Struct Message

pub struct Message {
    pub role: Role,
    pub content: Option<Content>,
    pub content_list: Option<Vec<Content>>,
    pub tool_calls: Option<Vec<ToolCall>>,
    pub refusal: Option<String>,
    pub annotations: Option<Vec<String>>,
}

Represents a message in an OpenAI conversation.

Messages are the core communication unit between users and OpenAI models. They can contain various types of content including text, images, tool calls, and metadata like refusals and annotations.

Content Types

A message can contain either:

• Single content (content field) - for simple text messages
• Multiple content items (content_list field) - for multi-modal messages

Fields

• role - The role of the message sender (User, Assistant, System, etc.)
• content - Optional single content item
• content_list - Optional list of content items for multi-modal messages
• tool_calls - Optional list of tool calls made by the assistant
• refusal - Optional refusal message if the model declined to respond
• annotations - Optional list of annotations or metadata

Examples

use openai_tools::common::message::{Message, Content};
use openai_tools::common::role::Role;

// Simple text message
let message = Message::from_string(Role::User, "Hello!");

// Multi-modal message with text and image
let contents = vec![
    Content::from_text("What's in this image?"),
    Content::from_image_url("https://example.com/image.png"),
];
let message = Message::from_message_array(Role::User, contents);

Fields

role: Role

The role of the message sender

content: Option<Content>

Optional single content item

content_list: Option<Vec<Content>>

Optional list of content items for multi-modal messages

tool_calls: Option<Vec<ToolCall>>

Optional list of tool calls made by the assistant

refusal: Option<String>

Optional refusal message if the model declined to respond

annotations: Option<Vec<String>>

Optional list of annotations or metadata

Implementations

impl Message

pub fn from_string<T: AsRef<str>>(role: Role, message: T) -> Self

Creates a new Message with a single text content.

This is a convenience method for creating simple text messages.

Arguments

• role - The role of the message sender
• message - The text content of the message

Returns

A new Message instance with the specified role and text content

Examples

use openai_tools::common::message::Message;
use openai_tools::common::role::Role;

let message = Message::from_string(Role::User, "Hello, how are you?");

pub fn from_message_array(role: Role, contents: Vec<Content>) -> Self

Creates a new Message with multiple content items.

This method is used for multi-modal messages that contain multiple types of content such as text and images.

Arguments

• role - The role of the message sender
• contents - Vector of content items to include in the message

Returns

A new Message instance with the specified role and content list

Examples

use openai_tools::common::message::{Message, Content};
use openai_tools::common::role::Role;

let contents = vec![
    Content::from_text("What's in this image?"),
    Content::from_image_url("https://example.com/image.png"),
];
let message = Message::from_message_array(Role::User, contents);

pub fn get_input_token_count(&self) -> usize

Calculates the approximate token count for the message content.

This method uses the tiktoken library to estimate the number of tokens that would be consumed by this message when sent to OpenAI’s API. Only text content is counted; images are not included in the calculation.

Returns

The estimated number of tokens for the text content in this message

Examples

use openai_tools::common::message::Message;
use openai_tools::common::role::Role;

let message = Message::from_string(Role::User, "Hello, world!");
let token_count = message.get_input_token_count();

Trait Implementations

impl Clone for Message

fn clone(&self) -> Message

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Message

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Message

Custom deserialization implementation for Message.

This implementation handles the deserialization of messages from OpenAI API responses, converting string content to Content objects and handling optional fields.

fn deserialize<D>(deserializer: D) -> Result<Message, D::Error>

where D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Message

Custom serialization implementation for Message.

This implementation ensures that messages are serialized correctly for the OpenAI API, handling the mutually exclusive nature of content and content_list fields. Either content or content_list must be present, but not both.

fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>

where S: Serializer,

Serialize this value into the given Serde serializer.