What a provider gets: the prompt, the tools, the response format and tool choice, the request's span and, for a conversation the provider keeps, the previous response's id and the prompt after it.
Attributes#
| Attribute | Type | Description |
|---|---|---|
prompt | P.Prompt | |
tools | tuple[AnyTool, ...] = () | |
response_format | TextFormat | JsonFormat = TextFormat() | |
tool_choice | ToolChoice = 'none' | |
span | Span | None = None | |
previous_response_id | str | None = None | |
incremental_prompt | P.Prompt | None = None | |
incremental_fallback | bool = False | |
tool_call_validation | ToolCallValidation = 'strict' |
Functions#
__init__#
__init__(prompt, tools=(), response_format=TextFormat(), tool_choice='none', span=None, previous_response_id=None, incremental_prompt=None, incremental_fallback=False, tool_call_validation='strict') -> None| Parameter | Type | Description |
|---|---|---|
prompt | P.Prompt | |
tools | tuple[AnyTool, ...] = () | |
response_format | TextFormat | JsonFormat = TextFormat() | |
tool_choice | ToolChoice = 'none' | |
span | Span | None = None | |
previous_response_id | str | None = None | |
incremental_prompt | P.Prompt | None = None | |
incremental_fallback | bool = False | |
tool_call_validation | ToolCallValidation = 'strict' |