Skip to content

dev.md

Latest commit

 

History

History
70 lines (49 loc) · 1.7 KB

dev.md

File metadata and controls

70 lines (49 loc) · 1.7 KB
Raw

Raven Developer Guide

Build & Run from Source

The project uses Python + uv with hatchling as the build backend.

1. Install dependencies

cd /path/to/raven
uv sync

This creates/updates .venv with all core dependencies from uv.lock.

For optional extras:

uv sync --extra channels   # messaging integrations (Telegram, Slack, etc.)
uv sync --extra sandbox    # boxlite sandbox execution
uv sync --extra tools      # web/readability tools
# or all at once:
uv sync --all-extras

2. Install the package in editable mode

uv pip install -e .

This wires up the raven CLI script to your source code so changes take effect immediately.

3. Run the CLI

# via uv run (uses .venv automatically, no activation needed):
uv run raven --help

# or activate the venv first:
source .venv/bin/activate
raven --help

4. First-time setup

uv run raven onboard

Creates ~/.raven/config.json and the workspace directory. Edit the config to add your API key (default provider: OpenRouter — get a key at https://openrouter.ai/keys).

5. Common commands

Command Description
raven agent Start interactive chat session
raven agent -m "Hello" Send a single message and exit
raven gateway Start full server (all channels + heartbeat + cron)
raven status Show config path, workspace, and API key status
raven channels status Show which messaging channels are enabled
raven provider login <name> Authenticate with an OAuth provider (e.g. openai-codex)

6. Run tests

uv run pytest tests/

Requires Python ≥ 3.11. Test configuration is in pyproject.toml (asyncio_mode = "auto").