Model hydration

load_as() loads configuration exactly like load(), then hydrates the result into a typed model instead of returning a loose dict — useful when you want validation, defaults, and IDE autocomplete instead of dot-access on an untyped mapping.

Pydantic

from pydantic import BaseModel
import yaconfiglib

class DBConfig(BaseModel):
    host: str
    port: int

db_settings = yaconfiglib.load_as(DBConfig, "config.yaml", loader="yaml")
print(db_settings.host)

Both Pydantic v1 (.parse_obj) and v2 (.model_validate) are supported automatically — load_as detects which API the model class exposes. Pydantic is a strictly optional dependency; it's only imported when load_as is actually called with a Pydantic model.

Dataclasses

from dataclasses import dataclass
import yaconfiglib

@dataclass
class DBConfig:
    host: str
    port: int = 5432

db_settings = yaconfiglib.load_as(DBConfig, "config.yaml")

Only keys matching the dataclass's declared fields are passed to the constructor — extra keys in the loaded document are silently ignored rather than raising a TypeError.

Plain classes

If model_cls is neither a Pydantic model nor a dataclass, load_as falls back to calling model_cls(**data) directly, so any class whose __init__ accepts the loaded document's keys as keyword arguments works without special integration.

Combining with other loader options

load_as accepts every keyword load() does — merging, interpolation, and multiple sources all work the same way:

settings = yaconfiglib.load_as(
    AppConfig,
    "base.yaml", "production.yaml",
    merge="deep",
    interpolate=True,
)

The loaded document must be a mapping (dict) — load_as raises TypeError if the merged/interpolated result isn't one.