/adk-style

Pydantic Patterns

ADK models use Pydantic v2.

Basic Model Structure

• Use Field() for validation, defaults, and descriptions.
• Use PrivateAttr() for internal state that must not be serialized.
• Use model_post_init() for setup logic, not __init__ — overriding __init__ on a Pydantic model bypasses validation ordering.
• Use model_dump() / model_dump_json(), not the v1 dict() / json().

Which mechanism to use

Field() with Constraints

Declare bounds on the field rather than writing a validator for them — it keeps the rule next to the data and shows up in the generated JSON schema.

compaction_interval: Optional[int] = Field(default=None, gt=0)
injected_latency_seconds: float = Field(default=0.0, le=120.0)

Documenting fields

The house style is an attribute docstring directly under the field, which Sphinx picks up:

model: Union[str, BaseLlm] = ''
"""The model to use for the agent.

When not set, the agent inherits the model from its ancestor.
"""

Those docstrings are documentation only. To also make them the field descriptions in the generated JSON schema, set use_attribute_docstrings=True in the model's ConfigDict; SerializedBaseModel already enables it.

class MyModel(BaseModel):
  model_config = ConfigDict(use_attribute_docstrings=True)

  field_name: str
  """Description of the field."""

On-Wire Models

A model that crosses a network or storage boundary — an API payload, a WebSocket message, a persisted event — should inherit from SerializedBaseModel in google.adk.utils._serialized_base_model rather than BaseModel. It sets alias_generator=to_camel with populate_by_name=True and defaults model_dump_json() to by_alias=True, so Python stays snake_case while the wire format stays camelCase without every call site remembering to pass by_alias.

field_validator — Single-Field Validation

Use @field_validator when a constraint needs logic that Field() cannot express.

@field_validator('max_llm_calls')
@classmethod
def validate_max_llm_calls(cls, value: int) -> int:
  if value <= 0:
    raise ValueError('max_llm_calls must be positive.')
  return value

Rules:

• Add @classmethod under the decorator. Pydantic v2 applies it implicitly, but ADK writes it out — every validator in the codebase does.
• Return the (possibly transformed) value.
• Raise ValueError with a message that names the field and the bound.
• The default mode is 'after', which runs post-coercion and is what you almost always want; omit the argument. Pass mode='before' only to intercept raw input.

model_validator — Cross-Field and Migration Validation

mode='before' — deprecation and field migration

Receives the raw input, usually a dict, before any field is parsed. Use it to rename or back-fill fields.

@model_validator(mode='before')
@classmethod
def check_for_deprecated_save_live_audio(cls, data: Any) -> Any:
  """If save_live_audio is passed, use it to set save_live_blob."""
  if isinstance(data, dict) and 'save_live_audio' in data:
    warnings.warn(
        'The `save_live_audio` config is deprecated, use `save_live_blob`.',
        DeprecationWarning,
        stacklevel=2,
    )
    if data['save_live_audio']:
      data['save_live_blob'] = True
  return data

Guard with isinstance(data, dict): the input can also arrive as an already constructed model instance, and indexing that raises.

mode='after' — cross-field consistency

Receives the constructed instance and must return it.

@model_validator(mode='after')
def _validate_parallel_worker_config(self) -> Node:
  if self.max_parallel_workers is not None and not self.parallel_worker:
    raise ValueError(
        'max_parallel_workers can only be set when parallel_worker is True.'
    )
  return self