refactor: remove useless function call and refine tests.

xllm/core/distributed_runtime/llm_master.cpp

@@ -439,19 +439,16 @@ std::shared_ptr<Request> LLMMaster::generate_request(
     std::optional<Call*> call,
     OutputCallback callback) {
   Timer timer;
-  std::optional<std::string> prompt;
-  if (sp.has_tools()) {
-    prompt = chat_template_->apply(messages, sp.tools, sp.chat_template_kwargs);
-  } else {
-    prompt = chat_template_->apply(messages, sp.chat_template_kwargs);
-  }
 
+  std::optional<std::string> prompt;
+  prompt = chat_template_->apply(messages, sp.tools, sp.chat_template_kwargs);
   if (!prompt.has_value()) {
     CALLBACK_WITH_ERROR(StatusCode::INVALID_ARGUMENT,
                         "Failed to construct prompt from messages");
     LOG(ERROR) << "Failed to construct prompt from messages";
     return nullptr;
   }
+
   COUNTER_ADD(chat_template_latency_seconds, timer.elapsed_seconds());
 
   return generate_request(

xllm/core/distributed_runtime/rec_master.cpp

@@ -524,19 +524,17 @@ void RecMaster::handle_request(
   }
 
   Timer timer;
+
   std::optional<std::string> prompt;
-  if (sp.has_tools()) {
-    prompt = chat_template_->apply(messages, sp.tools, sp.chat_template_kwargs);
-  } else {
-    prompt = chat_template_->apply(messages, sp.chat_template_kwargs);
-  }
+  prompt = chat_template_->apply(messages, sp.tools, sp.chat_template_kwargs);
 
   if (!prompt.has_value()) {
     CALLBACK_WITH_ERROR(StatusCode::INVALID_ARGUMENT,
                         "Failed to construct prompt from messages");
     LOG(ERROR) << "Failed to construct prompt from messages";
     return;
   }
+
   COUNTER_ADD(chat_template_latency_seconds, timer.elapsed_seconds());
 
   schedule_request(std::move(sp),

xllm/core/framework/chat_template/jinja_chat_template.h

@@ -35,16 +35,12 @@ class JinjaChatTemplate {
 
   std::optional<std::string> apply(const ChatMessages& messages) const;
 
-  std::optional<std::string> apply(
-      const ChatMessages& messages,
-      const nlohmann::ordered_json& chat_template_kwargs) const;
-
   std::optional<std::string> apply(
       const ChatMessages& messages,
       const std::vector<xllm::JsonTool>& json_tools,
       const nlohmann::ordered_json& chat_template_kwargs) const;
 
-  // expose this function for testing
+ protected:
   // apply the template to the values in the json object
   std::optional<std::string> apply(nlohmann::ordered_json& messages) const;
 
@@ -53,7 +49,10 @@ class JinjaChatTemplate {
       const nlohmann::ordered_json& tools,
       const nlohmann::ordered_json& chat_template_kwargs) const;
 
- private:
+  std::optional<std::string> apply(
+      const ChatMessages& messages,
+      const nlohmann::ordered_json& chat_template_kwargs) const;
+
   nlohmann::ordered_json get_mm_content(const MMContentVec& vec) const;
 
  private:

xllm/core/framework/chat_template/jinja_chat_template_test.cpp

@@ -19,6 +19,14 @@ limitations under the License.
 
 namespace xllm {
 
+class TestableJinjaChatTemplate : public JinjaChatTemplate {
+ public:
+  TestableJinjaChatTemplate(const TokenizerArgs& args)
+      : JinjaChatTemplate(args) {}
+
+  using JinjaChatTemplate::apply;
+};
+
 TEST(JinjaChatTemplate, OpenChatModel) {
   // clang-format off
   const std::string template_str =
@@ -46,7 +54,7 @@ TEST(JinjaChatTemplate, OpenChatModel) {
   args.chat_template(template_str);
   args.bos_token("");
   args.eos_token("<|end_of_turn|>");
-  JinjaChatTemplate template_(args);
+  TestableJinjaChatTemplate template_(args);
   auto result = template_.apply(messages);
   ASSERT_TRUE(result.has_value());
 

xllm/core/framework/request/request_params.h

@@ -137,8 +137,8 @@ struct RequestParams {
 
   // JSON-based tools (replacing proto_tools)
   std::vector<xllm::JsonTool> tools;
+
   std::string tool_choice = "auto";
-  bool has_tools() const { return !tools.empty(); }
 
   bool offline = false;