openai/codex
1
0
Fork 0
mirror of https://github.com/openai/codex.git synced 2026-08-23 13:09:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Route LM Studio requests through the shared HTTP client (#34678)

Browse Source 
## What changed

- Use the configured route-aware HTTP client pool for LM Studio server requests.
- Add connection-timeout support to `HttpClientBuilder` and route-aware pools, and keep LM Studio's five-second limit scoped to connection establishment.
- Verify that LM Studio accepts a response that arrives after the connection timeout has elapsed once the connection is established.

GitOrigin-RevId: c4300f4b5d37c4418822783ab09cb50d506ee423
This commit is contained in:
Celia Chen
2026-07-22 04:12:23 +00:00
committed by copyberry
parent 4f3852107e
commit 21db216db0
7 changed files with 161 additions and 118 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
codex-rs/lmstudio/Cargo.toml
View File

@@ -12,8 +12,8 @@ doctest = false
[dependencies]
codex-core = { path = "../core" }
codex-http-client = { workspace = true }
codex-model-provider-info = { path = "../model-provider-info" }
reqwest = { version = "0.12", features = ["json", "stream"] }
serde_json = "1"
tokio = { version = "1", features = ["rt"] }
tracing = { version = "0.1.44", features = ["log"] }

91
codex-rs/lmstudio/src/client.rs
View File

@@ -1,15 +1,19 @@
use codex_core::config::Config;
use codex_http_client::ClientRouteClass;
use codex_http_client::RouteAwareClientPool;
use codex_model_provider_info::LMSTUDIO_OSS_PROVIDER_ID;
use std::io;
use std::path::Path;
use std::time::Duration;
#[derive(Clone)]
pub struct LMStudioClient {
client: reqwest::Client,
client: RouteAwareClientPool,
base_url: String,
}
const LMSTUDIO_CONNECTION_ERROR: &str = "LM Studio is not responding. Install from https://lmstudio.ai/download and run 'lms server start'.";
const LMSTUDIO_CONNECTION_TIMEOUT: Duration = Duration::from_secs(5);
impl LMStudioClient {
pub async fn try_from_provider(config: &Config) -> std::io::Result<Self> {
@@ -29,10 +33,11 @@ impl LMStudioClient {
)
})?;
let client = reqwest::Client::builder()
.connect_timeout(std::time::Duration::from_secs(5))
.build()
.unwrap_or_else(|_| reqwest::Client::new());
let client = RouteAwareClientPool::with_connect_timeout(
config.http_client_factory(),
ClientRouteClass::Other,
LMSTUDIO_CONNECTION_TIMEOUT,
);
let client = LMStudioClient {
client,
@@ -188,19 +193,6 @@ impl LMStudioClient {
tracing::info!("Successfully downloaded model '{model}'");
Ok(())
}
/// Low-level constructor given a raw host root, e.g. "http://localhost:1234".
#[cfg(test)]
fn from_host_root(host_root: impl Into<String>) -> Self {
let client = reqwest::Client::builder()
.connect_timeout(std::time::Duration::from_secs(5))
.build()
.unwrap_or_else(|_| reqwest::Client::new());
Self {
client,
base_url: host_root.into(),
}
}
}
#[cfg(test)]
@@ -208,6 +200,23 @@ mod tests {
#![allow(clippy::expect_used, clippy::unwrap_used)]
use super::*;
fn client_from_host_root(
host_root: impl Into<String>,
connection_timeout: Duration,
) -> LMStudioClient {
let client = RouteAwareClientPool::with_connect_timeout(
codex_http_client::HttpClientFactory::new(
codex_http_client::OutboundProxyPolicy::ReqwestDefault,
),
ClientRouteClass::Other,
connection_timeout,
);
LMStudioClient {
client,
base_url: host_root.into(),
}
}
#[tokio::test]
async fn test_fetch_models_happy_path() {
if std::env::var(codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR).is_ok() {
@@ -235,7 +244,7 @@ mod tests {
.mount(&server)
.await;
let client = LMStudioClient::from_host_root(server.uri());
let client = client_from_host_root(server.uri(), LMSTUDIO_CONNECTION_TIMEOUT);
let models = client.fetch_models().await.expect("fetch models");
assert!(models.contains(&"openai/gpt-oss-20b".to_string()));
}
@@ -260,7 +269,7 @@ mod tests {
.mount(&server)
.await;
let client = LMStudioClient::from_host_root(server.uri());
let client = client_from_host_root(server.uri(), LMSTUDIO_CONNECTION_TIMEOUT);
let result = client.fetch_models().await;
assert!(result.is_err());
assert!(
@@ -288,7 +297,7 @@ mod tests {
.mount(&server)
.await;
let client = LMStudioClient::from_host_root(server.uri());
let client = client_from_host_root(server.uri(), LMSTUDIO_CONNECTION_TIMEOUT);
let result = client.fetch_models().await;
assert!(result.is_err());
assert!(
@@ -316,13 +325,40 @@ mod tests {
.mount(&server)
.await;
let client = LMStudioClient::from_host_root(server.uri());
let client = client_from_host_root(server.uri(), LMSTUDIO_CONNECTION_TIMEOUT);
client
.check_server()
.await
.expect("server check should pass");
}
#[tokio::test]
async fn test_check_server_allows_slow_response_after_connect() {
if std::env::var(codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR).is_ok() {
tracing::info!(
"{} is set; skipping test_check_server_allows_slow_response_after_connect",
codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR
);
return;
}
let server = wiremock::MockServer::start().await;
wiremock::Mock::given(wiremock::matchers::method("GET"))
.and(wiremock::matchers::path("/models"))
.respond_with(
wiremock::ResponseTemplate::new(200).set_delay(Duration::from_millis(250)),
)
.mount(&server)
.await;
let client = client_from_host_root(server.uri(), Duration::from_millis(100));
client
.check_server()
.await
.expect("server check should allow a slow response after connecting");
}
#[tokio::test]
async fn test_check_server_error() {
if std::env::var(codex_core::spawn::CODEX_SANDBOX_NETWORK_DISABLED_ENV_VAR).is_ok() {
@@ -340,7 +376,7 @@ mod tests {
.mount(&server)
.await;
let client = LMStudioClient::from_host_root(server.uri());
let client = client_from_host_root(server.uri(), LMSTUDIO_CONNECTION_TIMEOUT);
let result = client.check_server().await;
assert!(result.is_err());
assert!(
@@ -385,13 +421,4 @@ mod tests {
}
}
}
#[test]
fn test_from_host_root() {
let client = LMStudioClient::from_host_root("http://localhost:1234");
assert_eq!(client.base_url, "http://localhost:1234");
let client = LMStudioClient::from_host_root("https://example.com:8080/api");
assert_eq!(client.base_url, "https://example.com:8080/api");
}
}