Romanesco: A simple, flexible execution engine¶
What is Romanesco?¶
Romanesco is a python application for generic task execution. It can be run within a celery worker to provide a distributed batch job execution platform.
The application can run tasks in a variety of languages and environments, including python, R, spark, and docker, all via a single python or celery broker interface. Tasks can be chained together into workflows, and these workflows can actually span multiple languages and environments seamlessly. Data flowing between tasks can be automatically converted into a format understandable in the target environment. For example, a python object from a python task can be automatically converted into an R object for an R task at the next stage of a pipeline.
Romanesco defines a specification that prescribes a loose coupling between a task
and its runtime inputs and outputs. That specification is described in the API documentation
section. This specification is language-independent and instances of the spec are best
represented by a hierarchical data format such as JSON or YAML, or an equivalent
serializable type such as a
dict in python.
Romanesco is designed to be easily extended to new languages and environments, or to support new data types and formats, or modes of data transfer. This is accomplished via its plugin system, which is described in Plugins.
- Types and formats
- API documentation
- Developer documentation