vectus 0.1.38

A vector database implemented in Rust for learning purposes.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

use async_openai::{config::OpenAIConfig, types::CreateEmbeddingRequestArgs, Client};
use serde::{Deserialize, Serialize};
use std::{error::Error, sync::Arc};

pub enum ModelType {
    OpenAI(String),
}

pub struct Model {
    client: Arc<Client<OpenAIConfig>>,
}

impl Model {
    pub fn new(model_name: ModelType) -> Self {
        match model_name {
            ModelType::OpenAI(api_key) => Model {
                client: Arc::new(Client::with_config(
                    OpenAIConfig::new().with_api_key(api_key),
                )),
            },
        }
    }

    pub async fn get_embedding(&self, input: &String) -> Result<Vec<f32>, Box<dyn Error>> {
        let request = CreateEmbeddingRequestArgs::default()
            .model("text-embedding-3-large")
            .input([input])
            .build()
            .unwrap();
        let response = self.client.embeddings().create(request).await.unwrap();
        Ok(response.data[0].embedding.clone())
    }
}