chore: proper client extraction (#6996)

This commit is contained in:
jif-oai
2025-11-25 18:06:12 +00:00
committed by GitHub
parent 2845e2c006
commit 4502b1b263
45 changed files with 4893 additions and 2748 deletions

View File

@@ -0,0 +1,266 @@
use crate::ChatRequest;
use crate::auth::AuthProvider;
use crate::common::Prompt as ApiPrompt;
use crate::common::ResponseEvent;
use crate::common::ResponseStream;
use crate::endpoint::streaming::StreamingClient;
use crate::error::ApiError;
use crate::provider::Provider;
use crate::provider::WireApi;
use crate::sse::chat::spawn_chat_stream;
use crate::telemetry::SseTelemetry;
use codex_client::HttpTransport;
use codex_client::RequestTelemetry;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ReasoningItemContent;
use codex_protocol::models::ResponseItem;
use codex_protocol::protocol::SessionSource;
use futures::Stream;
use http::HeaderMap;
use serde_json::Value;
use std::collections::VecDeque;
use std::pin::Pin;
use std::sync::Arc;
use std::task::Context;
use std::task::Poll;
pub struct ChatClient<T: HttpTransport, A: AuthProvider> {
streaming: StreamingClient<T, A>,
}
impl<T: HttpTransport, A: AuthProvider> ChatClient<T, A> {
pub fn new(transport: T, provider: Provider, auth: A) -> Self {
Self {
streaming: StreamingClient::new(transport, provider, auth),
}
}
pub fn with_telemetry(
self,
request: Option<Arc<dyn RequestTelemetry>>,
sse: Option<Arc<dyn SseTelemetry>>,
) -> Self {
Self {
streaming: self.streaming.with_telemetry(request, sse),
}
}
pub async fn stream_request(&self, request: ChatRequest) -> Result<ResponseStream, ApiError> {
self.stream(request.body, request.headers).await
}
pub async fn stream_prompt(
&self,
model: &str,
prompt: &ApiPrompt,
conversation_id: Option<String>,
session_source: Option<SessionSource>,
) -> Result<ResponseStream, ApiError> {
use crate::requests::ChatRequestBuilder;
let request =
ChatRequestBuilder::new(model, &prompt.instructions, &prompt.input, &prompt.tools)
.conversation_id(conversation_id)
.session_source(session_source)
.build(self.streaming.provider())?;
self.stream_request(request).await
}
fn path(&self) -> &'static str {
match self.streaming.provider().wire {
WireApi::Chat => "chat/completions",
_ => "responses",
}
}
pub async fn stream(
&self,
body: Value,
extra_headers: HeaderMap,
) -> Result<ResponseStream, ApiError> {
self.streaming
.stream(self.path(), body, extra_headers, spawn_chat_stream)
.await
}
}
#[derive(Copy, Clone, Eq, PartialEq)]
pub enum AggregateMode {
AggregatedOnly,
Streaming,
}
/// Stream adapter that merges token deltas into a single assistant message per turn.
pub struct AggregatedStream {
inner: ResponseStream,
cumulative: String,
cumulative_reasoning: String,
pending: VecDeque<ResponseEvent>,
mode: AggregateMode,
}
impl Stream for AggregatedStream {
type Item = Result<ResponseEvent, ApiError>;
fn poll_next(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
let this = self.get_mut();
if let Some(ev) = this.pending.pop_front() {
return Poll::Ready(Some(Ok(ev)));
}
loop {
match Pin::new(&mut this.inner).poll_next(cx) {
Poll::Pending => return Poll::Pending,
Poll::Ready(None) => return Poll::Ready(None),
Poll::Ready(Some(Err(e))) => return Poll::Ready(Some(Err(e))),
Poll::Ready(Some(Ok(ResponseEvent::OutputItemDone(item)))) => {
let is_assistant_message = matches!(
