hobson

#!/bin/sh

# Manually patch and compile bash for systems like FC16 that don't have shellshock protection 
# availalbe through their package managers.

# sudo yum install patch byacc texinfo bison autoconf gettext ncurses-devel

VERSION1=4
VERSION2=2
VERSION_PATCH=52

hobson

#!/bin/sh
SHORTCUT="[Desktop Entry]
Name=Sublime Text 2
Comment=Edit text files
Exec=/usr/local/sublime-text-2/sublime_text
Icon=/usr/local/sublime-text-2/Icon/128x128/sublime_text.png
Terminal=false
Type=Application
Encoding=UTF-8
Categories=Utility;TextEditor;"

hobson

def walk_level(path, level=1):
    """Like os.walk, but takes `level` kwarg that indicates how deep the recursion will go.

    Notes:
      TODO: refactor `level`->`depth`

    References:
      http://stackoverflow.com/a/234329/623735

    Args:

hobson

"""Attempt to confirm results in paper by Miller and Sanjurjo 
["Confused by the Gambler's and Hot Hand Falacy?..."](http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2627354)
Calculate the probability of a streak of length 2 in a coin flip 
for a small sequence of attempts ("finite sequence of iid Bernoulli trials").
The fencepost counting issue means you have to ignore one of the flips.
Perhaps this is what causes Miller and Sanjurjo to get 0.4 instead of 0.5.
"the conditional relative frequency... is 0.4 not 0.5"

"""
from random import randint

hobson

hobson

r"""Faulty positive semidefinite matrix generator

Using a normally distributed random matrix it's pretty
rare that you'll get a semidefinite matrix from the product,
even for low dimensions. For 3 dimensions you have about a 0.001%% yield

>>> pos_semi_matrices = generate_positive_semidefinite_matrices(num_samples=1000)
  6.7% were positive semidefinite and not positive definite
 45.5% were positive definite
 52.2% were positive semidefinite

hobson

# Install python and some binary-dependencies from source in the user directory
# useful for Django webapps on shared hosts like Dreamhost

# BerkelyDB v 4.8 is supported by python 2.7 but not 5.3
# In addition, the bsddb module was deprecated in Python 2.6 and will be removed in 3.0
# So all this BDB stuff isn't useful, but does no harm
# linuxfromscratch.com has patch that can make python 2.7 work with BerkelyDB, if you need it
BDBVER=5.3.21
BDBVE=5.3
export LDFLAGS="-L$HOME/local/lib -L$HOME/local/BerkeleyDB.$BDBVE/lib"

hobson

# install pip & virtualenv
# create a virtual environment for your webapp
# install django in the new virtualenv

cd ~/tmp
PYVE=2.7
PYVER=2.7.3
STVER=0.6c11
# go directly to the source to get the latest and greatest
wget http://pypi.python.org/packages/$PYVE/s/setuptools/setuptools-$STVER-py$PYVE.egg#md5=fe1f997bc722265116870bc7919059ea

hobson

_autorebase                     000013-1
_update-info-dir                00236-1
alternatives                    1.3.30c-10
autobuild                       5.3-1
autoconf                        13-1
autoconf2.1                     2.13-12
autoconf2.5                     2.69-2
automake                        9-1
automake1.13                    1.13.4-1
automake1.14                    1.14-1

hobson

""" White balance (color balance)

Adjust colors to flatten color histogram peaks and broaden color spectrum for better color contrast.
This is also sometimes called white balancing or probability distribution whitening.

References:
  - Color balance algorithm in [OpenCV](http://www.morethantechnical.com/2015/01/14/simplest-color-balance-with-opencv-wcode/)
  - Ported to python by [DavidYKay](https://github.com/DavidYKay)
  - Optimized (`O(N^2)` -> `O(N)`) by @alxrsngartn funded in part by NSF grant number 1722399 to [Aira](http://github.com/aira) (@jmeyers-aira)
  - Incorporated into [nlpia](github.com/aira/nlpia) by @hobson