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
|
use super::{ExtraConfig, Model, OpenAIClient, PromptType, SendData};
use crate::{render::ReplyHandler, utils::PromptKind};
use anyhow::{anyhow, bail, Result};
use async_trait::async_trait;
use futures_util::StreamExt;
use reqwest::{Client as ReqwestClient, RequestBuilder};
use reqwest_eventsource::{Error as EventSourceError, Event, RequestBuilderExt};
use serde::Deserialize;
use serde_json::{json, Value};
const API_BASE: &str = "https://api.openai.com/v1";
const MODELS: [(&str, usize, &str); 8] = [
// https://platform.openai.com/docs/models
("gpt-3.5-turbo", 16385, "text"),
("gpt-3.5-turbo-1106", 16385, "text"),
("gpt-4-turbo", 128000, "text,vision"),
("gpt-4-turbo-preview", 128000, "text"),
("gpt-4-1106-preview", 128000, "text"),
("gpt-4-vision-preview", 128000, "text,vision"),
("gpt-4", 8192, "text"),
("gpt-4-32k", 32768, "text"),
];
#[derive(Debug, Clone, Deserialize, Default)]
pub struct OpenAIConfig {
pub name: Option<String>,
pub api_key: Option<String>,
pub api_base: Option<String>,
pub organization_id: Option<String>,
pub extra: Option<ExtraConfig>,
}
openai_compatible_client!(OpenAIClient);
impl OpenAIClient {
config_get_fn!(api_key, get_api_key);
config_get_fn!(api_base, get_api_base);
pub const PROMPTS: [PromptType<'static>; 1] =
[("api_key", "API Key:", true, PromptKind::String)];
pub fn list_models(local_config: &OpenAIConfig) -> Vec<Model> {
let client_name = Self::name(local_config);
MODELS
.into_iter()
.map(|(name, max_input_tokens, capabilities)| {
Model::new(client_name, name)
.set_capabilities(capabilities.into())
.set_max_input_tokens(Some(max_input_tokens))
})
.collect()
}
fn request_builder(&self, client: &ReqwestClient, data: SendData) -> Result<RequestBuilder> {
let api_key = self.get_api_key()?;
let api_base = self.get_api_base().unwrap_or_else(|_| API_BASE.to_string());
let body = openai_build_body(data, self.model.name.clone());
let url = format!("{api_base}/chat/completions");
debug!("OpenAI Request: {url} {body}");
let mut builder = client.post(url).bearer_auth(api_key).json(&body);
if let Some(organization_id) = &self.config.organization_id {
builder = builder.header("OpenAI-Organization", organization_id);
}
Ok(builder)
}
}
pub async fn openai_send_message(builder: RequestBuilder) -> Result<String> {
let data: Value = builder.send().await?.json().await?;
if let Some(err_msg) = data["error"]["message"].as_str() {
bail!("{err_msg}");
}
let output = data["choices"][0]["message"]["content"]
.as_str()
.ok_or_else(|| anyhow!("Invalid response data: {data}"))?;
Ok(output.to_string())
}
pub async fn openai_send_message_streaming(
builder: RequestBuilder,
handler: &mut ReplyHandler,
) -> Result<()> {
let mut es = builder.eventsource()?;
while let Some(event) = es.next().await {
match event {
Ok(Event::Open) => {}
Ok(Event::Message(message)) => {
if message.data == "[DONE]" {
break;
}
let data: Value = serde_json::from_str(&message.data)?;
if let Some(text) = data["choices"][0]["delta"]["content"].as_str() {
handler.text(text)?;
}
}
Err(err) => {
match err {
EventSourceError::InvalidStatusCode(_, res) => {
let data: Value = res.json().await?;
if let Some(err_msg) = data["error"]["message"].as_str() {
bail!("{err_msg}");
} else if let Some(err_msg) = data["message"].as_str() {
bail!("{err_msg}");
} else {
bail!("Request failed, {data}");
}
}
EventSourceError::StreamEnded => {}
_ => {
bail!("{}", err);
}
}
es.close();
}
}
}
Ok(())
}
pub fn openai_build_body(data: SendData, model: String) -> Value {
let SendData {
messages,
temperature,
stream,
} = data;
let mut body = json!({
"model": model,
"messages": messages,
});
// The default max_tokens of gpt-4-vision-preview is only 16, we need to make it larger
if model == "gpt-4-vision-preview" {
body["max_tokens"] = json!(4096);
}
if let Some(v) = temperature {
body["temperature"] = v.into();
}
if stream {
body["stream"] = true.into();
}
body
}
|
