Valeriy Pogrebitskiy's Quick-and-dirty Python test solutions

Python Test
===========

1) 	Write a function that, given a list of numbers, returns its arithmetic 
	average and its standard deviation.
   
	math_test.py
	============ 
	compute() function computes both numeric list's average and
	standard deviation - as a dictionary containing 'avg' and 'std' key/value pairs.
	This function computes standard deviation using following formula:
   
	std = sqrt(sum((x - mean)**2) / (n-1))  
   
	(Ref: http://www.ltcconline.net/greenl/courses/201/descstat/mean.htm)
   
	It uses generator expression (it can as well be using list comprehension, 
	but for large lists generator expressions save memory)
        

2) 	Write a quick-sort for lists. What can you tell about its complexity?

	qsort_test.py
	=============
	Implements quicksort using data comprehensions and randomized pivot.
	In general, use of randomized pivot is not necessary. In fact, it
	slows down performance when list is mostly unsorted, but... it significantly
	improves performance when list is (mostly) sorted. Performance loss
	for unsorted lists is very low - compared to performance gain for sorted lists
	
	IMPORTANT: since we perform data.pop() when taking a pivot, qsort() modifies
	original data. To prevent this, I am using @preservedList decorator - that
	executes qsort() but keeps original data unchanged 

3) 	How would you implement binary trees in Python? Write a function that visits a 
	binary tree in pre-order.
	
	btree_test.py
	=============
	This module implements two classes: Node and BTree. I have not provided *all*
	of the usual BTree methods - only the ones important for this particular assignment:
	insert() and preOrder(). preOrder() traverses all nodes of the BTree and prints
	value of a Node itself followed by traversing left sub-tree, then right sub-tree
	
	
Each of the 3 modules contains simple "doctest" tests.

To run tests, execute:

	python math_test.py -v
	python qsort_test.py -v
	python btree_test.py -v 

I could have provided "unittest" script(s) instead, but for given level of 
complexity - "doctest" works fine. Keep in mind that I have very good "unittest" experience ...