postfixcalc

Simple, stupid but easy and powerful infix expression evaluator using postfix notation.

User Guide

How to use?

from postfixcalc import Calc

expr = "(-1) ^ 2 + 3 - 4 * 8 - 2 ^ 3"
calc = Calc(expr)

print(calc.answer)
print(type(calc.answer))

-36
<class 'int'>

Explanation

expression: (-1) ^ 2 + 3 - 4 * 8 - 2 ^ 3

which with the math operator precedence is:

expression: ((-1) ^ 2) + 3 - (4 * 8) - (2 ^ 3)
            = (1) + 3 - (32) - (8)
            = 4 - 32 - 8
            = -28 - 8
            = -36

Initialization

calc = Calc(
    '(2 ^ 32) ^ (2 ^ 15) + -1',
    calc_timeout=1,         # timeout for the calculation (in seconds)
    str_repr_timeout=1.5    # timeout to generate the string representation (in seconds)
)

print(f"'(2 ^ 32) ^ (2 ^ 15) + -1's answer has '{len(calc.stranswer())}' digits")

'(2 ^ 32) ^ (2 ^ 15) + -1's answer has '315653' digits

print(f'answer is: {calc.stranswer(15, 15)}')

answer is: 674114012549907...068940335579135

Other Attributes

from rich.pretty import Pretty
from rich import print as rprint

rprint(calc.parsed)

<ast.BinOp object at 0x7fb65c3917b0>

rprint(calc.extracted)

[
    (
        [
            (
                [([2], <ast.Pow object at 0x7fb661991120>, [32])],
                <ast.Pow object at 0x7fb661991120>,
                [([2], <ast.Pow object at 0x7fb661991120>, [15])]
            )
        ],
        <ast.Add object at 0x7fb661990ee0>,
        [(<ast.USub object at 0x7fb661991540>, [1])]
    )
]

rprint(calc.flattened)

(
    (
        (([2], <ast.Pow object at 0x7fb661991120>, [32]),),
        <ast.Pow object at 0x7fb661991120>,
        (([2], <ast.Pow object at 0x7fb661991120>, [15]),)
    ),
    <ast.Add object at 0x7fb661990ee0>,
    (<ast.USub object at 0x7fb661991540>, [1])
)

rprint(calc.strparenthesized)

(((2 ^ 32)) ^ ((2 ^ 15))) + (-1)

rprint(calc.listparenthesized)

['(', '(', '(', 2, '^', 32, ')', ')', '^', '(', '(', 2, '^', 15, ')', ')', ')', '+', '(', '-1', ')']

rprint(calc.numerized)

['(', '(', '(', 2, '^', 32, ')', ')', '^', '(', '(', 2, '^', 15, ')', ')', ')', '+', '(', -1, ')']

rprint(calc.postfix)

[2, 32, '^', 2, 15, '^', '^', -1, '+']

rprint(f'{calc.stranswer(15, 15)}')

674114012549907...068940335579135

Calc Documentation

class Calc (expr: str, calc_timeout: int | float = 0.1, str_repr_timeout: int | float = 0.2)

• expr: infix math expression

• calc_timeout: the timeout of the math calculation, if a expression's calculation took longer than this time, it would be killed and a TimeoutError will be raised.

• str_repr_timeout: Calculating a expression like: (2 ^ 32) ^ (2 ^ 15) takes about a seconds, but the result has 315653 digits; so printing (getting the string representation) takes some time, this timeout is controlled by this parameter; and a TimeoutError will be raised.

Properties of Calc

Important Note

All the properties of Calc type (except stranswer) are cached properties. It means that they are calculated once and stored in the object and are not calculated every time you need them (if other properties and attributes are remained unchanged)

• Calc.parsed

Parse the object with ast.parse function and return the parsed expression. Because ast.parse uses the grammar of Python any syntax error will be raised here!

