chat_server

chat_server(self, input, output, session, *, model=DEFAULT_MODEL, api_key=None, url=None, system_prompt=DEFAULT_SYSTEM_PROMPT, temperature=DEFAULT_TEMPERATURE, text_input_placeholder=None, button_label='Ask', throttle=DEFAULT_THROTTLE, query_preprocessor=None, answer_preprocessor=None, debug=False)

Server portion of chatstream Shiny module.

Parameters

Name	Type	Description	Default
model	OpenAiModel | Callable[[], OpenAiModel]	OpenAI model to use. Can be a string or a function that returns a string.	DEFAULT_MODEL
api_key	str | Callable[[], str] | None	OpenAI API key to use (optional). Can be a string or a function that returns a string, or None. If None, then it will use the OPENAI_API_KEY environment variable for the key.	None
url	str | Callable[[], str] | None	OpenAI API endpoint to use (optional). Can be a string or a function that returns a string, or None. If None, then it will use the default OpenAI API endpoint.	None
system_prompt	str | Callable[[], str]	System prompt to use. Can be a string or a function that returns a string.	DEFAULT_SYSTEM_PROMPT
temperature	float | Callable[[], float]	Temperature to use. Can be a float or a function that returns a float.	DEFAULT_TEMPERATURE
text_input_placeholder	str | Callable[[], str] | None	Placeholder teext to use for the text input. Can be a string or a function that returns a string, or None for no placeholder.	None
throttle	float | Callable[[], float]	Throttle interval to use for incoming streaming messages. Can be a float or a function that returns a float.	DEFAULT_THROTTLE
button_label	str | Callable[[], str]	Label to use for the button. Can be a string or a function that returns a string.	'Ask'
query_preprocessor	Callable[[str], str] | Callable[[str], Awaitable[str]] | None	Function that takes a string and returns a string. This is run on the user’s query before it is sent to the OpenAI API. Note that is run only on the most recent query; previous messages in the chat history are not run through this function.	None
answer_preprocessor	Callable[[str], ui.TagChild] | Callable[[str], Awaitable[ui.TagChild]] | None	Function that tags a string and returns a TagChild. This is run on the answer from the AI assistant before it is displayed in the chat UI. Note that is run on streaming data. As each piece of streaming data comes in, the entire accumulated string is run through this function.	None
debug	bool	Whether to print debugging infromation to the console.	False

Attributes

Name	Type	Description
session_messages	reactive.Value[tuple[ChatMessageEnriched, …]]	All of the user and assistant messages in the conversation.
hide_query_ui	reactive.Value[bool]	This can be set to True to hide the query UI.
streaming_chat_string_pieces	reactive.Value[tuple[str, …]]	This is the current streaming chat content from the AI assistant, in the form of

a tuple of strings, one string from each message. When not streaming, it is empty. |

Methods

Name	Description
ask	Programmatically ask a question.
reset	Reset the state of this chat_server. Should not be called while streaming.

ask

chat_server.ask(self, query, delay=1)

Programmatically ask a question.

reset

chat_server.reset(self)

Reset the state of this chat_server. Should not be called while streaming.