From 51ae40306a9a607e935881a3fdfbccc46505cc84 Mon Sep 17 00:00:00 2001
From: Shuaiyi Zhang <576893949@qq.com>
Date: Tue, 8 Jul 2025 23:03:38 +0800
Subject: [PATCH] [router] forward stream_options in request (#7860)

---
 sgl-router/benches/request_processing.rs  |  3 +++
 sgl-router/src/openai_api_types.rs        | 14 ++++++++++++++
 sgl-router/tests/benchmark_integration.rs |  4 ++++
 3 files changed, 21 insertions(+)

diff --git a/sgl-router/benches/request_processing.rs b/sgl-router/benches/request_processing.rs
index 5761e8e1c..c2cee90d5 100644
--- a/sgl-router/benches/request_processing.rs
+++ b/sgl-router/benches/request_processing.rs
@@ -59,6 +59,7 @@ fn create_sample_chat_completion_request() -> ChatCompletionRequest {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         stop: None,
         presence_penalty: Some(0.0),
         frequency_penalty: Some(0.0),
@@ -86,6 +87,7 @@ fn create_sample_completion_request() -> CompletionRequest {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         logprobs: None,
         echo: false,
         stop: None,
@@ -130,6 +132,7 @@ fn create_large_chat_completion_request() -> ChatCompletionRequest {
         top_p: Some(0.95),
         n: Some(1),
         stream: false,
+        stream_options: None,
         stop: None,
         presence_penalty: Some(0.1),
         frequency_penalty: Some(0.1),
diff --git a/sgl-router/src/openai_api_types.rs b/sgl-router/src/openai_api_types.rs
index 808f8b46f..9870fd06b 100644
--- a/sgl-router/src/openai_api_types.rs
+++ b/sgl-router/src/openai_api_types.rs
@@ -52,6 +52,10 @@ pub struct CompletionRequest {
     #[serde(default)]
     pub stream: bool,
 
+    /// Options for streaming response
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub stream_options: Option<StreamOptions>,
+
     /// Include the log probabilities on the logprobs most likely tokens
     #[serde(skip_serializing_if = "Option::is_none")]
     pub logprobs: Option<u32>,
@@ -132,6 +136,10 @@ pub struct ChatCompletionRequest {
     #[serde(default)]
     pub stream: bool,
 
+    /// Options for streaming response
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub stream_options: Option<StreamOptions>,
+
     /// Up to 4 sequences where the API will stop generating further tokens
     #[serde(skip_serializing_if = "Option::is_none")]
     pub stop: Option<StringOrArray>,
@@ -258,6 +266,12 @@ pub struct ImageUrl {
     pub detail: Option<String>, // "auto", "low", or "high"
 }
 
+#[derive(Debug, Clone, Deserialize, Serialize)]
+pub struct StreamOptions {
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub include_usage: Option<bool>,
+}
+
 #[derive(Debug, Clone, Deserialize, Serialize)]
 #[serde(tag = "type")]
 pub enum ResponseFormat {
diff --git a/sgl-router/tests/benchmark_integration.rs b/sgl-router/tests/benchmark_integration.rs
index a57714b0d..b21c93fcf 100644
--- a/sgl-router/tests/benchmark_integration.rs
+++ b/sgl-router/tests/benchmark_integration.rs
@@ -50,6 +50,7 @@ fn test_benchmark_request_creation() {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         stop: None,
         presence_penalty: Some(0.0),
         frequency_penalty: Some(0.0),
@@ -75,6 +76,7 @@ fn test_benchmark_request_creation() {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         logprobs: None,
         echo: false,
         stop: None,
@@ -143,6 +145,7 @@ fn test_benchmark_request_adaptation() {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         stop: None,
         presence_penalty: Some(0.0),
         frequency_penalty: Some(0.0),
@@ -168,6 +171,7 @@ fn test_benchmark_request_adaptation() {
         top_p: Some(1.0),
         n: Some(1),
         stream: false,
+        stream_options: None,
         logprobs: None,
         echo: false,
         stop: None,