This commit is contained in:
jif-oai
2025-10-31 13:28:33 +01:00
parent 69341a04f4
commit cfe7fb825c
6 changed files with 67 additions and 39 deletions

View File

@@ -140,7 +140,7 @@ impl ApiClient for ChatCompletionsApiClient {
.get(reqwest::header::RETRY_AFTER)
.and_then(|v| v.to_str().ok())
.and_then(|s| s.parse::<u64>().ok())
.map(|s| Duration::from_secs(s));
.map(Duration::from_secs);
tokio::time::sleep(retry_after.unwrap_or_else(|| backoff(attempt))).await;
}
Err(error) => {
@@ -624,7 +624,7 @@ async fn process_chat_sse<S>(
.and_then(|v| v.as_array())
{
for call in tool_calls {
if let Some(index) = call.get("index").and_then(|i| i.as_u64())
if let Some(index) = call.get("index").and_then(serde_json::Value::as_u64)
&& index == 0
&& let Some(function) = call.get("function")
{
@@ -641,8 +641,7 @@ async fn process_chat_sse<S>(
if let Some(finish) = choice.get("finish_reason").and_then(|f| f.as_str())
&& finish == "tool_calls"
{
if let Some(name) = function_call_state.name.take() {
&& let Some(name) = function_call_state.name.take() {
let call_id =
function_call_state.call_id.take().unwrap_or_default();
let arguments = std::mem::take(&mut function_call_state.arguments);
@@ -655,7 +654,6 @@ async fn process_chat_sse<S>(
let _ =
tx_event.send(Ok(ResponseEvent::OutputItemDone(item))).await;
}
}
}
}
}

View File

@@ -120,11 +120,10 @@ impl ModelProviderInfo {
let url = self.get_full_url(effective_auth.as_ref());
let mut builder = client.post(url);
if let Some(context) = effective_auth.as_ref() {
if let Some(token) = context.bearer_token.as_ref() {
if let Some(context) = effective_auth.as_ref()
&& let Some(token) = context.bearer_token.as_ref() {
builder = builder.bearer_auth(token);
}
}
Ok(self.apply_http_headers(builder))
}

View File

@@ -74,13 +74,12 @@ impl ApiClient for ResponsesApiClient {
let mut payload_json = self.build_payload(&prompt)?;
if self.config.provider.is_azure_responses_endpoint() {
if let Some(input_value) = payload_json.get_mut("input")
if self.config.provider.is_azure_responses_endpoint()
&& let Some(input_value) = payload_json.get_mut("input")
&& let Some(array) = input_value.as_array_mut()
{
attach_item_ids_array(array, &prompt.input);
}
}
let max_attempts = self.config.provider.request_max_retries();
for attempt in 0..=max_attempts {
@@ -218,7 +217,7 @@ impl ResponsesApiClient {
.headers()
.get("cf-ray")
.and_then(|v| v.to_str().ok())
.map(|s| s.to_string());
.map(std::string::ToString::to_string);
}
match res {
@@ -260,8 +259,8 @@ impl ResponsesApiClient {
.and_then(|s| s.parse::<u64>().ok());
let retry_after = retry_after_secs.map(|s| Duration::from_millis(s * 1_000));
if status == StatusCode::UNAUTHORIZED {
if let Some(provider) = self.config.auth_provider.as_ref()
if status == StatusCode::UNAUTHORIZED
&& let Some(provider) = self.config.auth_provider.as_ref()
&& let Some(ctx) = auth.as_ref()
&& ctx.mode == AuthMode::ChatGPT
{
@@ -270,7 +269,6 @@ impl ResponsesApiClient {
.await
.map_err(|err| StreamAttemptError::Fatal(Error::Auth(err)))?;
}
}
if !(status == StatusCode::TOO_MANY_REQUESTS
|| status == StatusCode::UNAUTHORIZED
@@ -718,13 +716,13 @@ pub async fn stream_from_fixture(
let (tx_event, rx_event) = mpsc::channel::<Result<ResponseEvent>>(1600);
let display_path = path.as_ref().display().to_string();
let file = std::fs::File::open(path.as_ref())
.map_err(|e| Error::Other(format!("failed to open fixture {}: {}", display_path, e)))?;
.map_err(|e| Error::Other(format!("failed to open fixture {display_path}: {e}")))?;
let lines = std::io::BufReader::new(file).lines();
let mut content = String::new();
for line in lines {
let line = line
.map_err(|e| Error::Other(format!("failed to read fixture {}: {}", display_path, e)))?;
.map_err(|e| Error::Other(format!("failed to read fixture {display_path}: {e}")))?;
content.push_str(&line);
content.push_str("\n\n");
}

