Large Language Model Interaction Library

This Rust library provides easy-to-use functionality for interacting with various large language models (LLM). It offers methods for generating text completions and embedding documents using APIs of popular language models such as OpenAI's GPT-3.

Features

• Generate text completions
• Embed documents
• Embed queries
• Manage errors from OpenAI's API
• Chat models

Usage

Chat Generation

You can generate a Open AI text completions using the following:

let mut chat_llm = ChatOpenAI::default();
let text = chat_llm.execute("Hello, how are you?").await.unwrap();
println!("{}", text);

or a custom model

let mut chat_llm = ChatOpenAI::default().with_model(ChatModel::Gpt3_5Turbo16k);
let text = chat_llm.execute("Hello, how are you?").await.unwrap();
println!("{}", text);

more complex emaple

let chat_llm = ChatOpenAI::default().with_model(ChatModel::Gpt3_5Turbo16k);

 let json_str = r#"
    {
        "type": "user",
        "content": "Hello,how are you"
    }
    "#;

    // Deserialize JSON string into a HashMap
let messages: HashMap<String, String> = serde_json::from_str(json_str).unwrap();
let mut messages = messages_from_map(vec![mesagges]).unwrap();

let response =
    .chat_llm
    .generate(vec![messages.clone()])
    .await
    .unwrap();

let messages=messages_to_map(messages)
println!("{:?}", messages);

Document Embedding

use llm::embedding::embedder_trait::Embedder;
use llm::OpenAiEmbedder;

let embedder = OpenAiEmbedder::default();
let embeddings = embedder.embed_documents(vec!["Hello, how are you?".to_string()]).await.unwrap();
println!("{:?}", embeddings);

let query_embedding = embedder.embed_query("Hello, how are you?").await.unwrap();
println!("{:?}", query_embedding);

Note

You'll need to provide OpenAI's API key which can be set in the environment variable OPENAI_API_KEY or passed directly to the constructors.

Installation

[dependencies]
large-language-model-interaction = { version = "*", git = "https://github.com/Abraxas-365/llm_rust.git" }