runthis 0.0.1

Micro-utility for staying organized running experiments

runthis

Tired of your data science experiments becoming a mess? Don't want to pay $20,000 a year for a data science platform? Here's your poor man's version.

Usage

Name your experiment files like this:

 my_function?n=5&d=cat&init=[0.2,0.2,0.2].py

Then your experiment file looks like this:

if name=='main': import os kwargs = parse_kwargs(file.split(os.path.sep)[-1]) my_experiment(**kwargs)

This way what you do in the experiment never gets out of sync with the code.

Int, Float

Just call the experiment file

my_function?n=int:5&d=cat&init=[float:0.2,float:0.2,float:0.2].py

instead

Example

See here