handle prelight as a single endpoint

Author: tikikun

controllers/llamaCPP.cc

@@ -149,7 +149,7 @@ void llamaCPP::warmupModel() {
   return;
 }
 
-void llamaCPP::chatCompletionPrelight(
+void llamaCPP::handlePrelight(
     const HttpRequestPtr &req,
     std::function<void(const HttpResponsePtr &)> &&callback) {
   auto resp = drogon::HttpResponse::newHttpResponse();

controllers/llamaCPP.h

@@ -2530,18 +2530,17 @@ class llamaCPP : public drogon::HttpController<llamaCPP> {
 
   // Openai compatible path
   ADD_METHOD_TO(llamaCPP::chatCompletion, "/v1/chat/completions", Post);
-  ADD_METHOD_TO(llamaCPP::chatCompletionPrelight, "/v1/chat/completions",
-                Options);
+  ADD_METHOD_TO(llamaCPP::handlePrelight, "/v1/chat/completions", Options);
 
   ADD_METHOD_TO(llamaCPP::embedding, "/v1/embeddings", Post);
+  ADD_METHOD_TO(llamaCPP::handlePrelight, "/v1/embeddings", Options);
 
   // PATH_ADD("/llama/chat_completion", Post);
   METHOD_LIST_END
   void chatCompletion(const HttpRequestPtr &req,
                       std::function<void(const HttpResponsePtr &)> &&callback);
-  void chatCompletionPrelight(
-      const HttpRequestPtr &req,
-      std::function<void(const HttpResponsePtr &)> &&callback);
+  void handlePrelight(const HttpRequestPtr &req,
+                      std::function<void(const HttpResponsePtr &)> &&callback);
   void embedding(const HttpRequestPtr &req,
                  std::function<void(const HttpResponsePtr &)> &&callback);
   void loadModel(const HttpRequestPtr &req,