The purpose of this project is to create a generalized calculator engine for use in Python workshops teaching how to use git, pull requests, and other open-source contribution processes.

This calculator uses Reverse-Polish Notation to express functions.

Instantiate the calculator, then apply operations by calling push() and execute(). The only built-in functions are noop (no-operation), + (add), and / (floating-point division); additional desired functionality can be registered as needed (see Extending the Calculator).

import rpncalculator

calc = rpncalculator.Engine()
calc.push(1)
calc.execute('noop')
calc.push(2)
calc.push(3)
calc.execute('+')  # returns 5

To get a bare-bones calculator with no registered functions other than noop, pass auto_register=False to the constructor:

import rpncalculator

calc = rpncalculator.Engine(auto_register=False)
calc.execute('noop')  # succeeds
calc.execute('+')     # fails, function is not defined

You might need to install setuptools in your environment:

pip install setuptools

Once setuptools is installed, all testing is handled through setup.py.

The easiest way to run the test suite is using setup.py:

python setup.py test

If new tests need to be added to the project, put them in test/<module_name>.py and add from <module_name> import * to test/__init__.py.

Additional functions can be added to the calculator using the register(name, function) function:

def add(engine):
    engine.push(engine.pop() + engine.pop())

calc.register('+', add)

Registered functions must have the signature fn(engine) or the call to register will fail. An attempt to register the same function name twice will throw an error. The function may perform any sequence of calls to push, pop, and execute; however, to be RPN-correct it should only pop exactly the number of times needed to retrieve the necessary values and then push once (performing necessary algorithmic steps in-between).

Instead of interacting programmatically with the calculator, you can also supply input as a string or as an open file object:

import rpncalculator

# Note: this example assumes the '+' operation has been registered
# as previously demonstrated.

parser = rpncalculator.Parser()
result = parser.process('1 1 +')    # result = 2

with open(myinputfile) as f:
    result = parser.process(f)

Files can be arbitrarily large; the tokenization engine is smart about only tokenizing small chunks of the input at a time.