Introduction

The Rbbt system is a collection of functionalities (tools in the toolkit) organized into an API. These functionalities are powered by one another and by a sophisticated infrastructure of resources.

The Rbbt has the Workflow module to package functionalities. These can be accessed any of the following ways:

• Programmatically from Ruby
• From the command-line through the rbbt command
• Through the HTML interface
• Through the REST web server using a programmatic API

Implementing workflows is simplified thanks to the extensive collection of tools in Rbbt. The can be maintained independently from the base system, and managed through the rbbt command to install or update them from different sources, or to use them remotely.

Any software that uses Rbbt will tap into the same infrastructure, sharing not only the functionalities and resources, but also the caches, indexes and analysis results that get created over time.

Objectives

There are several objectives that the Rbbt framework tries to achieve:

• Complete reproducibility and provenance
• Complete re-usability
• Promoting of good development practices by using incentives
• Being flexible and un-intrusive; standards and practices can be ignored and worked around for particular cases
• Providing an executable domain specific language (DSL) that can best express the processes in bioinformatics.
• Allowing the best tool to be used for each task; i.e. R for data analysis, Sinatra, HAML & SASS for web development, shell scripts for system administration, basic Unix/Linux tools for efficient file processing, etc
• Achieving production-ready performance in basic operations, even over databases of millions of rows, without sacrificing simplicity

Achieving these objectives in Rbbt took many years and the development of several programming paradigms, which use extensively Ruby meta-programming. While the inner workings of the framework are very elaborate, the final DSL is very expressive and succinct.