refac

main.py

@@ -117,6 +117,7 @@ async def get_models():
 @app.post("/v1/chat/completions")
 async def generate_openai_chat_completion(form_data: OpenAIChatCompletionForm):
     user_message = get_last_user_message(form_data.messages)
+    messages = [message.model_dump() for message in form_data.messages]
 
     if form_data.model not in app.state.PIPELINES:
         return HTTPException(
@@ -133,7 +134,7 @@ async def generate_openai_chat_completion(form_data: OpenAIChatCompletionForm):
             def stream_content():
                 res = get_response(
                     user_message,
-                    messages=form_data.messages,
+                    messages=messages,
                     body=form_data.model_dump(),
                 )
 
@@ -186,7 +187,7 @@ async def generate_openai_chat_completion(form_data: OpenAIChatCompletionForm):
         else:
             res = get_response(
                 user_message,
-                messages=form_data.messages,
+                messages=messages,
                 body=form_data.model_dump(),
             )
             logging.info(f"stream:false:{res}")

pipelines/examples/applescript_pipeline.py

@@ -25,7 +25,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/azure_openai_pipeline.py

@@ -21,7 +21,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/haystack_pipeline.py

@@ -79,7 +79,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom RAG pipeline.
         # Typically, you would retrieve relevant information from your knowledge base and synthesize it to generate a response.

pipelines/examples/llama_cpp_pipeline.py

@@ -30,7 +30,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/llamaindex_ollama_github_pipeline.py

@@ -70,7 +70,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom RAG pipeline.
         # Typically, you would retrieve relevant information from your knowledge base and synthesize it to generate a response.

pipelines/examples/llamaindex_ollama_pipeline.py

@@ -30,7 +30,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom RAG pipeline.
         # Typically, you would retrieve relevant information from your knowledge base and synthesize it to generate a response.

pipelines/examples/llamaindex_pipeline.py

@@ -25,7 +25,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom RAG pipeline.
         # Typically, you would retrieve relevant information from your knowledge base and synthesize it to generate a response.

pipelines/examples/mlx_pipeline.py

@@ -73,7 +73,7 @@ class Pipeline:
                 print(f"Failed to terminate subprocess: {e}")
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/ollama_pipeline.py

@@ -21,7 +21,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/openai_pipeline.py

@@ -21,7 +21,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/pipeline_example.py

@@ -20,7 +20,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/examples/python_code_pipeline.py

@@ -31,7 +31,7 @@ class Pipeline:
             return e.output.strip(), e.returncode
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

pipelines/ollama_pipeline.py

@@ -21,7 +21,7 @@ class Pipeline:
         pass
 
     def get_response(
-        self, user_message: str, messages: List[OpenAIChatMessage], body: dict
+        self, user_message: str, messages: List[dict], body: dict
     ) -> Union[str, Generator, Iterator]:
         # This is where you can add your custom pipelines like RAG.'
         print(f"get_response:{__name__}")

schemas.py

@@ -12,6 +12,6 @@ class OpenAIChatMessage(BaseModel):
 class OpenAIChatCompletionForm(BaseModel):
     stream: bool = True
     model: str
-    messages: List[OpenAIChatMessage]
+    messages: List[dict]
 
     model_config = ConfigDict(extra="allow")

utils.py

@@ -22,7 +22,7 @@ def stream_message_template(model: str, message: str):
     }
 
 
-def get_last_user_message(messages: List[OpenAIChatMessage]) -> str:
+def get_last_user_message(messages: List[dict]) -> str:
     for message in reversed(messages):
         if message.role == "user":
             return message.content