193. Caching with functools

The functools module in Python provides several utilities to help with function decorators, and one of the most useful features is caching, which helps optimize the performance of functions by storing previously computed results. The functools.cache functions (lru_cache, cache_property, cache_on_self) are specifically designed for this purpose.

Here are 10 Python code snippets that demonstrate how to use caching with functools to improve the performance of various functions:

---

1. Basic Caching with lru_cache

from functools import lru_cache

@lru_cache(maxsize=None)  # No limit on the cache size
def slow_function(x):
    print("Computing...")
    return x * x

# Usage
print(slow_function(5))  # Computes and caches the result
print(slow_function(5))  # Retrieves the cached result

---

2. Caching with maxsize Parameter

from functools import lru_cache

@lru_cache(maxsize=3)  # Cache the last 3 calls
def slow_function(x):
    print("Computing...")
    return x * x

# Usage
print(slow_function(1))  # Computes and caches the result
print(slow_function(2))  # Computes and caches the result
print(slow_function(3))  # Computes and caches the result
print(slow_function(4))  # Computes and caches the result, evicts oldest (1)
print(slow_function(1))  # Computes again, as 1 was evicted

---

3. Clearing Cache Manually

from functools import lru_cache

@lru_cache(maxsize=3)
def slow_function(x):
    print("Computing...")
    return x * x

# Usage
print(slow_function(2))  # Computes and caches the result
slow_function.cache_clear()  # Clears the cache manually
print(slow_function(2))  # Computes again after cache is cleared

---

4. Using cache_property for Cached Properties

from functools import cached_property

class MyClass:
    @cached_property
    def expensive_computation(self):
        print("Computing expensive result...")
        return 42

# Usage
obj = MyClass()
print(obj.expensive_computation)  # Computes and caches the result
print(obj.expensive_computation)  # Retrieves the cached result

---

5. Using Cache with Mutable Arguments

from functools import lru_cache

@lru_cache(maxsize=2)
def slow_function(a, b):
    print("Computing...")
    return a + b

# Usage
print(slow_function(1, 2))  # Computes and caches the result
print(slow_function(1, 2))  # Retrieves the cached result
print(slow_function(2, 3))  # Computes and caches the result

---

6. Caching with Functions that Have Default Arguments

from functools import lru_cache

@lru_cache(maxsize=None)
def slow_function(a, b=2):
    print("Computing...")
    return a * b

# Usage
print(slow_function(3))  # Computes with default b=2
print(slow_function(3, 4))  # Computes with b=4
print(slow_function(3))  # Retrieves the cached result for a=3, b=2

---

7. Using cache_on_self for Method Caching

from functools import cache_on_self

class MyClass:
    @cache_on_self
    def method(self, x):
        print("Computing...")
        return x * 2

# Usage
obj = MyClass()
print(obj.method(5))  # Computes and caches the result
print(obj.method(5))  # Retrieves the cached result

---

8. Performance Boost with Cache for Recursion

from functools import lru_cache

@lru_cache(maxsize=None)
def fibonacci(n):
    if n <= 1:
        return n
    return fibonacci(n-1) + fibonacci(n-2)

# Usage
print(fibonacci(30))  # Fast computation with caching

---

9. LRU Cache with Custom Key Function

from functools import lru_cache

def custom_key(a, b):
    return a + b

@lru_cache(maxsize=None, key=custom_key)
def slow_function(a, b):
    print("Computing...")
    return a * b

# Usage
print(slow_function(2, 3))  # Computes and caches

---

10. Handling Cache Misses with Default Values

from functools import lru_cache

@lru_cache(maxsize=None)
def slow_function(a):
    print("Computing...")
    return a * 2

# Usage
print(slow_function(10))  # Computes and caches
print(slow_function(20))  # Computes and caches
print(slow_function(10))  # Retrieves the cached result for 10

---

In these snippets:

1. lru_cache(maxsize) is used for caching results of functions with positional and keyword arguments.

2. cached_property caches the result of a method call as a property, avoiding recomputation.

3. The cache size can be controlled with the maxsize parameter, and the cache can be cleared manually using cache_clear().

4. Caching significantly boosts performance when dealing with expensive or recursive computations by avoiding redundant calculations.