1348109517 Abaqus Script Debugging Checklist
Overview
Establish the first failing boundary before proposing a fix. Capture execution context, exact error text, repository state, and artifact state separately.
When to use
Use for reproducible scripting failures and surprising empty selections. Pair it with a domain skill only after the failing boundary is localized.
Inputs
- Exact command or GUI action and interpreter context
- Complete traceback or message
- Relevant configuration and source revision
- Expected and observed repository keys or artifacts
Outputs
Return a diagnosis containing reproduction, first failing operation, evidence, root cause confidence, smallest reversible fix, and verification command.
Workflow
- Reproduce with the smallest read-only or build-only command available.
- Confirm the intended Python interpreter and Abaqus execution mode.
- Inspect repository keys before indexing or selecting regions.
- Trace the failed value backward to its producer.
- Write a regression check before applying a code fix.
Safety gates
- Do not rerun an expensive or destructive job merely to obtain a traceback.
- Do not delete databases, lock files, or generated models during diagnosis.
- Do not replace an empty region with a broader region without geometry evidence.
- A successful import proves neither model construction nor solver correctness.
Example prompts
Diagnose why a surface lookup is empty after partitioning. Record repository keys and the producer-consumer chain. Do not broaden the selection or run a job.
Common failures
- Fixing the last exception instead of the first invalid state.
- Testing in regular Python when the script requires Abaqus modules.
- Indexing repositories without inspecting available keys.
- Confusing a missing result file with a failed field extraction.
Acceptance checklist
- Exact execution context and error text are recorded.
- The first failing boundary is identified.
- Root cause is separated from downstream symptoms.
- Proposed change is minimal and reversible.
- Verification can distinguish the fix from an unrelated success.