concoction 0.1.4

Library to inject configurations

Concoction

Concoction is a simple configuration injection library that allows you to easily inject configuration values into your Python applications. It supports both pydantic and dataclasses for defining configuration models. Unlike traditional configuration management tools, Concoction does not parse or layer configurations; it only works with a dictionary object as an input.

Inspiration

Concoction is inspired by the @ConfigurationProperties and @Value annotations from Spring Boot. These annotations allow for easy injection of configuration properties into Java applications, and Concoction aims to provide a similar experience for Python developers.

Key Features

• Flexibility with Plain Dictionaries: Simple and flexible configuration management using plain dictionaries.
• Injecting Whole Configuration Blocks: Inject entire configuration blocks for modular and decoupled settings.
• Injecting Individual Fields: Granular control with the ability to inject specific configuration fields.
• Merging Parent Configurations: Option to bind inherited configuration fields to parent configuration block

Installation

To install Concoction, run the following command:

pip install concoction

Usage

Here is an example of config.yaml file:

app:
  service:
    host: 0.0.0.0
    port: 8000

and injecting service block into our config

import yaml  # pip install pyyaml
from pydantic import BaseModel
from concoction import Configuration, set_global_config


# define config with injection

@Configuration("app.service")
class ServiceConfig(BaseModel):
    host: str
    port: int


# Load configuration from a YAML file
with open("config.yaml") as f:
    config = yaml.safe_load(f)

set_global_config(config)

# Create an instance of the configuration model
app_config = ServiceConfig()

or field-wise

from concoction.values.pydantic import Value


class ServiceConfig(BaseModel):
    host: str = Value("app.service.host")
    port: int = Value("app.service.port")

Contributing

Contributions are welcome! Please open an issue or submit a pull request on GitHub.