Run And Review¶
astraeus run executes one versioned RunSpec. A RunSpec defines the simulation
window, constellation, requested engineering stages, and the configuration those
stages consume. Astraeus validates the input, resolves the stage plan, executes
it, and writes a self-contained run directory.
flowchart LR
INPUT["RunSpec"] --> VALIDATE["Validate and resolve"]
VALIDATE --> EXECUTE["Execute stage plan"]
EXECUTE --> BUNDLE["runs/core/<run_id>/"]
BUNDLE --> REVIEW["Console, review, report, or direct analysis"]
Use a single run when one configuration answers the question. Use a sweep for related parameter variants or a batch for independently authored RunSpecs.
The Core Contract¶
Every completed run preserves:
- the canonical RunSpec that was executed
- execution status, stage plan, warnings, and provenance
- compact engineering and environment summaries
- stage-specific artifacts such as states, access, coverage, mission, transport, attitude, power, or propulsion results
The exact artifact families depend on the requested stages. Read Understanding outputs for the stable bundle contract and the model pages for stage-specific results.
Recommended Workflow¶
- Start from the nearest example.
- Keep only the stages needed for the engineering question.
- Validate with
astraeus run <runspec> --dry-run. - Execute with a deliberate
--run-idand output location. - Review the manifest and metrics before opening detailed artifacts.
- Use the Console, deterministic review, or report command as appropriate.
Guide Map¶
| Page | Purpose |
|---|---|
| RunSpec basics | Input structure and stage composition |
| Mission types | Service and observation mission behavior |
| Authoring RunSpecs | Building a study from examples and reusable profiles |
| Running simulations | Execution, validation, options, and failures |
| Run Monitor | Read-only progress and operational visibility |
| Understanding outputs | Run, sweep, and batch artifacts |
| Console and visualization | Local configuration, launch, and interactive result inspection |
For a guided first execution, use First core run.