices in src/ansys/dpf/composites/select_indices.py.

[greschd]

General question: are we planning to backport these fixes to the 0.6 release branch and 2024R2 pyansys metapackage?

[roosre]

General question: are we planning to backport these fixes to the 0.6 release branch and 2024R2 pyansys metapackage?

Hi @greschd , yes. That would make sense from my point of view.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 87.17949% with 15 lines in your changes missing coverage. Please review.

Project coverage is 91.66%. Comparing base (d8d8cc6) to head (83d1310).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/ansys/dpf/composites/_indexer.py	83.33%	15 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #512      +/-   ##
==========================================
- Coverage   92.21%   91.66%   -0.55%     
==========================================
  Files          50       50              
  Lines        3352     3397      +45     
==========================================
+ Hits         3091     3114      +23     
- Misses        261      283      +22     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.