&item,
ResponseItem::Message { role, .. } if role == "assistant"
);
if is_assistant_message {
match this.mode {
AggregateMode::AggregatedOnly => {
if this.cumulative.is_empty()
&& let ResponseItem::Message { content, .. } = &item
&& let Some(text) = content.iter().find_map(|c| match c {
ContentItem::OutputText { text } => Some(text),
_ => None,
})
{
this.cumulative.push_str(text);
}
continue;
}
AggregateMode::Streaming => {
if this.cumulative.is_empty() {
return Poll::Ready(Some(Ok(ResponseEvent::OutputItemDone(
item,
))));
} else {
continue;
}
}
}
}
return Poll::Ready(Some(Ok(ResponseEvent::OutputItemDone(item))));
}
Poll::Ready(Some(Ok(ResponseEvent::RateLimits(snapshot)))) => {
return Poll::Ready(Some(Ok(ResponseEvent::RateLimits(snapshot))));
}
Poll::Ready(Some(Ok(ResponseEvent::Completed {
response_id,
token_usage,
}))) => {
let mut emitted_any = false;
if !this.cumulative_reasoning.is_empty() {
let aggregated_reasoning = ResponseItem::Reasoning {
id: String::new(),
summary: Vec::new(),
content: Some(vec![ReasoningItemContent::ReasoningText {
text: std::mem::take(&mut this.cumulative_reasoning),
}]),
encrypted_content: None,
};
this.pending
.push_back(ResponseEvent::OutputItemDone(aggregated_reasoning));
emitted_any = true;
}
if !this.cumulative.is_empty() {
let aggregated_message = ResponseItem::Message {
id: None,
role: "assistant".to_string(),
content: vec![ContentItem::OutputText {
text: std::mem::take(&mut this.cumulative),
}],
};
this.pending
.push_back(ResponseEvent::OutputItemDone(aggregated_message));
emitted_any = true;
}
if emitted_any {
this.pending.push_back(ResponseEvent::Completed {
response_id: response_id.clone(),
token_usage: token_usage.clone(),
});
if let Some(ev) = this.pending.pop_front() {
return Poll::Ready(Some(Ok(ev)));
}
}
return Poll::Ready(Some(Ok(ResponseEvent::Completed {
response_id,
token_usage,
})));
}
Poll::Ready(Some(Ok(ResponseEvent::Created))) => {
continue;
}
Poll::Ready(Some(Ok(ResponseEvent::OutputTextDelta(delta)))) => {
this.cumulative.push_str(&delta);
if matches!(this.mode, AggregateMode::Streaming) {
return Poll::Ready(Some(Ok(ResponseEvent::OutputTextDelta(delta))));
} else {
continue;
}
}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta {
delta,
content_index,
}))) => {
this.cumulative_reasoning.push_str(&delta);
if matches!(this.mode, AggregateMode::Streaming) {
return Poll::Ready(Some(Ok(ResponseEvent::ReasoningContentDelta {
delta,
content_index,
})));
} else {
continue;
}
}
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryDelta { .. }))) => continue,
Poll::Ready(Some(Ok(ResponseEvent::ReasoningSummaryPartAdded { .. }))) => {
continue;
}
Poll::Ready(Some(Ok(ResponseEvent::OutputItemAdded(item)))) => {
return Poll::Ready(Some(Ok(ResponseEvent::OutputItemAdded(item))));
}
}
}
}
}
pub trait AggregateStreamExt {
fn aggregate(self) -> AggregatedStream;
fn streaming_mode(self) -> ResponseStream;
}
impl AggregateStreamExt for ResponseStream {
fn aggregate(self) -> AggregatedStream {
AggregatedStream::new(self, AggregateMode::AggregatedOnly)
}
fn streaming_mode(self) -> ResponseStream {
self
}
}
impl AggregatedStream {
fn new(inner: ResponseStream, mode: AggregateMode) -> Self {
AggregatedStream {
inner,
cumulative: String::new(),
cumulative_reasoning: String::new(),
pending: VecDeque::new(),
mode,
}
}
}

