Cache function

January 13, 2024

---

The functools module provides a @cache decorator to store the results of an expensive function; also known as memoize.

No function cache

This example calls a function several times. The function sleeps for 2 seconds to represent an expensive task being performed.

import time


def coeffs_one(method):
    """
    Get the a and b coefficients for a given method.
    """
    time.sleep(2)

    if method == "yaws":
        a, b = 1, 2
    elif method == "graham":
        a, b = 2, 4
    else:
        a, b = 0, 0

    return a, b


for _ in range(3):
    ti = time.perf_counter()
    a, b = coeffs_one("graham")
    tf = time.perf_counter()
    print(f"a is {a} | b is {b} | Elapsed {tf - ti:.4f} s")

The output is shown below. Notice the elapsed time for each run is 2 seconds.

a is 2 | b is 4 | Elapsed 2.0036 s
a is 2 | b is 4 | Elapsed 2.0032 s
a is 2 | b is 4 | Elapsed 2.0044 s

Using function cache

This example calls a function several times and uses the @cache decorator to store the results.

import time
from functools import cache


@cache
def coeffs_two(method):
    """
    Get the a and b coefficients for a given method.
    """
    time.sleep(2)

    if method == "yaws":
        a, b = 1, 2
    elif method == "graham":
        a, b = 2, 4
    else:
        a, b = 0, 0

    return a, b


for _ in range(3):
    ti = time.perf_counter()
    a, b = coeffs_two("graham")
    tf = time.perf_counter()
    print(f"a is {a} | b is {b} | Elapsed {tf - ti:.4f} s")

The output for this example is shown below. Notice the elapsed time for the second and third runs is zero because the function results are cached.

a is 2 | b is 4 | Elapsed 2.0001 s
a is 2 | b is 4 | Elapsed 0.0000 s
a is 2 | b is 4 | Elapsed 0.0000 s