Quick Stats hack for LMorePythonTHW

zedshaw · September 24, 2017, 7:33pm

I needed to demonstrate how to track your defect rates (or anything you need) while you’re programming so I cranked out a little hack script that helps you do it. The script is here:

github.com

zedshaw/learn-more-python-the-hard-way-solutions/blob/master/stats.py

#!/usr/bin/env python

import numpy
import sys
from bokeh import plotting
import os

command = sys.argv[1]
if command != "init":
    name = sys.argv[2]
    args = sys.argv[3:]

SAVE_DIR = ".morepystats"


if command == "add":
    # add a new stat
    file_name = f"{SAVE_DIR}/{name}.npy"
    assert os.path.exists(SAVE_DIR), f"No save dir {SAVE_DIR}. In the right place?"

This file has been truncated. show original

You use it like this:

$ ./stats.py init
$ ./stats.py add tests 1
$ ./stats.py add tests 3
$ ./stats.py add tests 2
$ ./stats.py add tests 8
$ ./stats.py add tests 6
$ ./stats.py stats tests
---
mean    4.0
stddev  2.6076809620810595
sd+/-1  1.3923190379189405, 6.6076809620810595
sd+/-2  -1.215361924162119, 9.215361924162119
q       2.0 3.0 6.0
IQR     4.0
$ ./stats.py graph tests

Which should, in theory, open a browser with a nice run chart of your stat:

This is an html file in .morepytstats/tests.html and you can open it and work with it in the browser. This is all provided by the bokeh project and numpy.