Skip to content

Support for Pydantic settings configuration file management

License

Notifications You must be signed in to change notification settings

jordantshaw/pydantic-config

Repository files navigation

Pydantic Config

Support for Pydantic settings configuration file loading

Installation

Pydantic Config can be installed via pip:

pip install pydantic-config

Pydantic Config is also available on conda under the conda-forge channel:

conda install pydantic-config -c conda-forge

Optional Dependencies

Pydantic-Config has the following optional dependencies:

  • yaml - pip install pydantic-config[yaml]
  • toml - pip install pydantic-config[toml] Only for python<3.11

You can install all the optional dependencies with pip install pydantic-config[all]

Usage

# config.toml
app_name = "Python Application"
description = "Test application description"
from pydantic_config import SettingsModel, SettingsConfig


class Settings(SettingsModel):
    app_id: str = 1
    app_name: str = None
    description: str = None
    log_level: str = 'INFO'
    
    model_config = SettingsConfig(
        config_file='config.toml',
    )


settings = Settings()
print(settings)
# app_id='1' app_name='Python Application' description='Test application description' log_level='INFO'

Using multiple config files

Multiple config files can be loaded by passing a list of file names. Files will be loaded in the order they are listed. Meaning later files in the list will take priority over earlier files.

# config.toml
app_name = "Python Application"
description = "Test application description"
// config.json
{
  "description": "Description from JSON file",
  "log_level": "WARNING"
}
from pydantic_config import SettingsModel, SettingsConfig


class Settings(SettingsModel):
    app_id: str = 1
    app_name: str = 'App Name'
    description: str = None
    log_level: str = 'INFO'
    
    model_config = SettingsConfig(
        config_file=['config.toml', 'config.json']  # The config.json file will take priority over config.toml
    )

settings = Settings()
print(settings)
# app_id='1' app_name='Python Application' description='Description from JSON file' log_level='WARNING'

Supported file formats

Currently, the following file formats are supported:

  • .yaml Requires pyyaml package
  • .toml Requires tomli package for python<3.11
  • .json
  • .ini

Using dotenv files

pydantic-config supports using dotenv files because pydantic-settings natively supports dotenv files. To use a dotenv file in conjunction with the config files simply set env_file parameter in SettingsConfig. The values in the dotenv file will take precedence over the values in the config files.

class Settings(SettingsModel):
  app_name: str = None
  description: str = None
    
    model_config = SettingsConfig(
        env_file='.env',
        config_file='config.toml',
    )

Requiring config files to load

Config files will attempt to be loaded from the specified file path. By default, if no file is found the file will simply not be loaded (no error occurs). This may be useful if you want to specify config files that may or may not exist. For example, you may have different config files for per environment: config-prod.toml and config-dev.toml.

To disable this behavior set config_file_required=True. This will cause an error to be raised if the specified config file(s) do not exist. Setting this to True will also prohibit the config_file parameter from being set to None or empty [].

Merging

If your configurations have existing list or dict variables the contents will be merged by default. To disable this behavior and override the contents instead you can set the config_merge option to False in the settings Config class.

# config.toml
[foo]
item1 = "value1"
# config2.toml
[foo]
item2 = "value2"
from pydantic_config import SettingsModel, SettingsConfig


class Settings(SettingsModel):
    foo: dict = {}
    
    model_config = SettingsConfig(
        config_file=['config.toml', 'config2.toml'],
        config_merge= True,
    )


settings = Settings()
print(settings)
# foo={'item1': 'value1', 'item2': 'value2'}

# If config_merge=False then config2.toml would override the values from config.toml
# foo={'item2': 'value2'}

Duplicate items in merged lists

By default, all list items will be merged into a single list regardless of duplicated items. To only keep unique list items, set config_merge_unique=True. This will only keep unique items in within a list.