Tasker is a lightweight Python framework for organizing and automating scientific data analysis. If you have lengthy computations involving several steps, over data in many directories, Tasker will help you
- Store your results to disk.
- Build up a complex analysis out of many intermediate steps, saving the results of earlier steps so you don't have to redo them.
- Only re-compute when necessary — when you change an input file or delete an intermediate result, only the computations that depended on it will be redone.
- Write code that works on data in many similar directories.
- Run said code in parallel.
Check out the quick (yet rather comprehensive) demonstration notebook included with the software.
Unit tests may be run with py.test.
pip install http://github.com/nkeim/tasker/zipball/master
If you've already installed Tasker and wish to upgrade:
pip install --upgrade http://github.com/nkeim/tasker/zipball/master
Written by Nathan Keim, while at the Penn Complex Fluids Lab.
Tasker is named after Tasker Street in Philadelphia, Pennsylvania.