configlite 0.2.5

Lightweight self-healing config manager

ConfigLite

A lightweight self-healing config handler.

What is ConfigLite?

ConfigLite aims to provide a simple "subclass and go" methodology for handling yaml-based file configs.

Subclass from the base BaseConfig object and add your defaults. Then, create an instance of your new class, providing a list of paths to search through. If a config is found at once of these paths, it is used, otherwise a default config is created.

You can then set import this from wherever is needed and access properties.

For more details, see Usage below.

Installation

pypi

ConfigLite is available on pypi!

Simply run pip install configlite to install it in your current environment.

development

If you want to make changes to how configs are handled, you can clone this repo and run pip install -e . to install it in editable mode.

[!note] Find a bug? Feel free to make an issue or pull request!

Usage

Creating a Config

There are two methods of creating a config object.

Direct Initialisation

The simplest method to create a config, simply init the BaseConfig class and provide your defaults as an argument:

from configlite import BaseConfig


CONFIG = BaseConfig(
    path = "config.yaml",
    defaults = {
        "value": 10,
        ...
    }
)

[!tip] Here, you may pass a single path, or a list of paths. The config will search through each path until it finds a match (See the section on Paths below for more info).

As a Class

If you wish to add extra methods, create a subclass of the base BaseConfig object, adding your parameters and their defaults to a defaults field.

For example:

from configlite import BaseConfig


class MyConfig(BaseConfig):
    defaults = {
        "value": 10,
        "name": "test",
        "pi": 3.14,
    }

You must still initialise the config, however:

CONFIG = MyConfig(path="config.yaml")

Access

To use your created config, there are two methods:

• "globally", where a single instance created.
• "locally", initialising the class where needed.

Global

To create a global config, set up the config as a parameter in the toplevel __init__.py:

CONFIG = MyConfig("./path/to/config.yaml")

Then in the rest of your code you may import this object:

from my_package import CONFIG

value = CONFIG.value

This is most useful if your code requires a single config file for everything.

Local

Or a local config, where you create an instance of your subclass wherever it is needed:

from my_package import MyConfig

config = MyConfig("./path/to/config.yaml")

value = config.value

This can be useful if you are juggling multiple different config files dynamically.

[!tip] This method is only recommended for subclasses, as each instance requires the same arguments.

Paths

The BaseConfig object can take either a single path, or a list of paths.

If a list is passed, the config will search each one in order, using the last one in the list if none are found.

Path Lists

Lets say you create a config this way:

class Config(BaseConfig):
    ...

CONFIG = Config(
    paths=[
        "config.yaml",
        "~/.config/app/config.yaml",
    ]
)

In this case, Config will take these steps:

1. Search for config.yaml in the current working directory.
2. If not found, search for ~/.config/app/config.yaml.
3. If still no file exists, create a default config at ~/.config/app/config.yaml

If we get to step 3, but then create a new config at ./config.yaml, this will take priority over the one found at ~/.config/app/config.yaml.

Environment Overrides

You may also set a prefix property, which enables environment overrides.

If we take our config from earlier, but add this feature:

class MyConfig(BaseConfig):
    defaults = {
        "value": 10,
        "name": "test",
        "pi": 3.14,
    }
    prefix = "test_prefix"

Now, on initialisation, the config will search for environment variables starting with test_prefix

If we export test_prefix_value="foo", we will see the changes:

cfg = MyConfig(...)

cfg.value
> "foo"

[!important] This is case-sensitive.

Example

See usage.py for an example.