This is a lightweight tool I built for organizing experiment output in python (e.g. data analysis & general machine learning work). Particularly useful for work on a remote server to avoid log/local/ipython notebooks.
This will make a timestamped and named directory for all your figures and redirect stdout log output to a file.
Example:
import numpy as np
from analysis_wrapper import AnalysisContext as Experiment
import matplotlib.pyplot as plt
with Experiment("plot_respect") as ex:
# Log output will get saved even if your stuff crashes
print("Doing some work")
fig = plt.figure()
# Do some plotting
ex.save_fig(fig, "My Figure")