rag 0.1.0

A Rust library and CLI for Retrieval-Augmented Generation
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
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165

use async_trait::async_trait;
use reqwest::Client;
use serde::{Deserialize, Serialize};

use crate::errors::{RagError, Result};

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct EmbeddingRequest {
    pub text: String,
    pub model: String,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct EmbeddingResponse {
    pub embedding: Vec<f32>,
    pub model: String,
}

#[async_trait]
pub trait EmbeddingModel: Send + Sync {
    async fn embed(&self, texts: Vec<String>) -> Result<Vec<Vec<f32>>>;
    async fn embed_single(&self, text: &str) -> Result<Vec<f32>> {
        let embeddings = self.embed(vec![text.to_string()]).await?;
        Ok(embeddings.into_iter().next().ok_or(RagError::EmbeddingError(
            "No embedding returned".to_string(),
        ))?)
    }
}

#[derive(Clone)]
pub struct OpenAIEmbeddingModel {
    client: Client,
    api_key: String,
    model: String,
    base_url: String,
}

impl OpenAIEmbeddingModel {
    pub fn new(api_key: String) -> Self {
        Self::with_model(api_key, "text-embedding-ada-002".to_string())
    }

    pub fn with_model(api_key: String, model: String) -> Self {
        Self {
            client: Client::new(),
            api_key,
            model,
            base_url: "https://api.openai.com/v1".to_string(),
        }
    }

    pub fn with_base_url(mut self, base_url: String) -> Self {
        self.base_url = base_url;
        self
    }
}

#[derive(Debug, Serialize)]
struct OpenAIRequest {
    input: Vec<String>,
    model: String,
}

#[derive(Debug, Deserialize)]
struct OpenAIResponse {
    data: Vec<OpenAIEmbeddingData>,
    #[allow(dead_code)]
    model: String,
}

#[derive(Debug, Deserialize)]
struct OpenAIEmbeddingData {
    embedding: Vec<f32>,
}

#[async_trait]
impl EmbeddingModel for OpenAIEmbeddingModel {
    async fn embed(&self, texts: Vec<String>) -> Result<Vec<Vec<f32>>> {
        let request = OpenAIRequest {
            input: texts.clone(),
            model: self.model.clone(),
        };

        let response = self
            .client
            .post(&format!("{}/embeddings", self.base_url))
            .header("Authorization", format!("Bearer {}", self.api_key))
            .json(&request)
            .send()
            .await?;

        if !response.status().is_success() {
            let error_text = response.text().await?;
            return Err(RagError::EmbeddingError(error_text));
        }

        let openai_response: OpenAIResponse = response.json().await?;

        Ok(openai_response.data.into_iter().map(|d| d.embedding).collect())
    }
}

#[derive(Clone)]
pub struct OllamaEmbeddingModel {
    client: Client,
    model: String,
    base_url: String,
}

impl OllamaEmbeddingModel {
    pub fn new(model: String) -> Self {
        Self {
            client: Client::new(),
            model,
            base_url: "http://localhost:11434".to_string(),
        }
    }

    pub fn with_base_url(mut self, base_url: String) -> Self {
        self.base_url = base_url;
        self
    }
}

#[derive(Debug, Serialize)]
struct OllamaRequest {
    model: String,
    prompt: String,
}

#[derive(Debug, Deserialize)]
struct OllamaResponse {
    embedding: Vec<f32>,
}

#[async_trait]
impl EmbeddingModel for OllamaEmbeddingModel {
    async fn embed(&self, texts: Vec<String>) -> Result<Vec<Vec<f32>>> {
        let mut embeddings = Vec::new();

        for text in texts {
            let request = OllamaRequest {
                model: self.model.clone(),
                prompt: text,
            };

            let response = self
                .client
                .post(&format!("{}/api/embeddings", self.base_url))
                .json(&request)
                .send()
                .await?;

            if !response.status().is_success() {
                let error_text = response.text().await?;
                return Err(RagError::EmbeddingError(error_text));
            }

            let ollama_response: OllamaResponse = response.json().await?;
            embeddings.push(ollama_response.embedding);
        }

        Ok(embeddings)
    }
}