Small repository that contains experiments around settings management in python.
Loading settings from a file. Configuration is defined in config.py and loaded in any other Python module you wish
by simply importing settings from config.
settings is a Dynaconf object but act as a normal Python dictionary.
from dynaconf import Dynaconf
settings = Dynaconf(
settings_files=["settings.yaml"], # load all files in this list
folder="config", # defines the folder from where to load the files
core_loaders=["yaml"], # defines the loader to use for the files ; Dynaconf support a lot of loaders
envvar_prefix="MYAPP_", # defines the prefix for the environment variables loaded
load_dotenv=True, # load the .env file
merge_enabled=True # merge all the loaded settings together
)Dynamic variables in settings files.
postgresql:
user: postgres
# password will be loaded from an environment variable called MYAPP_POSTGRES_PASSWORD
# usually, you inject those at deployment time
password: "@format {env[MYAPP_POSTGRES_PASSWORD]}"
host: localhost
port: 5432
# database variable will be set from an already defined value: postgresql.user
database: "@format {this.postgresql.user}"Validation of settings is also possible with Dynaconf but I prefer personally to use dataclasses or pydantic
to define validated data classes.
See the core implementation for more details.
Here's a list of references I used to implement this project: