easyconfig 0.1.1

Easy application configuration with yaml files

easyconfig

Easy application configuration with yaml files

Description

Easyconfig simplifies the configuration management for small application.

Validation and parsing of the configuration file is done through pydantic and easyconfig builds on that. It's possible to use all pydantic features and model features so every exotic use case should be covered. If you have previously worked with pydantic you should feel right at home

Why not pydantic settings

A pydantic settings object is a non mutable object. With easyconfig you can create a global configuration and just import it into your modules. When your application starts you can read the configuration e.g. from a settings file and the object will change the values accordingly.

Usage

Instead of using BaseModel from pydantic there are different Models which have to be used. They just add additional behaviour to make the object mutable.

from easyconfig import ConfigModel     # Use this instead of BaseModel
from easyconfig import PathModel       # Use this for path configuration
from easyconfig import AppConfigModel  # Use this as a topmost Model

Simple example

from easyconfig import AppConfigModel


class MyAppSimpleConfig(AppConfigModel):
    retries: int = 5
    url: str = 'localhost'
    port: int = 443


# Create a global variable which then can be used throughout your code
CONFIG = MyAppSimpleConfig()

# Use with type hints and auto complete
print(CONFIG.port)

# Load configuration file from disk.
# If the file does not exist it will be created
# Loading will also change all values of CONFIG accordingly
CONFIG.load_file('/my/configuration/file.yml')

Created configuration file:

retries: 5
url: localhost
port: 443

Nested example

from pydantic import Field
from easyconfig import AppConfigModel, ConfigModel


class HttpConfig(ConfigModel):
    retries: int = 5
    url: str = 'localhost'
    port: int = 443


class MyAppSimpleConfig(AppConfigModel):
    run_at: int = Field(12, alias='run at')  # use alias to load from/create a different key
    http = HttpConfig()


CONFIG = MyAppSimpleConfig()
CONFIG.load_file('/my/configuration/file.yml')

Created configuration file:

run at: 12
http:
  retries: 5
  url: localhost
  port: 443

Comments

It's possible to specify a description through the pydantic Field. The description will be created as a comment in the .yml file

from pydantic import Field
from easyconfig import AppConfigModel


class MyAppSimpleConfig(AppConfigModel):
    retries: int = Field(5, description='Amount of retries on error')
    url: str = 'localhost'
    port: int = 443


CONFIG = MyAppSimpleConfig()
CONFIG.load_file('/my/configuration/file.yml')

Created configuration file:

retries: 5  # Amount of retries on error
url: localhost
port: 443

Callbacks

It's possible to register callbacks that will get executed when a value changes or when the configuration gets loaded for the first time. A useful feature if the application allows dynamic reloading of the configuration file (e.g. through a file watcher).

from easyconfig import AppConfigModel


class MyAppSimpleConfig(AppConfigModel):
    retries: int = 5
    url: str = 'localhost'
    port: int = 443


def setup_http():
    create_my_http_client(CONFIG.url, CONFIG.port)


CONFIG = MyAppSimpleConfig()
CONFIG.load_file('/my/configuration/file.yml')

# setup_http will be automatically called if a value changes in the MyAppSimpleConfig
# during a subsequent call to CONFIG.load_file() or when the config gets loaded for the first time
sub = CONFIG.subscribe_for_changes(setup_http)

# It's possible to cancel the subscription again
sub.cancel()

Model documentation

ConfigModel

from easyconfig import ConfigModel
from pydantic.fields import PrivateAttr

class MyConfigModel(ConfigModel):
    a = 10

    # Use PrivateAttr to declare variable which are not part of the configuration
    my_var: int = PrivateAttr()

    # implement this to do something when all values have been set
    def on_all_values_set(self):
        print('all values have been set')
        self.my_var = self.a * self.a

PathModel

PathModel inherits from ConfigModel so everything that works for ConfigModel also works for PathModel

from pathlib import Path
from easyconfig import PathModel, AppConfigModel


# Path objects will be automatically resolved to absolute paths
class PathContainer(PathModel):
    folder = Path('folder_path')
    exec = Path('./rel_path')


class MyAppSimpleConfig(AppConfigModel):
    folders = PathContainer()


CONFIG = MyAppSimpleConfig()
CONFIG.load_file('/my/configuration/file.yml')

CONFIG.folders.exec  # <-- /my/configuration/rel_path

# base path can be set per PathModel and is valid for all sub items
CONFIG.folders.set_file_path('/other/path')
CONFIG.load_file()

CONFIG.folders.exec  # <-- /other/path/rel_path

AppConfigModel

AppConfigModel inherits from PathModel so everything that works for PathModel also works for AppConfigModel

from easyconfig import AppConfigModel


class MyAppSimpleConfig(AppConfigModel):
    retries: int = 5
    url: str = 'localhost'
    port: int = 443


CONFIG = MyAppSimpleConfig()

# It's possible to set the file path without loading
CONFIG.set_file_path('/my/configuration/file.yml')

# Path can be omitted if it was passed once in load_file or set_file_path
CONFIG.load_file()

# It's also possible to load from a python dictionary
CONFIG.load_dict({'my': 'dict'})

Changelog

0.1.1 (26.02.2022)

• Fixed an issue with dynamic defaults
• Optional values with default None will not be created in the yaml file

0.1.0 (10.01.2022)

• Updated requirements

0.0.2 (16.09.2021)

• Validate user defaults
• Use json representation of values to get native yaml data types
• Use enum values instead of enum types

0.0.1 (14.09.2021)

• Initial release