Status:
2024-01-22
Autopilot is in maintenance-only mode - development has been paused as we take the long way around towards building a new kind of p2p networking module to support a reworked autopilot 2.0. We will write a more detailed blogpost about lessons learned from autopilot soon.
Autopilot is not dead, it is merely resting <3
-jonny
Docs | Paper | Forum | Wiki |
---|---|---|---|
Autopilot is a Python framework for performing complex, hardware-intensive behavioral experiments with swarms of networked Raspberry Pis. As a tool, it provides researchers with a toolkit of flexible modules to design experiments without rigid programming & API limitations. As a vision, it dreams of bridging the chaotic hacky creativity of scientific programmers with a standardized, communally developed library of reproducible experiment prototypes.
Autopilot was developed with three primary design principles:
Autopilot's premise is simple: to scale experiments, just use more computers.
Autopilot systems consist of multiple "Agents" -- computers with specialized roles in the swarm. One user-facing "Terminal" agent allows a researcher to control many "Pilots," or computers that perform experiments (typically the beloved Raspberry Pi). Each Pilot can coordinate one or many "Children" to offload subsets of an experiment's computational or hardware requirements. Users can use and misuse Autopilot's flexible modules to make whatever agent topology they need <3.
Autopilot divides the logical structure of experiments into independent1 modules:
Module | |
---|---|
Agents - Pilot & Terminal Runtime classes that encapsulate a computer/Pi's role in the swarm. Terminals provide the user interface and coordinate subjects and tasks, Pilots do the experiments. Formalizing the Agent API to allow additional agents like Compute or Surveillance agents is a major short-term development goal! | |
Hardware - Control your tools! Extensible classes to control whatever hardware you've got. | |
Stimuli - Stimulus management and presentation. Parametric sound generation with a realtime audio server built on Jackd. Stubs are present for future development of visual stimuli using Psychopy. | |
Tasks - Build experiments! Write some basic metadata to describe data, plots, and hardware and the rest is up to you :) | |
Subject - Data management with hdf5 and pyTables. Abstraction layer for keeping obsessive records of subject history and system configuration | |
Transforms - Composable data transformations. Need to control the pitch of a sound with a video? build a transformation pipeline to connect your objects | |
UI - UI for controlling swarms of Pilots using Qt5/PySide2 | |
Visualization - (Mostly Prototypes) to do common visualizations |
1 a continual work in progress!
All documentation is hosted at https://docs.auto-pi-lot.com
Installation is simple, just install with pip and use Autopilot's guided setup to configure your environment and preferences. The initial setup routine uses a CLI interface that is SSH friendly :)
pip3 install auto-pi-lot
python3 -m autopilot.setup.setup
All of Autopilot is quite new, so bugs, incomplete documentation, missing features are very much expected! Don't be shy about raising issues or asking questions in the forum.
Jonny is trying to graduate! Autopilot will be slow and maybe a little chaotic until then!
We're working on a formal contribution system, pardon the mess! Until we get that and our CI coverage up, main
will lag a bit behind the development branches:
dev
- main development branch that collects hotfixes, PRs, etc. Unstable but usually has lots of extra goodieshotfix
- branches from dev
for building and testing hotfixes, PRs back to dev
.lab
- branches from dev
but doesn't necessarily PR back, the local branch used in the maintaining (Wehr) labparallax
- experimental departure from dev
to implement a particular experiment and rebuild a lot of components along the way, will eventually return to dev
<3See the short-term development goals in our version milestones:
v0.4.0
- Implement registries to separate user code extensions like tasks and local hardware devices in a user directory, preserve source code in produced data so local development isn't lost. v0.5.0
- Make a unitary inheritance structure from a root Autopilot object such that a) common operations like logging and networking are implemented only once, b) the plugin system for v0.4.0
can not only add new objects, but replace core objects while maintaining provenance (ie. no monkey patching needed), c) object behavior that requires coordination across multiple instances gets much easier, making some magical things like self-healing self-discovering networking possible. This will also include a major refactoring of the code structure, finally breaking up some of the truly monstrous thousand-line modules in core
into an actually modular system we can build from <3Autopilot's extended development goals, in their full extravagance, can be found at the Autopilot Development Todo
After much ado, we're releasing Autopilot's first major upgrade. Cameras, Continuous data, DeepLabCut, and a lot more!
OS
Python Versions
Raspberry Pi Versions