Configuration Serde

A common use case is to describe the configuration for an application as dataclasses, then deserialize it from a JSON, YAMl or TOML file.

from __future__ import annotations
import databind.json
import dataclasses
import tomlib
from pathlib import Path

@dataclasses.dataclass
class ServerConfig:
    host: str
    port: int = 8080

@dataclasses.dataclass
class MainConfig:
    server: ServerConfig

    @staticmethod
    def load(path: Path | str) -> MainConfig:
        data = tomlib.loads(Path(path).read_text())
        return databind.json.load(data, MainConfig, filename=path)

config = MainConfig.load_toml("config.toml")

An example config TOML file that can be parsed with the above configuration:

[server]
host = "localhost"
port = 8080

Note that any extra keys that are not expected per the schema will raise a databind.core.converter.ConversionError.

Danger

Databind uses Python runtime type annotation introspection using the typeapi package. This requires that all type annotations that databind comes in contact with must be valid expressions in the current Python version, even if from __future__ import annotations is used.

This means if your code needs to be compatible with Python versions lower than 3.10 or 3.9 that you can not use the new type union syntax (a | b) or built-in generic aliases (such as t.List[int]) and need to continue to use typing.Union, typing.Optional and typing.List, etc.