View File

@@ -101,7 +101,7 @@ impl ResponsesBackend {
self.client
.stream(prompt)
.await
.map(|stream| stream.boxed())
.map(futures::StreamExt::boxed)
}
}
@@ -181,8 +181,8 @@ impl ModelClient {
pub async fn stream(&self, payload: &StreamPayload) -> Result<ResponseStream> {
let mut prompt = payload.prompt.clone();
self.populate_prompt(&mut prompt);
if self.provider.wire_api == WireApi::Responses {
if let Some(path) = &*CODEX_RS_SSE_FIXTURE {
if self.provider.wire_api == WireApi::Responses
&& let Some(path) = &*CODEX_RS_SSE_FIXTURE {
warn!(path, "Streaming from fixture");
let stream = stream_from_fixture(
path,
@@ -194,7 +194,6 @@ impl ModelClient {
.boxed();
return Ok(wrap_stream(stream));
}
}
let backend = self
.backend

View File

@@ -303,7 +303,7 @@ pub(crate) struct SessionConfiguration {
provider: ModelProviderInfo,
/// If not specified, server will use its default model.
model: String,
pub(crate) model: String,
model_reasoning_effort: Option<ReasoningEffortConfig>,
model_reasoning_summary: ReasoningSummaryConfig,
@@ -315,7 +315,7 @@ pub(crate) struct SessionConfiguration {
user_instructions: Option<String>,
/// Base instructions override.
base_instructions: Option<String>,
pub(crate) base_instructions: Option<String>,
/// Compact prompt override.
compact_prompt: Option<String>,
@@ -335,7 +335,7 @@ pub(crate) struct SessionConfiguration {
cwd: PathBuf,
/// Set of feature flags for this session
features: Features,
pub(crate) features: Features,
// TODO(pakrym): Remove config from here
original_config_do_not_use: Arc<Config>,
@@ -588,8 +588,9 @@ impl Session {
config.active_profile.clone(),
);
// Create the mutable state for the Session.
let state = SessionState::new(session_configuration.clone());
let model_family = find_family_for_model(&session_configuration.model)
.unwrap_or_else(|| config.model_family.clone());
let state = SessionState::new(session_configuration.clone(), model_family);
let services = SessionServices {
mcp_connection_manager,
@@ -696,7 +697,6 @@ impl Session {
pub(crate) async fn update_settings(&self, updates: SessionSettingsUpdate) {
let mut state = self.state.lock().await;
state.session_configuration = state.session_configuration.apply(&updates);
}
@@ -942,12 +942,6 @@ impl Session {
self.send_raw_response_items(turn_context, items).await;
}
async fn prompt_for_turn(&self, turn_context: &TurnContext) -> Prompt {
let supports_chain = turn_context.client.supports_responses_api_chaining();
let mut state = self.state.lock().await;
state.prompt_for_turn(supports_chain)
}
fn reconstruct_history_from_rollout(
&self,
turn_context: &TurnContext,
@@ -1801,7 +1795,8 @@ pub(crate) async fn run_task(
// Construct the input that we will send to the model.
sess.record_conversation_items(&turn_context, &pending_input)
.await;
let prompt = sess.prompt_for_turn(&turn_context).await;
let mut state = sess.state.lock().await;
let prompt = state.prompt_for_turn();
let turn_input_messages: Vec<String> = {
prompt
@@ -2594,7 +2589,9 @@ mod tests {
session_source: SessionSource::Exec,
};
let state = SessionState::new(session_configuration.clone());
let model_family = find_family_for_model(&session_configuration.model)
.unwrap_or_else(|| config.model_family.clone());
let state = SessionState::new(session_configuration.clone(), model_family);
let services = SessionServices {
mcp_connection_manager: McpConnectionManager::default(),
@@ -2670,7 +2667,9 @@ mod tests {
session_source: SessionSource::Exec,
};
let state = SessionState::new(session_configuration.clone());
let model_family = find_family_for_model(&session_configuration.model)
.unwrap_or_else(|| config.model_family.clone());
let state = SessionState::new(session_configuration.clone(), model_family);
let services = SessionServices {
mcp_connection_manager: McpConnectionManager::default(),

