Write a calculator library.

Requirements:

• Handle binary operations, e.g. 5 ­ 6, 8 * 4
• The calculator should be able to handle a full arithmetic expression at once, for example: 1 + 1 ­ 4 * 4
• The library must take Order of Operations into account
• No need to handle parentheses, e.g. (1 + 1 ­ 4) * 4
• Must be extendable so that it accepts user­defined operands

Objectives:

• OOP
• Best practices
• Algorithms

Need to convert infix notation into a tree or find another alternative. Dijkstra's shunting yard method can be used here; it might be shorter to convert to RPN instead and process a stack than build and traverse a tree.

From an OOP perspective, we might consider operators to be tokenized into a subclass of a general "Operator", each having specific attributes regarding symbol and precedence/weight, etc. From a practical perspective, this might be overkill.

For simplicity's sake, let's assume that we'll only accept positive integers.

In addition to listed requirements, library should be self-contained for ease of demonstration/review.

Choosing Python to solve problem since it is great for modeling concepts. The runtime environment also comes with a built-in unit test library, making it easy to show that the project works (and doesn't work) under various scenarios.

Assumes Python 2.7 is installed and available in the PATH. Assumes GNU Make is installed.

Run make to execute tests/demos.

To try using the library interactively, run python2.7 from this directory and try this:

>>> from infix import eval_expr
>>> eval_expr('4 + 4 * 8 / 2 + 10')
30