# AgriMesh Federated Privacy

AgriMesh is a Python package for privacy-preserving coordination across neighboring farms. It models local irrigation and resource constraints, exchanges aggregated signals, records tamper-evident plan deltas in SQLite, and produces coordinated irrigation plans for drought conditions.

## What It Includes

- Pydantic models for farms, local problems, shared signals, regional constraints, adapters, and plan deltas.
- A SQLite governance ledger for farms, contracts, adapters, events, and offline delta sync.
- An ADMM-lite irrigation allocator that respects farm budgets and regional water and energy caps.
- A canonical mapping layer that converts AgriMesh inputs into a CatOpt-style intermediate representation.
- A drought demo for 2 to 3 farms.

## Package Layout

- `src/idea170_agrimesh_federated_privacy/models.py` - validated data models.
- `src/idea170_agrimesh_federated_privacy/ledger.py` - SQLite governance ledger.
- `src/idea170_agrimesh_federated_privacy/solver.py` - cross-farm allocator.
- `src/idea170_agrimesh_federated_privacy/canonical.py` - CatOpt mapping.
- `src/idea170_agrimesh_federated_privacy/demo.py` - drought scenario builder.

## Quick Start

```python
from idea170_agrimesh_federated_privacy import build_drought_demo

result = build_drought_demo()
for delta in result.deltas:
    print(delta.farm_id, delta.irrigation_liters)
```

```bash
bash test.sh
python3 -m idea170_agrimesh_federated_privacy
```

## Testing

The repository is wired for:

- `bash test.sh`
- `pytest -q`
- `python3 -m build`

## Packaging

This project is published as `idea170-agrimesh-federated-privacy` and uses `README.md` as its package description source.