View File

@@ -0,0 +1,162 @@
use crate::auth::AuthProvider;
use crate::auth::add_auth_headers;
use crate::common::CompactionInput;
use crate::error::ApiError;
use crate::provider::Provider;
use crate::provider::WireApi;
use crate::telemetry::run_with_request_telemetry;
use codex_client::HttpTransport;
use codex_client::RequestTelemetry;
use codex_protocol::models::ResponseItem;
use http::HeaderMap;
use http::Method;
use serde::Deserialize;
use serde_json::to_value;
use std::sync::Arc;
pub struct CompactClient<T: HttpTransport, A: AuthProvider> {
transport: T,
provider: Provider,
auth: A,
request_telemetry: Option<Arc<dyn RequestTelemetry>>,
}
impl<T: HttpTransport, A: AuthProvider> CompactClient<T, A> {
pub fn new(transport: T, provider: Provider, auth: A) -> Self {
Self {
transport,
provider,
auth,
request_telemetry: None,
}
}
pub fn with_telemetry(mut self, request: Option<Arc<dyn RequestTelemetry>>) -> Self {
self.request_telemetry = request;
self
}
fn path(&self) -> Result<&'static str, ApiError> {
match self.provider.wire {
WireApi::Compact | WireApi::Responses => Ok("responses/compact"),
WireApi::Chat => Err(ApiError::Stream(
"compact endpoint requires responses wire api".to_string(),
)),
}
}
pub async fn compact(
&self,
body: serde_json::Value,
extra_headers: HeaderMap,
) -> Result<Vec<ResponseItem>, ApiError> {
let path = self.path()?;
let builder = || {
let mut req = self.provider.build_request(Method::POST, path);
req.headers.extend(extra_headers.clone());
req.body = Some(body.clone());
add_auth_headers(&self.auth, req)
};
let resp = run_with_request_telemetry(
self.provider.retry.to_policy(),
self.request_telemetry.clone(),
builder,
|req| self.transport.execute(req),
)
.await?;
let parsed: CompactHistoryResponse =
serde_json::from_slice(&resp.body).map_err(|e| ApiError::Stream(e.to_string()))?;
Ok(parsed.output)
}
pub async fn compact_input(
&self,
input: &CompactionInput<'_>,
extra_headers: HeaderMap,
) -> Result<Vec<ResponseItem>, ApiError> {
let body = to_value(input)
.map_err(|e| ApiError::Stream(format!("failed to encode compaction input: {e}")))?;
self.compact(body, extra_headers).await
}
}
#[derive(Debug, Deserialize)]
struct CompactHistoryResponse {
output: Vec<ResponseItem>,
}
#[cfg(test)]
mod tests {
use super::*;
use crate::provider::RetryConfig;
use async_trait::async_trait;
use codex_client::Request;
use codex_client::Response;
use codex_client::StreamResponse;
use codex_client::TransportError;
use http::HeaderMap;
use std::time::Duration;
#[derive(Clone, Default)]
struct DummyTransport;
#[async_trait]
impl HttpTransport for DummyTransport {
async fn execute(&self, _req: Request) -> Result<Response, TransportError> {
Err(TransportError::Build("execute should not run".to_string()))
}
async fn stream(&self, _req: Request) -> Result<StreamResponse, TransportError> {
Err(TransportError::Build("stream should not run".to_string()))
}
}
#[derive(Clone, Default)]
struct DummyAuth;
impl AuthProvider for DummyAuth {
fn bearer_token(&self) -> Option<String> {
None
}
}
fn provider(wire: WireApi) -> Provider {
Provider {
name: "test".to_string(),
base_url: "https://example.com/v1".to_string(),
query_params: None,
wire,
headers: HeaderMap::new(),
retry: RetryConfig {
max_attempts: 1,
base_delay: Duration::from_millis(1),
retry_429: false,
retry_5xx: true,
retry_transport: true,
},
stream_idle_timeout: Duration::from_secs(1),
}
}
#[tokio::test]
async fn errors_when_wire_is_chat() {
let client = CompactClient::new(DummyTransport, provider(WireApi::Chat), DummyAuth);
let input = CompactionInput {
model: "gpt-test",
input: &[],
instructions: "inst",
};
let err = client
.compact_input(&input, HeaderMap::new())
.await
.expect_err("expected wire mismatch to fail");
match err {
ApiError::Stream(msg) => {
assert_eq!(msg, "compact endpoint requires responses wire api");
}
other => panic!("unexpected error: {other:?}"),
}
}
}

View File

@@ -0,0 +1,4 @@
pub mod chat;
pub mod compact;
pub mod responses;
mod streaming;

