Try the balanced binary search tree interview problem!

Balanced Binary Search Tree

Given a sorted list of numbers `a`, write a function `balanced_bst(a)` to create a balanced binary search tree. A balanced [Binary Search Tree](https://en.wikipedia.org/wiki/Binary_search_tree) has no more than one level of depth difference between the right and left sides of the tree. 

Each value in the list `a` should correspond to a node value. The return value of `balanced_bst()` will be the root node of the balanced tree. An empty array passed to `balanced_bst()` should return `None`.

For example, given a list `a = [1,2,3,4,5,6,7,8]`, you want to create a balanced tree that may resemble the following:
```pseudo
       5
     /   \
    3     7
   / \   / \
  2   4 6   8
 /
1
```
The above figure represents a balanced tree because there is at most 1 level of difference between the depths of each side of the tree.

For this challenge you are given the class `TreeNode` with the members:
- `value`: the node value
- `left`: the left child node; defaults to `None`
- `right`: the right child node; defaults to `None`

The `__str__()` function is also implemented in the class `TreeNode`, so at any time you can print the root node to see a basic representation of the tree.

This challenge and variations of it were reported to have been asked at interviews with Google. If you've covered the material in [Pass the Technical Interview with Python](https://www.codecademy.com/learn/paths/pass-the-technical-interview-with-python) or an equivalent, you should be able to solve this challenge. If you have trouble, try refreshing your knowledge there first.
