# NovaPlan MVP

A minimal, open-source MVP for decentralized, privacy-preserving multi-agent mission planning in deep-space robotic constellations.

- Offline-first, privacy-aware coordination across heterogeneous fleets (rovers, drones, habitat bots).
- Local problem solving with a tiny ADMM-like core and federation of agents.
- A lightweight mission ledger that can anchor decisions when ground links are available.
- Lightweight adapters for common hardware and simulation-ready interfaces.
- A clear test and packaging path to verify end-to-end viability.

This repository focuses on a small, well-scoped subset of the NovaPlan ecosystem to demonstrate core ideas and enable further expansion.

How to run tests
- Run: `./test.sh`
- This will execute unit tests and verify packaging with `python -m build`.

MVP extensions (planned, small scope)
- Add a minimal DSL (nova_plan.dsl) and a delta-synchronization helper (delta_sync) to scaffold federated optimization flows.
- Introduce a lightweight ContractRegistry to enable versioning and interoperability of data contracts.
- Provide a tiny DSL-to-LocalProblem translator to bootstrap CatOpt-style integration with adapters.
- Expand ledger anchoring with a deterministic reconciliation log for outages.

Directory layout
- nova_plan/          Core MVP implementation (planner, contracts, ledger, adapters)
- tests/              Unit tests for core workflow and contracts
- adapters/           Stubs for rover and habitat modules
- README.md, AGENTS.md  Documentation and governance for the project
- pyproject.toml        Build metadata for packaging
- AGENTS.md            Architecture and contribution guidelines
- READY_TO_PUBLISH     (created after the repo is ready for publishing)

Note: This is a minimal MVP intended for demonstration and testing; it is not a production-ready system.