View File

@@ -0,0 +1,107 @@
use crate::auth::AuthProvider;
use crate::common::Prompt as ApiPrompt;
use crate::common::Reasoning;
use crate::common::ResponseStream;
use crate::common::TextControls;
use crate::endpoint::streaming::StreamingClient;
use crate::error::ApiError;
use crate::provider::Provider;
use crate::provider::WireApi;
use crate::requests::ResponsesRequest;
use crate::requests::ResponsesRequestBuilder;
use crate::sse::spawn_response_stream;
use crate::telemetry::SseTelemetry;
use codex_client::HttpTransport;
use codex_client::RequestTelemetry;
use codex_protocol::protocol::SessionSource;
use http::HeaderMap;
use serde_json::Value;
use std::sync::Arc;
pub struct ResponsesClient<T: HttpTransport, A: AuthProvider> {
streaming: StreamingClient<T, A>,
}
#[derive(Default)]
pub struct ResponsesOptions {
pub reasoning: Option<Reasoning>,
pub include: Vec<String>,
pub prompt_cache_key: Option<String>,
pub text: Option<TextControls>,
pub store_override: Option<bool>,
pub conversation_id: Option<String>,
pub session_source: Option<SessionSource>,
}
impl<T: HttpTransport, A: AuthProvider> ResponsesClient<T, A> {
pub fn new(transport: T, provider: Provider, auth: A) -> Self {
Self {
streaming: StreamingClient::new(transport, provider, auth),
}
}
pub fn with_telemetry(
self,
request: Option<Arc<dyn RequestTelemetry>>,
sse: Option<Arc<dyn SseTelemetry>>,
) -> Self {
Self {
streaming: self.streaming.with_telemetry(request, sse),
}
}
pub async fn stream_request(
&self,
request: ResponsesRequest,
) -> Result<ResponseStream, ApiError> {
self.stream(request.body, request.headers).await
}
pub async fn stream_prompt(
&self,
model: &str,
prompt: &ApiPrompt,
options: ResponsesOptions,
) -> Result<ResponseStream, ApiError> {
let ResponsesOptions {
reasoning,
include,
prompt_cache_key,
text,
store_override,
conversation_id,
session_source,
} = options;
let request = ResponsesRequestBuilder::new(model, &prompt.instructions, &prompt.input)
.tools(&prompt.tools)
.parallel_tool_calls(prompt.parallel_tool_calls)
.reasoning(reasoning)
.include(include)
.prompt_cache_key(prompt_cache_key)
.text(text)
.conversation(conversation_id)
.session_source(session_source)
.store_override(store_override)
.build(self.streaming.provider())?;
self.stream_request(request).await
}
fn path(&self) -> &'static str {
match self.streaming.provider().wire {
WireApi::Responses | WireApi::Compact => "responses",
WireApi::Chat => "chat/completions",
}
}
pub async fn stream(
&self,
body: Value,
extra_headers: HeaderMap,
) -> Result<ResponseStream, ApiError> {
self.streaming
.stream(self.path(), body, extra_headers, spawn_response_stream)
.await
}
}

View File

@@ -0,0 +1,82 @@
use crate::auth::AuthProvider;
use crate::auth::add_auth_headers;
use crate::common::ResponseStream;
use crate::error::ApiError;
use crate::provider::Provider;
use crate::telemetry::SseTelemetry;
use crate::telemetry::run_with_request_telemetry;
use codex_client::HttpTransport;
use codex_client::RequestTelemetry;
use codex_client::StreamResponse;
use http::HeaderMap;
use http::Method;
use serde_json::Value;
use std::sync::Arc;
use std::time::Duration;
pub(crate) struct StreamingClient<T: HttpTransport, A: AuthProvider> {
transport: T,
provider: Provider,
auth: A,
request_telemetry: Option<Arc<dyn RequestTelemetry>>,
sse_telemetry: Option<Arc<dyn SseTelemetry>>,
}
impl<T: HttpTransport, A: AuthProvider> StreamingClient<T, A> {
pub(crate) fn new(transport: T, provider: Provider, auth: A) -> Self {
Self {
transport,
provider,
auth,
request_telemetry: None,
sse_telemetry: None,
}
}
pub(crate) fn with_telemetry(
mut self,
request: Option<Arc<dyn RequestTelemetry>>,
sse: Option<Arc<dyn SseTelemetry>>,
) -> Self {
self.request_telemetry = request;
self.sse_telemetry = sse;
self
}
pub(crate) fn provider(&self) -> &Provider {
&self.provider
}
pub(crate) async fn stream(
&self,
path: &str,
body: Value,
extra_headers: HeaderMap,
spawner: fn(StreamResponse, Duration, Option<Arc<dyn SseTelemetry>>) -> ResponseStream,
) -> Result<ResponseStream, ApiError> {
let builder = || {
let mut req = self.provider.build_request(Method::POST, path);
req.headers.extend(extra_headers.clone());
req.headers.insert(
http::header::ACCEPT,
http::HeaderValue::from_static("text/event-stream"),
);
req.body = Some(body.clone());
add_auth_headers(&self.auth, req)
};
let stream_response = run_with_request_telemetry(
self.provider.retry.to_policy(),
self.request_telemetry.clone(),
builder,
|req| self.transport.stream(req),
)
.await?;
Ok(spawner(
stream_response,
self.provider.stream_idle_timeout,
self.sse_telemetry.clone(),
))
}
}