yapeco: Yet Another Python Envvar Config Object

A positively miniscule utility module to access .env (and otherwise environment-defined) config values in a structured way through a Python object. Created out of annoyance with not being able to autocomplete said envvars in my editor :^)

Features & Limitations

• Case-insensitive + snake-case (i.e. SNAKE_case) field names
• Primitives such as str, bool, int and float are supported (no guarantees with float though, because, well... floating point)
• Assuming use of the above primitives, supports Optional[*] types (and by extension Union[*,None]), but no others from typing
• Default values through class variable assignment; assumed to be None for optional types
• Will (intentionally) raise a RuntimeError if there is no value set and no default value
• Common boolean config formats (i.e. VAR=0/1/true/false/True/False) work as expected

Usage

Installation (PyPI):

pip3 install yapeco

A simple example:

# -- [mock environment] ---------------------------
API_KEY=abc123
DELAY_MSEC=18
FEATURE_A_ENABLED=false
FEATURE_B_ENABLED=1

# --- contrived_config.py -------------------------

from yapeco import BaseEnvironment as Env

class Config(Env):
    api_key: str
    delay_msec: int
    feature_a_enabled: bool
    feature_a_flags: str = "-a -b -c"
    feature_b_enabled: bool
    feature_b_flags: Optional[str]


# --- main.py -------------------------------------

from config import Config

Config.api_key # "abc123"
Config.delay_msec # 18
Config.feature_a_enabled # False
Config.feature_a_flags # "-a -b -c"
Config.feature_b_enabled # True
Config.feature_b_flags # None

# ...
# API_KEY=def456
# FEATURE_B_ENABLED=false

Config.refresh() # update environment

Config.api_key # "def456"
Config.feature_b_enabled # False

Development

Requires poetry.

poetry install  # install dependencies
poetry build  # build package
poetry run pytest .  # run tests
poetry run pyright .  # run type checks

Feature Backburner

• Support implied types (type conversion currently depends on cls.__annotations__)

Extra

Pedantic note:

As in Smalltalk, classes themselves are objects.

—The Literal Python Documentation