WARG Autonomy Monorepo

This repository is organized as a flat monorepo for WARG autonomy projects. Each top-level directory is a project, and every project is registered from the root projects.toml file.

Layout

.
├── README.md
├── projects.toml
└── <project>/
    ├── warg.toml
    └── ...

Projects own their local source, tests, dependencies, and commands. The root of the repo owns only shared repository documentation and the project registry.

Project registry

Register every project in projects.toml:

[projects.camera]
path = "camera"

[projects.mavlink_comm]
path = "mavlink_comm"

[projects.gesture_control]
path = "gesture_control"

The registry is intentionally explicit so dependency resolution and sparse checkout paths stay predictable.

Project manifests

Each project should include a warg.toml manifest:

name = "gesture_control"
description = "Gesture control pipeline for camera-based command input."

depends_on = ["camera", "mavlink_comm"]

[commands]
setup = "uv sync"
test = "uv run pytest"
"test:unit" = "uv run pytest tests/unit"
"test:integration" = "uv run pytest tests/integration"
run = "uv run python -m gesture_control"
lint = "uv run ruff check ."

Commands are project-defined, similar to scripts in package.json. Keep project-specific setup and workflows in the project manifest rather than in the root README.

Projects

• warg_cli: developer CLI for materializing projects, inspecting manifests, and running project-defined commands. See warg_cli/README.md.

• gnc: Guidance, Navigation, and Control module.

• perception: Computer vision and target tracking algorithms module.

• IMS: Intelligent Monitoring System, a dashboard displaying MAVLink message feed, drone state, and live camera feed.

• Camera: A Hardware Abstraction Layer for cameras. Normalizes frames from ArduCam, Oak-D, and simulation.

• Utils: Shared enums and dataclasses used across all Airside modules.

• Mav_comms: Wrapper around pymavlink.

• airside: ROS 2 Humble workspace for the airside architecture. See airside/README.md.