AssistantTurn

AssistantTurn(
    contents,
    *,
    tokens=None,
    finish_reason=None,
    completion=None,
    cost=None,
    partial_reason=None,
    **kwargs,
)

Assistant turn - represents model response with additional metadata

Parameters

Name	Type	Description	Default
contents	str | Sequence[Content | str]	A list of Content objects.	required
tokens	Optional[tuple[int, int, int] | list[int]]	A numeric vector of length 3 representing the number of input, output, and cached tokens (respectively) used in this turn.	None
finish_reason	Optional[FinishReason | str]	The (normalized) reason why the turn ended, e.g. "success", "tool_use", "max_tokens", "content_filter", "context_window", or "stop_sequence". Providers may also surface a reason chatlas doesn’t yet recognize, in which case the raw string is passed through unchanged.	None
completion	Optional[CompletionT]	The completion object returned by the provider. This is useful if there’s information returned by the provider that chatlas doesn’t otherwise expose.	None
cost	Optional[float]	The cost of this turn in USD. This is computed when the turn is created based on the token usage and pricing information (including service tier).	None
partial_reason	Optional[str]	If set, indicates this turn is incomplete (e.g., the stream was interrupted or cancelled). The value describes the reason for the partial state.	None

See Also

• Turn: The base class for all turn types.

Attributes

Name	Description
is_partial	Whether this turn is a partial (interrupted/cancelled) turn.

Methods

Name	Description
validate_tokens	Convert list to tuple for JSON deserialization compatibility.

validate_tokens

AssistantTurn.validate_tokens(v)

Convert list to tuple for JSON deserialization compatibility.