Cachetools is a Python module which provides various memoizing collections and decorators. It also includes variants from the functools’ @lru_cache decorator. To use it, first, we need to install it using pip.
pip install cachetools
Cachetools provides us five main function.
- cached
- LRUCache
- TTLCache
- LFUCache
- RRCache
Let’s look at each of the following functions in detail and with examples.
Cached
cached is used as a decorator. When we call cache, it will cache the function for later use. This will by default perform a simple cache.
Syntax:
@cached(cache = {})
def some_fun():
pass
Example: Let’s see it using an example. We are going to use the time module to see the efficiency of our module.
from cachetools import cached import time # Without cached def fib(n): return n if n<2 else fib(n-1) + fib(n-2) s = time.time() print(old_fib(35)) print("Time Taken: ", time.time() - s) # Now using cached s = time.time() # Use this decorator to enable caching @cached(cache ={}) def fib(n): return n if n<2 else fib(n-1) + fib(n-2) print(fib(35)) print("Time Taken(cached): ", time.time() - s) |
Output:
9227465 Time Taken: 4.553245782852173 9227465 Time Taken(cached): 0.0003821849822998047
LRUCache
LRUCache is used inside the cached decorator. LRU cache means “Least Recently Used” cache. It takes a parameter “maxsize” which states that how recent functions should be cached.
Syntax:
@cached(cache= LRUCache(maxsize= 3))
def some_fun():
pass
Example:
from cachetools import cached, LRUCache import time # cache using LRUCache @cached(cache = LRUCache(maxsize = 3)) def myfun(n): # This delay resembles some task s = time.time() time.sleep(n) print("\nTime Taken: ", time.time() - s) return (f"I am executed: {n}") # Takes 3 seconds print(myfun(3)) # Takes no time print(myfun(3)) # Takes 2 seconds print(myfun(2)) # Takes 1 second print(myfun(1)) # Takes 4 seconds print(myfun(4)) # Takes no time print(myfun(1)) # Takes 3 seconds because maxsize = 3 # and the 3 recent used functions had 1, # 2 and 4. print(myfun(3)) |
Output:
Time Taken: 3.0030977725982666 I am executed: 3 I am executed: 3 Time Taken: 2.002072334289551 I am executed: 2 Time Taken: 1.001115083694458 I am executed: 1 Time Taken: 4.001702070236206 I am executed: 4 I am executed: 1 Time Taken: 3.0030171871185303 I am executed: 3
Note: LRUCache can also be called from the standard Python package – functools. It can see imported as
from functools import lru_cache
@lru_cache
def myfunc():
pass
TTLCache
TTLCache or “Time To Live” cache is the third function that is included in cachetools module. It takes two parameters – “maxsize” and “TTL”. The use of “maxsize” is the same as LRUCache but here the value of “TTL” states for how long the cache should be stored. The value is in seconds.
Syntax:
@cached(cache= TTLCache(maxsize= 33, ttl = 600))
def some_fun():
pass
Example:
from cachetools import cached, TTLCache import time # Here recent 32 functions # will we stored for 1 minutes @cached(cache = TTLCache(maxsize = 32, ttl = 60)) def myfun(n): # This delay resembles some task s = time.time() time.sleep(n) print("\nTime Taken: ", time.time() - s) return (f"I am executed: {n}") print(myfun(3)) print(myfun(3)) time.sleep(61) print(myfun(3)) |
Output:
Time Taken: 3.0031025409698486 I am executed: 3 I am executed: 3 Time Taken: 3.0029332637786865 I am executed: 3
LFUCache
LFUCache or “Least Frequently Used” cache is another type of caching technique that retrieves how often an item is called. It discards the items which are called least often to make space when necessary. It takes one parameter – “maxsize” which is the same as in LRUCache.
Syntax:
@cached(cache= LFUCache(maxsize= 33))
def some_fun():
pass
Example:
from cachetools import cached, LFUCache import time # Here if a particular item is not called # within 5 successive call of the function, # it will be discarded @cached(cache = LFUCache(maxsize = 5)) def myfun(n): # This delay resembles some task s = time.time() time.sleep(n) print("\nTime Taken: ", time.time() - s) return (f"I am executed: {n}") print(myfun(3)) print(myfun(3)) print(myfun(2)) print(myfun(4)) print(myfun(1)) print(myfun(1)) print(myfun(3)) print(myfun(3)) print(myfun(4)) |
Output:
Time Taken: 3.002413272857666 I am executed: 3 I am executed: 3 Time Taken: 2.002107620239258 I am executed: 2 Time Taken: 4.003819465637207 I am executed: 4 Time Taken: 1.0010886192321777 I am executed: 1 I am executed: 1 I am executed: 3 I am executed: 3 I am executed: 4
RRCache
RRCache or “Random Replacement” cache is another type of caching technique that randomly chooses items in the cache and discards them to free up space when necessary. It takes one parameter – “maxsize” which is the same as in LRUCache. It also has a parameter choice which is by default set to “random.choice”.
Syntax:
@cached(cache= RRCache(maxsize= 33))
def some_fun():
pass
Example:
from cachetools import cached, RRCache import time # Here if a particular item is not called # within 5 successive call of the function, # it will be discarded @cached(cache = RRCache(maxsize = 5)) def myfun(n): # This delay resembles some task s = time.time() time.sleep(n) print("\nTime Taken: ", time.time() - s) return (f"I am executed: {n}") print(myfun(3)) print(myfun(3)) print(myfun(2)) print(myfun(4)) print(myfun(1)) print(myfun(1)) print(myfun(3)) print(myfun(2)) print(myfun(3)) |
Output:
Time Taken: 3.003124713897705 I am executed: 3 I am executed: 3 Time Taken: 2.0021231174468994 I am executed: 2 Time Taken: 4.004120588302612 I am executed: 4 Time Taken: 1.0011250972747803 I am executed: 1 I am executed: 1 I am executed: 3 I am executed: 2 I am executed: 3
