Crates.io | openai-api |
lib.rs | openai-api |
version | 0.1.4 |
source | src |
created_at | 2020-12-06 01:08:53.366091 |
updated_at | 2020-12-22 06:35:13.743516 |
description | OpenAI API library for rust |
homepage | https://github.com/deontologician/openai-api-rust/ |
repository | https://github.com/deontologician/openai-api-rust/ |
max_upload_size | |
id | 320029 |
size | 105,146 |
A simple rust client for OpenAI API.
Has a few conveniences, but is mostly at the level of the API itself.
$ cargo add openai-api
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let api_token = std::env::var("OPENAI_SK")?;
let client = openai_api::Client::new(&api_token);
let prompt = String::from("Once upon a time,");
println!(
"{}{}",
prompt,
client.complete(prompt.as_str()).await?
);
Ok(())
}
For simple demos and debugging, you can do a completion and use the Display
instance of a Completion
object to convert it to a string:
let response = client.complete_prompt("Once upon a time").await?;
println!("{}", response);
To configure the prompt more explicitly, you can use the CompletionArgs
builder:
let args = openai_api::api::CompletionArgs::builder()
.prompt("Once upon a time,")
.engine(Engine::Davinci)
.max_tokens(20)
.temperature(0.7)
.top_p(0.9)
.stop(vec!["\n".into()]);
let completion = client.complete_prompt(args).await?;
println!("Response: {}", response.choices[0].text);
println!("Model used: {}", response.model);
See examples/