Managing and Documenting Legacy Scientific Workflows

Scientific legacy workflows are often developed over many years, poorly documented and implemented with scripting languages. In the context of our cross-disciplinary projects we face the problem of maintaining such scientific workflows. This paper presents the Workflow Instrumentation for Structure...

Full description

Bibliographic Details
Main Authors: Acuña Ruben, Chomilier Jacques, Lacroix Zoé
Format: Article
Language:English
Published: De Gruyter 2015-09-01
Series:Journal of Integrative Bioinformatics
Online Access:https://doi.org/10.1515/jib-2015-277
Description
Summary:Scientific legacy workflows are often developed over many years, poorly documented and implemented with scripting languages. In the context of our cross-disciplinary projects we face the problem of maintaining such scientific workflows. This paper presents the Workflow Instrumentation for Structure Extraction (WISE) method used to process several ad-hoc legacy workflows written in Python and automatically produce their workflow structural skeleton. Unlike many existing methods, WISE does not assume input workflows to be preprocessed in a known workflow formalism. It is also able to identify and analyze calls to external tools. We present the method and report its results on several scientific workflows.
ISSN:1613-4516