Celebi
Celebi is a data analysis management toolkit designed for high-energy physics research. It provides a structured environment for organizing projects, tasks, algorithms, and data, enabling reproducible, traceable, and collaborative scientific workflows.
Celebi is particularly suited for complex analysis chains common in HEP, where dependency tracking, provenance, and re-execution are critical.
Key Features and Benefits
Structured Organization Clear separation of projects, data, algorithms, and tasks with a well-defined hierarchy.
Dependency Tracking Automatic tracking of relationships between data, algorithms, and tasks, forming a directed acyclic graph (DAG).
Impressions (Versioning) Built-in impressions system to snapshot important results, configurations, and object states over time.
Reproducibility Complete capture of workflow structure, parameters, inputs, and execution environments.
Adaptability Modify algorithms or parameters and re-run only affected downstream tasks.
Collaboration Share projects and workflows consistently across users and environments.
Core Concepts
Project – A self-contained analysis workspace.
Data – Raw or derived datasets registered and managed by Celebi.
Algorithm – A reusable, self-contained piece of code or script.
Task – A concrete execution instance of an algorithm with specific inputs and parameters.
Runner – An execution backend (local, batch system, remote, etc.).
Impression – A recorded snapshot of key outputs or analysis states.
Getting Started
To initialize a Celebi project and start the CLI:
celebi init
celebi <command>
Documentation
For detailed usage, see the following pages:
Additional Resources
Online Docs: Celebi Documentation
Ask DeepWiki: DeepWiki Celebi Page
License
Apache License, Version 2.0