kwaak 0.1.1

Run a team of autonomous agents on your code, right from your terminal
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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189

use super::ApiKey;
use anyhow::{Context as _, Result};
use serde::{Deserialize, Serialize};
use swiftide::{
    chat_completion::ChatCompletion,
    integrations,
    traits::{EmbeddingModel, SimplePrompt},
};
use url::Url;

#[derive(Debug, Clone, Deserialize, Serialize)]
#[serde(untagged)]
#[allow(clippy::large_enum_variant)] // Parent is always on the heap in config
pub enum LLMConfigurations {
    Single(LLMConfiguration),
    Multiple {
        // TODO: Should probably be with reduced attrs on needed per item
        indexing: LLMConfiguration,
        embedding: LLMConfiguration,
        query: LLMConfiguration,
    },
}

#[derive(Debug, Clone, Deserialize, Serialize)]
#[serde(tag = "provider")]
pub enum LLMConfiguration {
    OpenAI {
        api_key: ApiKey,
        #[serde(default)]
        prompt_model: OpenAIPromptModel,
        #[serde(default)]
        embedding_model: OpenAIEmbeddingModel,
        base_url: Option<Url>,
    },
    // Groq {
    //     api_key: SecretString,
    //     prompt_model: String,
    // },
    // Ollama {
    //     prompt_model: Option<String>,
    //     embedding_model: Option<String>,
    //     vector_size: Option<usize>,
    // },
    // AWSBedrock {
    //     prompt_model: String,
    // },
    // FastEmbed {
    //     embedding_model: String,
    //     vector_size: usize,
    // },
}

impl LLMConfiguration {
    pub(crate) fn vector_size(&self) -> i32 {
        match self {
            LLMConfiguration::OpenAI {
                embedding_model, ..
            } => match embedding_model {
                OpenAIEmbeddingModel::TextEmbedding3Small => 1536,
                OpenAIEmbeddingModel::TextEmbedding3Large => 3072,
            },
        }
    }
}

#[derive(
    Debug,
    Clone,
    Deserialize,
    Serialize,
    PartialEq,
    strum_macros::EnumString,
    strum_macros::Display,
    Default,
)]
pub enum OpenAIPromptModel {
    #[strum(serialize = "gpt-4o-mini")]
    #[serde(rename = "gpt-4o-mini")]
    #[default]
    GPT4OMini,
    #[strum(serialize = "gpt-4o")]
    #[serde(rename = "gpt-4o")]
    GPT4O,
}

#[derive(
    Debug,
    Clone,
    Deserialize,
    Serialize,
    strum_macros::EnumString,
    strum_macros::Display,
    PartialEq,
    Default,
)]
pub enum OpenAIEmbeddingModel {
    #[strum(serialize = "text-embedding-3-small")]
    #[serde(rename = "text-embedding-3-small")]
    TextEmbedding3Small,
    #[strum(serialize = "text-embedding-3-large")]
    #[serde(rename = "text-embedding-3-large")]
    #[default]
    TextEmbedding3Large,
}

fn build_openai(
    api_key: &ApiKey,
    embedding_model: &OpenAIEmbeddingModel,
    prompt_model: &OpenAIPromptModel,
    base_url: Option<&Url>,
) -> Result<integrations::openai::OpenAI> {
    let mut config =
        async_openai::config::OpenAIConfig::default().with_api_key(api_key.expose_secret());

    if let Some(base_url) = base_url {
        config = config.with_api_base(base_url.to_string());
    };

    integrations::openai::OpenAI::builder()
        .client(async_openai::Client::with_config(config))
        .default_prompt_model(prompt_model.to_string())
        .default_embed_model(embedding_model.to_string())
        .build()
        .context("Failed to build OpenAI client")
}

impl TryInto<Box<dyn EmbeddingModel>> for &LLMConfiguration {
    type Error = anyhow::Error;

    fn try_into(self) -> std::result::Result<Box<dyn EmbeddingModel>, Self::Error> {
        let boxed = match self {
            LLMConfiguration::OpenAI {
                api_key,
                embedding_model,
                prompt_model,
                base_url,
            } => Box::new(build_openai(
                api_key,
                embedding_model,
                prompt_model,
                base_url.as_ref(),
            )?),
        };

        Ok(boxed)
    }
}

impl TryInto<Box<dyn SimplePrompt>> for &LLMConfiguration {
    type Error = anyhow::Error;
    fn try_into(self) -> std::result::Result<Box<dyn SimplePrompt>, Self::Error> {
        let boxed = match self {
            LLMConfiguration::OpenAI {
                api_key,
                embedding_model,
                prompt_model,
                base_url,
            } => Box::new(build_openai(
                api_key,
                embedding_model,
                prompt_model,
                base_url.as_ref(),
            )?),
        };

        Ok(boxed)
    }
}

impl TryInto<Box<dyn ChatCompletion>> for &LLMConfiguration {
    type Error = anyhow::Error;
    fn try_into(self) -> std::result::Result<Box<dyn ChatCompletion>, Self::Error> {
        let boxed = match self {
            LLMConfiguration::OpenAI {
                api_key,
                embedding_model,
                prompt_model,
                base_url,
            } => Box::new(build_openai(
                api_key,
                embedding_model,
                prompt_model,
                base_url.as_ref(),
            )?),
        };

        Ok(boxed)
    }
}