Skip to main content

Tools for writing adaptive plans

Project description

Build Status Latest PyPI version BSD 3-Clause License

Bluesky Adaptive is a component of the Bluesky ecosystem designed to introduce adaptivity and intelligent decision-making into experimental workflows. It provides a flexible API that supports a wide range of adaptive algorithms, from simple rule-based adjustments to sophisticated artificial intelligence and machine learning models. The package is designed to provide a framework and set of harnesses for “bring-your-own-agent” development. It has no opinions on the algorithms that underpin your adaptivity, and it is designed to be agnostic to the specifics of the domain in which it is used.

PyPI

pip install bluesky-adaptive

Source code

https://github.com/bluesky/bluesky-adaptive

Documentation

https://blueskyproject.io/bluesky-adaptive

Features

  • Adaptive harness for lockstep agents to interact with the RunEngine.

  • Base classes for developing distributed agents to interact with the RunEngine Manager through Queue Server.

  • Example agents using optional requirements (BoTorch, GPyTorch, Scikit Learn).

  • Server for running managed agents with a FAST API.

  • Adjudicators for acting as meta-agents that consume suggestions from many agents and gatekeep the Queue Server.

Getting Started

To get started with Bluesky-Adaptive, refer to the Getting Started Tutorial. This tutorial will guide you through the installation process, setting up the Bluesky stack, and running your first adaptive experiment.

Usage

Bluesky Adaptive is designed to be integrated into your experimental workflows to provide adaptive capabilities. It offers two primary approaches:

  • Lockstep Approach: For synchronous operations where each experimental step waits for the previous one to complete.

  • Asynchronous Approach: Allows for independent operation from the main execution thread, suitable for complex setups.

Detailed guides on how to extend your existing Python tools for use with Bluesky Adaptive and how to run your agent as a service can be found in the documentation.

Contributing

We welcome contributions from the community, whether it’s adding new features, improving documentation, or reporting bugs. Please see our contribution guidelines for more information on how to get involved.

License

Bluesky-Adaptive is distributed under the BSD-3-Clause license. See the LICENSE file for more details.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

bluesky_adaptive-0.3.2.tar.gz (91.7 kB view details)

Uploaded Source

Built Distribution

bluesky_adaptive-0.3.2-py3-none-any.whl (75.7 kB view details)

Uploaded Python 3

File details

Details for the file bluesky_adaptive-0.3.2.tar.gz.

File metadata

  • Download URL: bluesky_adaptive-0.3.2.tar.gz
  • Upload date:
  • Size: 91.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.3

File hashes

Hashes for bluesky_adaptive-0.3.2.tar.gz
Algorithm Hash digest
SHA256 00809e52e170e5b962f94ecfbcbc56aa5ac515d878809c8fd0ca121c380c4ebe
MD5 e707fec114a013688fb1b8cdc05a7f06
BLAKE2b-256 18245e80ab7323dde796e9781fe6bd34d11e854b3a330b144a12c588192414de

See more details on using hashes here.

File details

Details for the file bluesky_adaptive-0.3.2-py3-none-any.whl.

File metadata

File hashes

Hashes for bluesky_adaptive-0.3.2-py3-none-any.whl
Algorithm Hash digest
SHA256 b9aa52475d78ffc2aa6fc758c35da981ec3ef83093b3a44c9e869ea02481140d
MD5 1d64789ffb787d95cf8e54bd2474258b
BLAKE2b-256 92026b06969a043215b257e617ad0ccf75de7285e5b2f0570d1b67a009a3efe3

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page