https://github.com/akr4/openai-api

Rust client for OpenAI API supporting streaming mode of the chat completion endpoint.
https://github.com/akr4/openai-api

openai-api rust

Last synced: about 1 month ago
JSON representation

Rust client for OpenAI API supporting streaming mode of the chat completion endpoint.

• Host: GitHub
• URL: https://github.com/akr4/openai-api
• Owner: akr4
• License: mit
• Created: 2023-03-27T00:08:07.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-05-14T10:58:36.000Z (6 months ago)
• Last Synced: 2024-05-14T11:56:47.412Z (6 months ago)
• Topics: openai-api, rust
• Language: Rust
• Homepage:
• Size: 13.7 KB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# openai-api

Rust client for OpenAI API supporting streaming mode of the chat completion endpoint.

## Supported Endpoints

- Completion: post https://api.openai.com/v1/completions

- Chat Completion: post https://api.openai.com/v1/chat/completions (with streaming mode support)

- Embeddings: post https://api.openai.com/v1/embeddings

## Getting Started

Add the following to your `Cargo.toml`:

```toml:file=Cargo.toml

openai-api = { git = "https://github.com/akr4/openai-api.git" }

```

Put this in your crate root:

```rust

use std::env;

use openai_api::chat;

#[tokio::main]

async fn main() {

    // create request

    let request = chat::CompletionRequest {

        model: chat::Model::Gpt35Turbo,

        temperature: Some(1.0),

        messages: vec![

            chat::Message {

                role: chat::MessageRole::User,

                content: "Hello".to_string(),

            }

        ],

    };

    // call completion endpoint

    let response = chat::completion(

        env::var("OPENAI_API_KEY").expect("environment variable OPENAI_API_KEY is not found."),

        &request).await.unwrap();

    // show response text

    println!("{}", response.choices[0].message.content);

}

```

### Streaming Mode

```rust

use std::env;

use futures_util::StreamExt;

use openai_api::{chat, chat_stream};

#[tokio::main]

async fn main() {

    // create request

    let request = chat::CompletionRequest {

        model: chat::Model::Gpt35Turbo,

        temperature: Some(1.0),

        messages: vec![

            chat::Message {

                role: chat::MessageRole::User,

                content: "Hello".to_string(),

            }

        ],

    };

    // call completion endpoint

    let mut response = chat_stream::completion(

        env::var("OPENAI_API_KEY").expect("environment variable OPENAI_API_KEY is not found."),

        &request).await.unwrap();

    // receive response

    let mut response_text = String::new();

    while let Some(response) = response.next().await {

        match response {

            Ok(response) => {

                if let Some(text_chunk) = response.choices[0].delta.content.clone() {

                    print!("{text_chunk}");

                    response_text.push_str(&text_chunk);

                }

            }

            Err(err) => {

                println!("{:?}", err);

                break;

            }

        }

    }

    // show response text

    println!("\nResponse Text: {response_text}");

}

```