scimate-ai/hpc-skills/hpc-paraview

HPC ParaView

Treat ParaView as a pipeline-driven post-processing and visualization stack with both GUI and scripted execution paths.

Start

1. Read references/workflow-manual.md before creating or repairing a ParaView workflow.
2. Read references/readers-filters-and-pipeline.md when mapping datasets to readers, filters, and pipeline objects.
3. Read references/color-maps-layout-and-views.md when tuning representations, coloring, views, screenshots, or layout behavior.
4. Read references/pvpython-pvbatch-and-traces.md when choosing between GUI tracing, pvpython, and pvbatch.
5. Read references/state-files-animation-and-extracts.md when using .pvsm, Python state, screenshots, or saved data products.
6. Read references/remote-and-parallel-visualization.md when the workflow needs pvserver, reverse connections, SSH tunnels, or distributed rendering.
7. Read references/cluster-execution-playbook.md when staging a ParaView workflow for scheduler-backed cluster post-processing.
8. Read references/error-recovery.md when readers, filters, scripts, state files, or remote connections fail.

Reusable Templates

Use assets/templates/ when a concrete starting scaffold is needed, especially:

• pvpython_screenshot.py
• pvbatch_slice_extract.py
• paraview-pvbatch-slurm.sh
• paraview-pvserver-slurm.sh

Guardrails

• Do not guess reader or filter semantics from file extensions alone when metadata is available.
• Do not use pvbatch and remote Connect() logic in the same script.
• Do not save brittle state files with absolute data paths unless that is intentional.
• Do not run heavy rendering or batch extraction on login nodes.

Outputs

Summarize:

• input datasets and readers
• key filters and representations
• chosen execution path such as GUI, pvpython, pvbatch, or pvserver
• expected screenshots, extracts, animations, or saved data products