# underlying function

from postfixcalc.parser import parse

expr = '-1 ^ 2'
parsed = parse(expr)

rprint(parsed)

<ast.UnaryOp object at 0x7fb6603566b0>

• Calc.extracted

Return a list of extracted numbers and operators from the parsed object

# underlying function
from postfixcalc.parser import extract_nums_and_ops

extracted = extract_nums_and_ops(parsed)
rprint(extracted)

[(<ast.USub object at 0x7fb661991540>, [([1], <ast.Pow object at 0x7fb661991120>, [2])])]

• Calc.flattened

Flatten the numbers and operators list, this will reduce the nested lists and tuples

# underlying function
from postfixcalc.parser import flatten_nodes

flattened = flatten_nodes(extracted)
rprint(flattened)

(<ast.USub object at 0x7fb661991540>, ([1], <ast.Pow object at 0x7fb661991120>, [2]))

• Calc.strparenthesized

Generate a parenthesized version of the expression passed in initialization according to the math operator precedence

# underlying function
from postfixcalc.parser import restrexpression

rprint(restrexpression(flattened))

-(1 ^ 2)

• Calc.listparenthesized

Return the digits and parenthesis and operators in a list that will be used to generate the postfix list

# underlying function
from postfixcalc.parser import relistexpression

listed = relistexpression(flattened)
rprint(listed)

['-', '(', 1, '^', 2, ')']

• Calc.numerized

Numerize the string numbers returned by listparenthesized method. In some cased like: (-1) ^ 2) the listparenthesized looks like this: ['(', '-1', ')', '^', 2], so we have to make those strings numbers

# underlying function
from postfixcalc.parser import make_num

numerized = make_num(listed)
rprint(numerized)

['-', '(', 1, '^', 2, ')']

• Calc.postfix

Return a list with the postfix notation of the expression

# underlying function
from postfixcalc.parser import infix_to_postfix

postfixed = infix_to_postfix(numerized)
rprint(postfixed)

[1, 2, '^', '-']

• Calc.answer

Calculate the answer respecting the calc_timeout

IMPORTANT NOTE: DON'T CALL print ON THE RESULT OF THIS METHOD.

This is because for instance calculating (2 ^ 32) ^ (2 ^ 15) is done under calc_timeout BUT generating the string representation WILL TAKE MUCH LONGER!!! (it has 315653 digits)

If you want to print the result, use stranswer method

method

@cached_property
def answer(self):
    process = multiprocessing.Process(target=evaluate, args=(self.postfix,))
    process.start()
    process.join(timeout=self.calc_timeout)
    if process.is_alive():
        process.terminate()
        raise TimeoutError(
            f"Calculations of {self.strparenthesized!r} took longer than {self.calc_timeout} seconds",
        ) from None
    return evaluate(self.postfix)

# underlying function
from postfixcalc.pyeval import evaluate

rprint(evaluate(postfixed))

-1

• Calc.stranswer

Return the string representation of the Calc.answer respecting the str_repr_timeout

method

def stranswer(
    self,
    beginning: Optional[int] = None,
    ending: Optional[int] = None,
) -> str:
    """Return the string representation of the answer with the respect to the `str_repr_timeout`

    beginning: see the first n digits of the answer: '982734...'
    ending: see the last n digits of the answer: '...982734'

    if both are specified: '987234...873242'
    """
    process = multiprocessing.Process(target=str, args=(self.answer,))
    process.start()
    process.join(self.str_repr_timeout)
    if process.is_alive():
        process.terminate()
        raise TimeoutError(
            f"Generating a string representation of {self.expr!r} took longer than {self.str_repr_timeout} seconds",
        ) from None
    try:
        answer = str(self.answer)
        match (beginning, ending):
            case (None, None):
                return answer
            case (x, None):
                return f"{answer[:x]}..."
            case (None, x):
                return f"...{answer[-x:]}"
            case (x, y):
                return f"{answer[:x]}...{answer[-y:]}"
            case _:
                raise ValueError("Confusing beginning and ending")
    except ValueError:
        raise TimeoutError(
            f"Generating a string representation of {self.expr!r} took longer than {self.str_repr_timeout} seconds",
        ) from None

How the answer is calculated

Calc type has cached properties which rely on each others returned values, it means that when we access the Calc.answer property this sequential method and function calls, happen:

print(calc.answer)

1. calc.answer needs the postfix: calc.postfix gets called
2. calc.postfix needs the numerized list of the digits and operators: calc.numerized gets called
3. calc.numerized needs the parenthesized list of the digits and operators: calc.listparenthsized gets called
4. calc.listprenthesized needs the flattened list of extracted nums and ops: calc.flattened gets called.
5. calc.flattened needs the extracted nums and ops: calc.extracted gets called.
6. calc.extracted needs the parsed expression of the input expression: calc.parsed gets called.

Call graph for getting the answer