vtcode-llm 0.173.0

LLM provider abstraction, client implementations, and streaming for VT Code
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

const WEBSOCKET_ACTIVE_RESPONSE_ERROR_PREFIX: &str = "Conversation already has an active response in progress:";
const OPENAI_BETA_RESPONSES_WEBSOCKET_V2: &str = "responses=v2";
const WEBSOCKET_CONNECTION_LIMIT_REACHED_CODE: &str = "websocket_connection_limit_reached";
const PREVIOUS_RESPONSE_NOT_FOUND_CODE: &str = "previous_response_not_found";
const WEBSOCKET_AUTH_RETRY_STATUSES: [&str; 2] = ["401", "403"];

// Retained custom WebSocket transport while VTCode migrates the optional path
// to Rig. Rig 0.40 exposes `client.responses_websocket()` behind its
// `websocket` feature, including warm-up requests, but it does not yet cover
// VTCode's custom auth refresh and retry policy. Keep this path separate from
// normal HTTP/SSE request shaping: it may use `previous_response_id` only as
// connection-local state, and every websocket request still sends `store=false`.
use crate::provider::LLMError;

fn is_websocket_active_response_error(err: &LLMError) -> bool {
    let message = match err {
        LLMError::Provider { message, .. } | LLMError::Network { message, .. } => message,
        LLMError::Authentication { .. } | LLMError::RateLimit { .. } | LLMError::InvalidRequest { .. } => return false,
    };

    message.contains(WEBSOCKET_ACTIVE_RESPONSE_ERROR_PREFIX)
}

fn is_websocket_connection_limit_error(err: &LLMError) -> bool {
    let message = match err {
        LLMError::Provider { message, .. } | LLMError::Network { message, .. } => message,
        LLMError::Authentication { .. } | LLMError::RateLimit { .. } | LLMError::InvalidRequest { .. } => return false,
    };

    message.contains(WEBSOCKET_CONNECTION_LIMIT_REACHED_CODE)
}

fn is_websocket_previous_response_not_found_error(err: &LLMError) -> bool {
    let message = match err {
        LLMError::Provider { message, .. } | LLMError::Network { message, .. } => message,
        LLMError::Authentication { .. } | LLMError::RateLimit { .. } | LLMError::InvalidRequest { .. } => return false,
    };

    message.contains(PREVIOUS_RESPONSE_NOT_FOUND_CODE)
}

fn is_websocket_reconnect_error(err: &LLMError) -> bool {
    matches!(err, LLMError::Network { .. })
}

fn is_websocket_auth_retryable(error: &tokio_tungstenite::tungstenite::Error) -> bool {
    let message = error.to_string();
    WEBSOCKET_AUTH_RETRY_STATUSES.iter().any(|status| message.contains(status))
}

mod events;
mod provider;
mod session;

#[cfg(test)]
use events::apply_generate_mode;
use events::{
    canonical_history, completed_replay_output, format_network_error, format_provider_error, input_is_incremental,
    prepare_websocket_event, read_websocket_event, responses_websocket_url,
};
use session::PreparedWebSocketEvent;
pub(super) use session::{OpenAIResponsesWebSocketContinuationCache, OpenAIResponsesWebSocketSession};
use session::{WebSocketLease, completed_websocket_stream, websocket_deadline_error};

#[cfg(test)]
mod tests;