Pysettings YAML

This library extends over the concepts of Decouple and Split Settings to create a library that is simple to use and expresses your settings in a yaml file for higher expressivity.

Installation

pip install pysettings-yaml

Usage

Define a YAML file with the definition of your settings, for example, a settings.yaml in the same path as your settings module:

settings:
  SAMPLE_SETTING_BOOL:  # Each entry inside the settings section is the name of your setting
    origins:
      - name: env
      - name: direct
        value: true
  SAMPLE_SETTING_STR:
    origins:
      - name: direct
        value: banana
      - name: env

Then you can use this file like this in a settings.py:

from split_settings.tools import optional
from pathlib import Path

from pysettings_yaml import get_config

BASE_DIR = Path(__file__).parent

setting_files = [
    BASE_DIR / "settings.yaml",
]


config = get_config(setting_files)

SAMPLE_SETTING_BOOL = config("SAMPLE_SETTING_BOOL", cast=bool)
SAMPLE_SETTING_STR = config("SAMPLE_SETTING_STR")

print(SAMPLE_SETTING_BOOL)
print(SAMPLE_SETTING_STR)  # This will print banana

By default the library supports two origin names: env (to receive the value of your setting using the decouple library) or direct (to provide a constant value), but you can add your own providers easily (refer to the "Adding custom providers" section).

Origin are evaluated in strict order. In those examples, SAMPLE_SETTING_BOOL will evaluate first the env origin and, if that value is None, it will process the next origin, which would be direct. In contrast, for SAMPLE_SETTING_STR, the first origin is the direct origin. Since direct can never return None, the env origin will never be evaluated.

Overriding settings

You can define several setting files to be processed in order. All of them will be merged together, so any subsequent setting file will override the settings with the same name defined in a previous file of the list. Very useful if you need to define different origins for your settings depending on your deployment environment:

from split_settings.tools import optional
from pathlib import Path
import os

from pysettings_yaml import get_config
ENVIRONMENT = os.environ.get("ENVIRONMENT", "dev")

BASE_DIR = Path(__file__).parent

setting_files = [
    BASE_DIR / "settings.yaml",
    BASE_DIR / f"settings.{ENVIRONMENT}.yaml",
    optional(BASE_DIR / "settings.nonexistent.yaml")
]


config = get_config(setting_files)

SAMPLE_SETTING_BOOL = config("SAMPLE_SETTING_BOOL", cast=bool)
SAMPLE_SETTING_STR = config("SAMPLE_SETTING_STR")

Any path wrapped with the optional function will not raise an exception should that file not exist in the path you specify.

Adding custom providers

If you need other providers (for example for getting your settings from a vault like Amazon SSM) you can implement your own providers like this:

from typing import Union, Optional

from pysettings_yaml.providers.interfaces import SettingsProvider, OriginModel, NoValue


class AWSModel(OriginModel):
    path: str
    decrypt: bool


class SampleAWSSettingsProvider(SettingsProvider):
    name = "aws"
    schema = AWSModel

    def get(
        self, setting_name: str, origin_data: AWSModel
    ) -> Union[Optional[str], NoValue]:
        return f"path: {origin_data.path}, decrypt: {origin_data.decrypt}"

Just make sure to:

• Provide the name and a pydantic schema
• Implement the get method

Then you can obtain your config providing your SettingProvider and use it in your yaml:

config = get_config(setting_files, additional_providers=[SampleAWSSettingsProvider()])

settings:
  SAMPLE_SETTING:
    origins:
      - name: aws
        path: some/path
        decrypt: true
      - name: env