Nested models

Being able to use Pydantic models to parse nested configurations on our YAML files is maybe the strongest point of DriConfig.

Let's say we have a YAML config.yaml file looking like this:

# config.yaml

timeout: 1000
min_date: 2021-04-17
model_parameters:
  alpha: 2
  beta: 0.1
  gamma: 30

In this case, model_parameters is what we call a nested configuration, being a dictionary of dictionaries.

Then, we would have two options:

• Parse model_parameters as a dict type.
• Parse it as a Pydantic model, and performing specific type validation for each of its values.

Arbitrary types

Let's parse the model_parameters configuration as an arbitrary dictionary.

from datetime import date
from typing import Any, Dict

from driconfig import DriConfig, DriConfigConfigDict


class AppConfig(DriConfig):
    """Configuration class to parse the config.yaml file contents."""

    """Configure AppConfig to point at the config.yaml file."""
    model_config = DriConfigConfigDict(
        config_folder=".",
        config_file_name="config.yaml",
    )

    timeout: int
    min_date: date
    model_parameters: Dict[str, Any]

app_config = AppConfig()
print(app_config.model_dump_json(indent=4))
"""
{
    "timeout": 1000,
    "min_date": "2021-04-17",
    "model_parameters": {
        "alpha": 2,
        "beta": 0.1,
        "gamma": 30
    }
}
"""
print(type(app_config.model_parameters))
"""
<class 'dict'>
"""

Note that no type validation is performed on the values of model_paramters. Instead, it is stored as an arbitrary dictionary.

Nested model

Now, let's properly parse model_parameters as a Pydantic model.

from datetime import date

from pydantic import BaseModel

from driconfig import DriConfig, DriConfigConfigDict


class ModelParameters(BaseModel):
    """Pydantic model for `model_parameters`."""

    alpha: int
    beta: float
    gamma: int


class AppConfig(DriConfig):
    """Configuration class to parse the config.yaml file contents."""

    """Configure AppConfig to point at the config.yaml file."""
    model_config = DriConfigConfigDict(
        config_folder=".",
        config_file_name="config.yaml",
    )

    timeout: int
    min_date: date
    model_parameters: ModelParameters

app_config = AppConfig()
print(app_config.model_dump_json(indent=4))
"""
{
    "timeout": 1000,
    "min_date": "2021-04-17",
    "model_parameters": {
        "alpha": 2,
        "beta": 0.1,
        "gamma": 30
    }
}
"""
print(type(app_config.model_parameters))
"""
<class '__main__.ModelParameters'>
"""

Note how, in this case, type validation is performed for alpha, beta and gamma.

We could mimic this pattern to build even deeper nested configuration parsers, holding all of Pydantic's validation power.