ansys/pymapdl

PyMAPDL Reader Independence

Opened this issue · 1 comments

germa89 commented

Summary

Currently there is duplication regarding the post-processing of MAPDL results. There are three libraries which can do this job:

Intended Outcome

Effort should be merged toward only one single library for post-processing. The current general agreement is to go with DPF.

Any post-processing reference and function inside pymapdl should point to DPF equivalent.

How will it work?

Retro-compatibility should be ensured. This process should be transparent to the user.
Use DPF to implement post_processing module. It requires to change 3-4 functions which are doing most of the heavy work.

Stages/steps

  • Prepare example with DPF for me to get familiar with the library. #1279
  • Use DPF for the mesh module in PyMAPDL. Tracked in #1299
    So far, we are pending on the VTK parser. That is key as mentioned in ansys/pydpf-core#331
  • Post-processing - It does not need to be migrated to DPF because it only uses the reader for the mesh, which should be fixed in the previous bullet point.
  • Result. This is going to be tricky. Tracked in: #1300

Notes

This issue replace ansys/pymapdl_roadmap#2

natter1 commented

Imho that would be a huge improvement. I'm not sure how much is covered via #1299, but one major issue forcing me again and again to go back to PyMAPDL Reader or PyMAPDL is the difference in plotting capabilities. I'm not sure if this is still correct, but e.g. animations are only possible (with lots of limitations) using animate_nodal_displacement. The capabilities of formating plots in general is most limited using DPF.