sim-plugin-abaqus 0.1.4

Dassault Systemes SIMULIA Abaqus driver for sim-cli, distributed as an out-of-tree plugin

sim-plugin-abaqus

Use Codex, Claude Code, or another AI agent to work with Dassault Systemes SIMULIA Abaqus input decks and Abaqus/CAE Python scripts through sim-cli.

sim-plugin-abaqus gives an agent practical Abaqus control paths: run .inp decks, run Abaqus/CAE Python scripts in Abaqus's embedded interpreter, keep an iterative noGUI CAE authoring session, inspect runtime health, collect model summaries, read job diagnostics, and report generated artifacts such as .dat, .msg, .sta, .cae, and .odb files.

The Abaqus application is not bundled. Bring your own Abaqus installation and license. See LICENSE-NOTICE.md.

This plugin is for Dassault SIMULIA Abaqus, not Ansys Mechanical. Use the Mechanical plugin for PyMechanical and Ansys Mechanical sessions.

What an agent can do with Abaqus

• Run complete Abaqus input decks with sim run --solver abaqus model.inp.
• Run Abaqus/CAE Python scripts through Abaqus's embedded Python.
• Build and debug CAE models incrementally with connect, exec, and inspect.
• Use a file-backed authoring session for conservative step-by-step work.
• Use the optional noGUI bridge backend for lower-latency trusted local loops.
• Inspect session.health, cae.model_summary, workdir.files, and job.latest before continuing a workflow.
• Treat solver completion as one signal, then validate against engineering acceptance criteria.

Choose the right Abaqus workflow

1. Batch input-deck execution

Use this when the model is already fully specified as an Abaqus .inp deck:

sim lint path/to/model.inp
sim run --solver abaqus path/to/model.inp

Abaqus writes output artifacts next to the deck. Inspect .msg, .sta, .dat, and .odb outputs before treating the result as accepted.

2. Abaqus/CAE Python script execution

Use this when the model is authored by a complete Abaqus/CAE Python script:

sim lint path/to/model.py
sim run --solver abaqus path/to/model.py

The plugin invokes abaqus cae noGUI=<script>, so Abaqus modules are imported inside Abaqus's embedded interpreter, not inside the sim-cli Python process.

3. Iterative CAE authoring session

Use this when an agent should build, inspect, debug, and report a model across multiple bounded snippets:

sim connect --solver abaqus --mode cae --ui-mode no_gui
sim inspect session.health
sim exec --file step.py
sim inspect cae.model_summary
sim inspect job.latest

The default file-backed backend starts Abaqus/CAE for each snippet, loads the session .cae database, mutates it, saves it, and returns diagnostics. This is the safest default for public agent workflows.

For trusted local workspaces that need faster repeated snippets, request the bridge backend:

sim connect --solver abaqus --mode cae --ui-mode no_gui --driver-option backend=bridge

Bridge mode starts one long-lived noGUI CAE process on localhost with a per-session token. Use it only in workspaces trusted by the user running the session.

Prerequisites

Install these before asking an agent to use this plugin:

• Python 3.10 or newer.
• uv for Python environment and package installs.
• sim-cli or a project environment where sim-cli can be installed.
• A local Abaqus installation available through abaqus.bat, abqNNNN.bat, or an explicit launcher path.

If Abaqus is installed but not discoverable, point the driver at the launcher:

$env:SIM_ABAQUS_COMMAND = 'C:\SIMULIA\Commands\abq2026.bat'
sim check abaqus

The plugin does not include Abaqus, vendor binaries, vendor SDKs, or licensed example content. It installs the Python adapter and bundled agent guidance only.

Install

For most users and agents, install the latest published PyPI version:

uv pip install sim-plugin-abaqus

You can also install through sim-cli's plugin command:

sim plugin install sim-plugin-abaqus

For quick testing of the current source branch, install from GitHub:

uv pip install "git+https://github.com/svd-ai-lab/sim-plugin-abaqus.git@main"

For a reproducible agent run, pin a commit SHA:

uv pip install "git+https://github.com/svd-ai-lab/sim-plugin-abaqus.git@<commit-sha>"

Verify Install

After installation, sim-cli should auto-discover the driver and bundled skill:

sim check abaqus

If sim check abaqus reports that Abaqus itself is unavailable, first confirm the Python package installed correctly, then fix the local Abaqus launcher or license prerequisites.

Connect And Inspect Health

Use a noGUI authoring session for bounded agent-driven CAE work:

sim connect --solver abaqus --mode cae --ui-mode no_gui
sim inspect session.health
sim inspect cae.model_summary
sim inspect workdir.files

Use one-shot execution for complete decks or scripts:

sim run --solver abaqus path/to/job.inp
sim run --solver abaqus path/to/cae_script.py

Common Agent Workflow

1. Confirm sim check abaqus is ok.
2. Choose batch .inp, complete CAE script, or iterative CAE authoring.
3. Gather geometry, materials, loads, boundary conditions, analysis type, and acceptance criteria before changing the model.
4. Run one bounded deck, script, or CAE snippet.
5. Inspect session.health, cae.model_summary, workdir.files, and job.latest.
6. Validate solver artifacts against physics-based acceptance criteria, not just exit code.

Troubleshooting

• sim command not found: install sim-cli in the same Python environment.
• Driver not discovered: reinstall the plugin and run sim check abaqus.
• Abaqus not detected: set SIM_ABAQUS_COMMAND to the exact launcher path.
• A .py script imports fail in normal Python: run it through sim run --solver abaqus script.py; Abaqus modules live in Abaqus's embedded interpreter.
• A job exits with code 0 but results look wrong: inspect .msg, .sta, .dat, and .odb outputs and compare against engineering criteria.
• Bridge backend stops responding: inspect session.health, disconnect, and reconnect with the default file-backed backend before continuing.

Agent quickstart

Give an agent this instruction when the task is about Abaqus:

Use the bundled Abaqus skill from sim-plugin-abaqus. First identify whether the
task needs a batch .inp run, a complete Abaqus/CAE Python script run, or an
iterative noGUI CAE authoring session. Confirm `sim check abaqus` first. For
iterative work, connect with `sim connect --solver abaqus --mode cae --ui-mode
no_gui`, run one bounded snippet at a time, then inspect `session.health`,
`cae.model_summary`, `workdir.files`, and `job.latest`. Validate against
physics-based acceptance criteria, not only exit code.

The bundled skill entry point is:

src/sim_plugin_abaqus/_skills/abaqus/SKILL.md

How it relates to sim-cli

sim-plugin-abaqus extends sim-cli with the Abaqus-specific driver and bundled Abaqus skill. sim-cli supplies the common runtime surface (check, lint, run, connect, exec, and inspect), while this plugin supplies Abaqus detection, subprocess execution, CAE authoring, health checks, model summaries, job diagnostics, and bundled Abaqus agent guidance.

The plugin registers three entry-point groups:

[project.entry-points."sim.drivers"]
abaqus = "sim_plugin_abaqus:AbaqusDriver"

[project.entry-points."sim.skills"]
abaqus = "sim_plugin_abaqus:skills_dir"

[project.entry-points."sim.plugins"]
abaqus = "sim_plugin_abaqus:plugin_info"

Develop

git clone https://github.com/svd-ai-lab/sim-plugin-abaqus
cd sim-plugin-abaqus
uv sync
uv run pytest tests -m "not integration"

End-to-end tests require a local Abaqus installation and are skipped unless their prerequisites are available.

License

Apache-2.0. See LICENSE and LICENSE-NOTICE.md.