Last updated:
0 purchases
pyflexconfig 1.0.2
pyflexconfig
You prefer using direct Python files to provide configuration data to your Python application,
just because you can provide any object type you'd prefer in your configuration file.
You want to place your custom configuration file anywhere you'd prefer, in traditional places like
/etc or wherever you prefer.
You want your app to provide some default values to some or all configuration data.
But you may prefer something else than Python files (YAML, JSON, XML, etc.), just provide your own
parser.
pyflexconfig is just another kid in the block in the settings / options / configuration data
management game that fulfills these requirements.
Features
Provide the default configuration programmatically, with a default file in your package or through
environment variables.
Override the default configuration options with a custom configuration file. You can override one,
two or all default configuration options.
You don't like Python configuration files for onr reason or another, even when executed in a
sandbox. Okay, provide your own parser.
The configuration container is just a global object on which you can add attributes. I recommend
using a types.SimpleNamespace object but anything may do the job.
You may provide the custom configuration file either through conventional places
(ie. /etc/my_custom_config.py), explicitly through a dedicated command line option, or through an
environment variable.
The default filter keeps only UPPERCASED options not starting with an underscore. This allows to
use temporary variables with lowercase names in your configuration files, that are not exposed in
the resulting configuration object. You may provide your own filter if you want other rules or add
other rules to the default one.
You may finally provide a validator if you need to check the type / value of some or all
configuration options.
Installation
Using pyflexconfig
pyflexconfig requires Python 3.6 or later. Install it with the usual pip dance :
pip install pyflexconfig
Note that pyflexconfig runs on any operating system and does not require packages outside the
stdlib.
Develop on a fork of pyflexconfig
Of course, provide and activate a dedicated virtual environment with python 3.6 or later, fork
pyflexconfig, then:
git clone <your fork URL>
cd pyflexconfig
pip install -e .[dev]
Please work from the develop branch that supposed to include the latest validated developments.
Please rebase your fork on the develop branch and fix the unit tests (run pytest) before issuing
a pull / merge request on the develop branch of the original Git repository.
Ah ! And of course file an issue that explains your changes.
Using pyflexconfig
A basic usage :
Your default config file defaultsettings.py:
# Warning: do not import here somethong that's not in the stdlib
ONE = 1
TWO = 2
# ...
The settings.py module in the same directory:
import pathlib
from types import SimpleNamespace
from pyflexconfig import bootstrap
config = SimpleNamespace(
# "hardcoded" default options
THREE = 3,
# ...
)
# Load the default config that ships in the package
_default_config_path = pathlib.Path(__file__).resolve().parent / "defaultsettings.py"
bootstrap(config, default_path=_default_config_path)
And your main application:
from .settings import config
def main():
print(config.ONE)
print(config.TWO)
print(config.THREE)
if __name__ == "__main__":
main()
For more examples, please browse the
demo/ directory of the repository
as well as the tests.py module.
View the full API issuing in the console:
python -c "import pyflexconfig; help(pyflexconfig)"
License
This software is provided under the terms of the MIT license. See the LICENSE file in the Git
repository or here.
Author
This software comes from Gilles Lenfant's damaged brain.
The development of pyflexconfig is sponsored by the Caisse des
Depots.
Links
Github
Issues tracker
History
1.0.2 - 2020/7/5
Bugfix: Accepting pathlib.Path objects as well as str paths.
1.0.1 - 2020/6/19
Removed python-dotenv from requirements (not used)
Some typos in README.md
History file
1.0.0 - 2020/6/18
initial release
For personal and professional use. You cannot resell or redistribute these repositories in their original state.
There are no reviews.