/connected

An experimental data-sharing library

Primary LanguagePythonMIT LicenseMIT

connected

Experiments in Pipeline/Workflow syntax

This project is an exploration of Python syntax and recent language features, intended to implement an inheritable base class for applications that require sharing data across invocations and between objects, such as AI pipelines and workflows. Its intention is primarily a personal learning exercise, but hopes to find a novel solution to the problem through experimentation. It draws inspiration from langchain's LCEL (LangChanin Expression Language).

Getting Started

To get started with this project, follow these steps:

  • Clone the repository: git clone https://github.com/martinwalsh/connected.git
  • Run rye sync (for more information on rye check out the rye documentation)
  • Explore the code and experiment with the Context and LazyRef objects.
  • Take a look in the tests directory for examples.
  • Share your feedback and contribute to the project if you find it interesting!

Resources

License

This project is licensed under the MIT License. See the LICENSE file for more information.