testkraut – test and evaluate heterogeneous data processing pipelines

This is a framework for software testing. That being said, testkraut tries to minimize the overlap with the scopes of unit testing, regression testing, and continuous integration testing. Instead, it aims to complement these kinds of testing, and is able to re-use them, or can be integrated with them.

In a nutshell testkraut helps to facilitate statistical analysis of test results. In particular, it focuses on two main scenarios:

  • Comparing results of a single (test) implementation across different or changing computational environments (think: different operating systems, different hardware, or the same machine before an after a software upgrade).
  • Comparing results of different (test) implementations generating similar output from identical input (think: performance of various signal detection algorithms).

While such things can be done using other available tools as well, testkraut aims to provide a lightweight, yet comprehensive description of a test run. Such a description allows for decoupling test result generation and analysis – opening up the opportunity to “crowd-source” software testing efforts, and aggregate results beyond the scope of a single project, lab, company, or site.

Package availability chart
Distribution Base version Our version Architectures
Debian GNU/Linux 7.0 (wheezy)   0.0.1-1~nd70+1 i386, amd64, sparc, armel
Debian GNU/Linux 8.0 (jessie)   0.0.1-1~nd70+1 i386, amd64, sparc, armel
Debian unstable (sid)   0.0.1-1~nd+1 i386, amd64, sparc, armel
Ubuntu 12.04 LTS “Precise Pangolin” (precise)   0.0.1-1~nd12.04+1 i386, amd64, sparc
Ubuntu 14.04 “Trusty Tahr” (trusty)   0.0.1-1~nd12.10+1+nd13.04+1+nd13.10+1+nd14.04+1 i386, amd64, sparc

Comments

blog comments powered by Disqus