fix: fix chat and optimize deploy

source/infrastructure/package.json

@@ -6,7 +6,7 @@
     "clobber": "npx projen clobber",
     "compile": "npx projen compile",
     "default": "npx projen default",
-    "deploy": "npx cdk deploy",
+    "deploy": "npx cdk deploy --all",
     "destroy": "npx projen destroy",
     "diff": "npx projen diff",
     "eject": "npx projen eject",
@@ -100,4 +100,4 @@
     }
   },
   "//": "~~ Generated by projen. To modify, edit .projenrc.js and run \"npx projen\"."
-}
+}

source/lambda/online/common_logic/langchain_integration/chat_models/__init__.py

@@ -1,7 +1,9 @@
 """
 chat models build in command pattern
 """
-from common_logic.common_utils.constant import LLMModelType,ModelProvider
+
+from common_logic.common_utils.constant import LLMModelType, ModelProvider
+
 from ..model_config import MODEL_CONFIGS
 
 
@@ -10,7 +12,7 @@ class ModeMixins:
     def convert_messages_role(messages: list[dict], role_map: dict):
         """
         Args:
-            messages (list[dict]): 
+            messages (list[dict]):
             role_map (dict): {"current_role":"targe_role"}
 
         Returns:
@@ -20,18 +22,22 @@ def convert_messages_role(messages: list[dict], role_map: dict):
         new_messages = []
         for message in messages:
             message = {**message}
-            role = message['role']
+            role = message["role"]
             assert role in valid_roles, (role, valid_roles, messages)
-            message['role'] = role_map[role]
+            message["role"] = role_map[role]
             new_messages.append(message)
         return new_messages
 
 
 class ModelMeta(type):
     def __new__(cls, name, bases, attrs):
         new_cls = type.__new__(cls, name, bases, attrs)
-        if name == "Model" or new_cls.model_id is None \
-            or name.endswith("BaseModel") or name.lower().endswith("basemodel"):
+        if (
+            name == "Model"
+            or new_cls.model_id is None
+            or name.endswith("BaseModel")
+            or name.lower().endswith("basemodel")
+        ):
             return new_cls
         new_cls.model_map[new_cls.get_model_id()] = new_cls
         return new_cls
@@ -48,38 +54,45 @@ class Model(ModeMixins, metaclass=ModelMeta):
     @classmethod
     def create_model(cls, model_kwargs=None, **kwargs):
         raise NotImplementedError
-    
+
     @classmethod
-    def get_model_id(cls):
-        return f"{cls.model_id}__{cls.model_provider}"
+    def get_model_id(cls, model_id=None, model_provider=None):
+        if model_id is None:
+            model_id = cls.model_id
+        if model_provider is None:
+            model_provider = cls.model_provider
+        return f"{model_id}__{model_provider}"
+
     @classmethod
     def get_model(cls, model_id, model_kwargs=None, **kwargs):
+        model_provider = kwargs["provider"]
         # dynamic load module
         _load_module(model_id)
-        return cls.model_map[cls.get_model_id()].create_model(model_kwargs=model_kwargs, **kwargs)
+        return cls.model_map[cls.get_model_id(model_id=model_id, model_provider=model_provider)].create_model(
+            model_kwargs=model_kwargs, **kwargs
+        )
 
     @classmethod
     def model_id_to_class_name(cls, model_id: str) -> str:
         """Convert model ID to a valid Python class name.
-        
+
         Examples:
             anthropic.claude-3-haiku-20240307-v1:0 -> Claude3Haiku20240307V1Model
         """
         # Remove version numbers and vendor prefixes
-        name = str(model_id).split(':')[0]
-        name = name.split('.')[-1]
-        parts = name.replace('_', '-').split('-')
+        name = str(model_id).split(":")[0]
+        name = name.split(".")[-1]
+        parts = name.replace("_", "-").split("-")
 
         cleaned_parts = []
         for part in parts:
             if any(c.isdigit() for c in part):
-                cleaned = ''.join(c.upper() if i == 0 or part[i-1] in '- ' else c
-                                  for i, c in enumerate(part))
+                cleaned = "".join(c.upper() if i == 0 or part[i - 1] in "- " else c for i, c in enumerate(part))
             else:
                 cleaned = part.capitalize()
             cleaned_parts.append(cleaned)
 
-        return ''.join(cleaned_parts) + "Model"
+        return "".join(cleaned_parts) + "Model"
 
     @classmethod
     def create_for_model(cls, model_id: str):
@@ -95,13 +108,14 @@ def create_for_model(cls, model_id: str):
                 "default_model_kwargs": config.default_model_kwargs,
                 "enable_any_tool_choice": config.enable_any_tool_choice,
                 "enable_prefill": config.enable_prefill,
-            }
+            },
         )
         return model_class
 
 
 def _import_bedrock_models():
     from .bedrock_models import model_classes
+
     # from .bedrock_models import (
     #     Claude2,
     #     ClaudeInstance,
@@ -119,20 +133,15 @@ def _import_bedrock_models():
 
 
 def _import_openai_models():
-    from .openai_models import (
-        ChatGPT35,
-        ChatGPT4Turbo,
-        ChatGPT4o
-    )
+    from .openai_models import ChatGPT4o, ChatGPT4Turbo, ChatGPT35
+
 
 def _import_dmaa_models():
     from . import dmaa_models
-
 
 
 def _load_module(model_id):
-    assert model_id in MODEL_MODULE_LOAD_FN_MAP, (
-        model_id, MODEL_MODULE_LOAD_FN_MAP)
+    assert model_id in MODEL_MODULE_LOAD_FN_MAP, (model_id, MODEL_MODULE_LOAD_FN_MAP)
     MODEL_MODULE_LOAD_FN_MAP[model_id]()
 
 
@@ -167,5 +176,5 @@ def _load_module(model_id):
     LLMModelType.CLAUDE_3_5_SONNET_APAC: _import_bedrock_models,
     LLMModelType.CLAUDE_3_HAIKU_APAC: _import_bedrock_models,
     LLMModelType.LLAMA3_1_70B_INSTRUCT_US: _import_bedrock_models,
-    LLMModelType.QWEN25_INSTRUCT_72B_AWQ: _import_dmaa_models
+    LLMModelType.QWEN25_INSTRUCT_72B_AWQ: _import_dmaa_models,
 }