Merge branch 'enable_thinking' of github.com:matteoserva/llama.cpp into enable_thinking

Author: matteoserva

common/chat.cpp

@@ -1583,7 +1583,7 @@ static common_chat_msg common_chat_parse_hermes_2_pro(const std::string& input,
 
 static common_chat_params common_chat_params_init_without_tools(const common_chat_template & tmpl, const struct templates_params & inputs) {
     common_chat_params data;
-    data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt,inputs.extra_context);
+    data.prompt = apply(tmpl, inputs.messages, inputs.tools.empty() ? json() : inputs.tools, inputs.add_generation_prompt, inputs.extra_context);
     data.format = COMMON_CHAT_FORMAT_CONTENT_ONLY;
     data.grammar_lazy = false;
     if (!inputs.json_schema.is_null()) {
@@ -1615,7 +1615,7 @@ static common_chat_params common_chat_templates_apply_jinja(
     params.grammar = inputs.grammar;
     params.now = inputs.now;
 
-    for (auto el: inputs.chat_template_kwargs) {
+    for (auto el : inputs.chat_template_kwargs) {
         params.extra_context[el.first] = json::parse(el.second);
     }
 

common/chat.h

@@ -74,7 +74,7 @@ struct common_chat_templates_inputs {
     bool parallel_tool_calls = false;
     bool extract_reasoning     = true;
     std::chrono::system_clock::time_point now = std::chrono::system_clock::now();
-    std::map<std::string,std::string> chat_template_kwargs;
+    std::map<std::string, std::string> chat_template_kwargs;
 };
 
 struct common_chat_params {

common/common.h

@@ -375,7 +375,7 @@ struct common_params {
     std::string ssl_file_key  = "";                                                                         // NOLINT
     std::string ssl_file_cert = "";                                                                         // NOLINT
 
-    std::map<std::string,std::string> default_template_kwargs;
+    std::map<std::string, std::string> default_template_kwargs;
 
     // "advanced" endpoints are disabled by default for better security
     bool webui            = true;

tools/server/utils.hpp

@@ -584,7 +584,7 @@ static json oaicompat_completion_params_parse(
     const json & body, /* openai api json semantics */
     bool use_jinja,
     common_reasoning_format reasoning_format,
-    const std::map<std::string,std::string> default_template_kwargs,
+    const std::map<std::string, std::string> & default_template_kwargs,
     const struct common_chat_templates * tmpls,
     bool allow_non_text,
     std::vector<raw_buffer> & out_files)
@@ -729,7 +729,7 @@ static json oaicompat_completion_params_parse(
     inputs.add_generation_prompt = json_value(body, "add_generation_prompt", true);
 
     auto chat_template_kwargs_object = json_value(body, "chat_template_kwargs", json::object());
-    for (const auto & item: default_template_kwargs) {
+    for (const auto & item : default_template_kwargs) {
         inputs.chat_template_kwargs[item.first] = item.second;
     }
     for (const auto & item : chat_template_kwargs_object.items()) {