add multi lora support

src/engine.py

@@ -125,6 +125,25 @@ def __init__(self, vllm_engine):
         super().__init__(vllm_engine)
         self.served_model_name = os.getenv("OPENAI_SERVED_MODEL_NAME_OVERRIDE") or self.engine_args.model
         self.response_role = os.getenv("OPENAI_RESPONSE_ROLE") or "assistant"
+
+        adapters = os.getenv("LORA_MODULES", [])
+
+        try:
+            adapters = json.loads(adapters)
+        except Exception as e:
+            logging.info(f"---Initialized adapter json load error: {e}")
+            adapters = []
+
+        self.lora_adapters = []
+        for adapter in adapters:
+            try:
+                lora:LoRAModulePath = LoRAModulePath(**adapter)
+                self.lora_adapters.append(lora)
+                logging.info(f"---Initialized adapter: {adapter}")
+            except Exception as e:
+                logging.info(f"---Initialized adapter not worked: {e}")
+                continue
+
         asyncio.run(self._initialize_engines())
         self.raw_openai_output = bool(int(os.getenv("RAW_OPENAI_OUTPUT", 1)))
 
@@ -134,19 +153,11 @@ async def _initialize_engines(self):
             BaseModelPath(name=self.engine_args.model, model_path=self.engine_args.model)
         ]
 
-        lora_modules = os.getenv('LORA_MODULES', None)
-        if lora_modules is not None:
-            try:
-                lora_modules = json.loads(lora_modules)
-                lora_modules = [LoRAModulePath(**lora_modules)]
-            except:
-                lora_modules = None
-
         self.serving_models = OpenAIServingModels(
             engine_client=self.llm,
             model_config=self.model_config,
             base_model_paths=self.base_model_paths,
-            lora_modules=None,
+            lora_modules=self.lora_adapters,
             prompt_adapters=None,
         )