blog - Thoughts On Python

Intro

This is a listing of things I like and dislike about Python:

packages & dependencies

pydantic - A library for data validation and serialization FastAPI - Powered by Pydantic and offers an easy-to-use framework for building APIs

syntax

Likes: - Context Managers (e.g. with some_resource() as r: ) - A language construct for setup and shutdown, very handy for abstractions that have a lot of setup and need to tidy up after use - comprehensions

Idioms & Patterns

An idiomatic way to represent in-memory caching objects in Python is to use the class obj pattern. Since all classes are imported once per run-time by the import system in Python, you can design your caching logic like so:

class Cache():
    _cache = None
    def __init__(self, ):
        print("cache_constructed")
    

    @classmethod
    def populate_cache(cls, *args, **kwargs):
        """doc-string"""
        if not cls._cache:
            print("")
            cls._cache = [i for i in range(0, 20, 2)]
        else:
            print("already populated")
    
    @classmethod
    def get_nums(cls, ) -> Generator:
        """
            returns nums cache
        """
        yield from cls._cache

then when you import this class in one module, all subsequent imports are all reference the same cache object:

#module_a
from cache import Cache


Cache.populate_cache()
print(list(Cache.get_nums()))


class A():
    def __init__(self) -> None:
        pass

#module_b
from cache import Cache


Cache.populate_cache()

class B():
    def __init__(self) -> None:
        pass

then when you import both modules in module_c, Cache was already imported in module_a:

#module_c
from module_a import A
from module_b import B

print("module C")

when you run module_c, you get this output:

path/to/dir$ uv run module_c.py

[0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
already populated
[0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
module C

References