Struct OpenAIClientState

Source

pub struct OpenAIClientState {
    pub prompt: VecDeque<Message>,
    pub client: OpenAIClient,
    pub entry_limit: Option<u64>,
}

Expand description

Represents a client state with a prompt history.

Fields§

§prompt: VecDeque<Message>

Conversation history messages.

§client: OpenAIClient

Reference to the OpenAIClient.

§entry_limit: Option<u64>

Implementations§

Source §

impl<'a> OpenAIClientState

Source

pub async fn add(&mut self, messages: Vec<Message>) -> &mut Self

Add messages to the conversation prompt.

§Arguments

messages - A vector of messages to add.

§Returns

A mutable reference to self.

Source

pub async fn add_last(&mut self, messages: Vec<Message>) -> &mut Self

Source

pub async fn set_entry_limit(&mut self, limit: u64) -> &mut Self

Set the maximum number of entries in the conversation prompt.

§Arguments

limit - The maximum number of entries.

§Returns

A mutable reference to self.

Source

pub async fn clear(&mut self) -> &mut Self

Clear all messages from the conversation prompt.

§Returns

A mutable reference to self.

Source

pub async fn last(&mut self) -> Option<&Message>

Retrieve the last message in the prompt.

§Returns

An Option containing a reference to the last Message.

Source

pub async fn generate( &mut self, model: Option<&ModelConfig>, ) -> Result<GenerateResponse, ClientError>

Generate an AI response.

This method sends the prompt to the API and, upon successful response, adds the assistant’s message to the prompt.

§Arguments

model - The model configuration.

§Returns

An APIResult with the API response or a ClientError.

Source

pub async fn generate_can_use_tool<F>( &mut self, model: Option<&ModelConfig>, show_call: Option<F>, ) -> Result<GenerateResponse, ClientError>
where F: Fn(&str, &Value),

Generate an AI response, possibly calling a tool.

If the API response includes a function call, it will run the corresponding tool.

§Arguments

model - The model configuration.
show_call - Optional callback function to show the tool call.(eg, show_call("tool_name", "args"))

§Returns

An APIResult with the API response or a ClientError.

Source

pub async fn generate_use_tool<F>( &mut self, model: Option<&ModelConfig>, show_call: Option<F>, ) -> Result<GenerateResponse, ClientError>
where F: Fn(&str, &Value),

Generate an AI response while forcing the use of a specific tool.

If the response includes a function call, the specified tool will be executed

§Arguments

model - The model configuration.
tool_name - The name of the tool to use.
show_call - Optional callback function to show the tool call.(eg, show_call("tool_name", "args"))

§Returns

An APIResult with the API response or a ClientError.

Source

pub async fn generate_with_tool<F>( &mut self, model: Option<&ModelConfig>, tool_name: &str, show_call: Option<F>, ) -> Result<GenerateResponse, ClientError>
where F: Fn(&str, &Value),

Generate an AI response while forcing the use of a specific tool.

If the response includes a function call, the specified tool will be executed.

§Arguments

model - The model configuration.
tool_name - The name of the tool to use.
show_call - Optional callback function to show the tool call.(eg, show_call("tool_name", "args"))