Crab AI

The Crab AI library provides convenient access to the OpenAI REST API from any Rust application. This library aims to follow the implementation of the official OpenAI SDKs for Python and Node.js as closely as possible. However, it is unofficial and not maintained by OpenAI.

Note: This project is still under development, should not be used in production, and many aspects of the structure may change.

Documentation

The REST API documentation can be found on platform.openai.com. The full API of this library can be found in docs.rs.

Installation

Add this to your Cargo.toml:

[dependencies]
crab_ai = "0.1.9"

Usage

The full API of this library can be found in the documentation.

use crab_ai::{OpenAI, ClientOptions};
use crab_ai::resources::chat::{ChatCompletionContent::{Multiple, Text},
    ChatCompletionContentPart::Image, ChatCompletionCreateParams, Detail,
    ChatCompletionMessageParam::{Assistant, System, User}, ImageURL,
};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let openai = OpenAI::new(ClientOptions::new())?;

    let completion = openai.chat.completions.create(ChatCompletionCreateParams {
        messages: vec![System{ content: "You are a helpful assistant.", name: None },
            User{ content: Text("Who won the world series in 2020?"), name: None },
            Assistant{ content: Some("The Los Angeles Dodgers won the World Series in 2020."), name: None, tool_calls: None },
            User{ content: Text("Where was it played?"), name: None }],
        model: "gpt-4o-mini",
        ..Default::default()
    }).await?;

    println!("{:?}", completion);
    Ok(())
}

While you can provide an api_key directly, we recommend using environment variables to keep your API key secure.

Examples

Refer to the examples directory for usage examples.

OpenAI Assistant Beta

The Crab AI library includes support for the OpenAI Assistant API, which is currently in beta. This feature allows you to create and manage threads, messages, and runs with the Assistant. The Assistant API is designed to help build conversational agents that can interact with users in a more dynamic and context-aware manner.

Example: Creating a Thread and a Message

Below is an example demonstrating how to create a thread and a message within that thread using the Assistant API:

let thread = openai.beta.threads.create(ThreadCreateParams::default()).await?;

let message = openai.beta.threads.messages.create(
    &thread.id,
    MessageCreateParams {
        role: message_create_params::Role::User,
        content: message_create_params::Content::Text("I need to solve the equation `3x + 11 = 14`. Can you help me?".to_string()),
        ..Default::default()
    },
    None,
).await?;

This example showcases how to initialize a conversation with a thread and add a message to it.

Creating a Run and Polling for Completion

Here's an example of how to create a run using the Assistant API and poll until it reaches a terminal state:

let run = openai.beta.threads.runs.create_and_poll(
    &thread.id,
    RunCreateParams {
        assistant_id: "asst_ABcDEFgH12345678910xZZZz".to_string(),
        instructions: Some("Please address the user as Jane Doe. The user has a premium account.".to_string()),
        ..Default::default()
    },
    None
).await?;

if run.status == RunStatus::Completed {
    let messages = openai.beta.threads.messages.list(&run.thread_id, None, None).await?;

    for message in messages.data.iter().rev() {
        match &message.content.first().unwrap() {
            messages::MessageContent::Text { text } => {
                println!("{:?} > {:?}", message.role, text.value);
            }
            _ => {}
        }
    }
}

By integrating these features, the Crab AI library provides a robust interface for utilizing the latest capabilities of the OpenAI API, including the Assistant API currently in beta.