ReproMatch

Tool Details

Sumatra

Sumatra is a Python library and tool that manages and tracks projects by capturing and maintaining information about each run of the experiment. Sumatra requires knowledge of the programming language to get library dependencies. In addition, users need to configure the location of input and output files, and for experiments that use or support Python scripting, there is an API that includes capture within the code based on user-defined annotations.

Website:
http://neuralensemble.org/sumatra/
Source code:
https://github.com/open-research/sumatra
Publication:
http://dx.doi.org/10.1109/MCSE.2012.41
Tags:
programming tool software package Python Matlab OS Support OS X: True Windows: False Linux: True Web-based: False Provenance Capture: True OS-Based: False Code-Based: True Workflow-Based: False Data-Based: False Representation: False Replicability: True Modifiability: True Portability: Low Note: Sumatra does not offer any particular support for making computations portable: experiments are only guaranteed to be repeated in the original environment. Longevity: False Document linkage: True By Reference: True Inline: False Note: Sumatra provides document linkage by reference by including both a LaTeX package and Sphinx directives that enable embedding output data files (e.g., images) from previously-captured experiments into documents, with verification of file contents using SHA1 digests (i.e., a form of cryptographically secure hashing) and hyperlinks to a full record containing the captured provenance. Experiment sharing: False