View File

@@ -3,27 +3,39 @@
use codex_protocol::models::ResponseItem;
use crate::client_common::Prompt;
use crate::client_common::compute_full_instructions;
use crate::codex::SessionConfiguration;
use crate::conversation_history::ConversationHistory;
use crate::conversation_history::ResponsesApiChainState;
use crate::conversation_history::format_prompt_items;
use crate::model_family::ModelFamily;
use crate::protocol::RateLimitSnapshot;
use crate::protocol::TokenUsage;
use crate::protocol::TokenUsageInfo;
use crate::tools::spec::ToolsConfig;
use crate::tools::spec::ToolsConfigParams;
use crate::tools::spec::build_specs;
use crate::tools::spec::tools_metadata_for_prompt;
/// Persistent, session-scoped state previously stored directly on `Session`.
pub(crate) struct SessionState {
pub(crate) session_configuration: SessionConfiguration,
pub(crate) history: ConversationHistory,
pub(crate) latest_rate_limits: Option<RateLimitSnapshot>,
pub(crate) model_family: ModelFamily,
}
impl SessionState {
/// Create a new session state mirroring previous `State::default()` semantics.
pub(crate) fn new(session_configuration: SessionConfiguration) -> Self {
pub(crate) fn new(
session_configuration: SessionConfiguration,
model_family: ModelFamily,
) -> Self {
Self {
session_configuration,
history: ConversationHistory::new(),
latest_rate_limits: None,
model_family,
}
}
@@ -79,13 +91,36 @@ impl SessionState {
self.history.set_token_usage_full(context_window);
}
pub(crate) fn prompt_for_turn(&mut self, supports_responses_api_chaining: bool) -> Prompt {
pub(crate) fn prompt_for_turn(&mut self) -> Prompt {
let tools_config = ToolsConfig::new(&ToolsConfigParams {
model_family: &self.model_family,
features: &self.session_configuration.features,
});
let (tool_specs, _registry) = build_specs(&tools_config, None).build();
let tool_specs = tool_specs.into_iter().map(|c| c.spec).collect::<Vec<_>>();
let prompt_items = self.history.get_history_for_prompt();
let chain_state = self.history.responses_api_chain();
let (prompt, reset_chain) = build_prompt_from_items(prompt_items, chain_state.as_ref());
let (mut prompt, reset_chain) = build_prompt_from_items(prompt_items, chain_state.as_ref());
if reset_chain {
self.reset_responses_api_chain();
}
// Populate prompt fields that depend only on session state.
let (tools_json, has_freeform_apply_patch) =
tools_metadata_for_prompt(&tool_specs).expect("tool specs serialization");
format_prompt_items(&mut prompt.input, has_freeform_apply_patch);
let apply_patch_present = tool_specs.iter().any(|spec| spec.name() == "apply_patch");
let base_override = self.session_configuration.base_instructions.as_deref();
let instructions =
compute_full_instructions(base_override, &self.model_family, apply_patch_present)
.into_owned();
prompt.instructions = instructions;
prompt.tools = tools_json;
prompt.parallel_tool_calls = self.model_family.supports_parallel_tool_calls;
prompt
}
}