39 lines
1.3 KiB
Python
39 lines
1.3 KiB
Python
from functools import partial
|
|
from typing import List
|
|
|
|
from fastapi import APIRouter
|
|
from pydantic import BaseModel, Field
|
|
|
|
from memgpt.models.pydantic_models import LLMConfigModel
|
|
from memgpt.server.rest_api.auth_token import get_current_user
|
|
from memgpt.server.rest_api.interface import QueuingInterface
|
|
from memgpt.server.server import SyncServer
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
class ListModelsResponse(BaseModel):
|
|
models: List[LLMConfigModel] = Field(..., description="List of model configurations.")
|
|
|
|
|
|
def setup_models_index_router(server: SyncServer, interface: QueuingInterface, password: str):
|
|
partial(partial(get_current_user, server), password)
|
|
|
|
@router.get("/models", tags=["models"], response_model=ListModelsResponse)
|
|
async def list_models():
|
|
# Clear the interface
|
|
interface.clear()
|
|
|
|
# currently, the server only supports one model, however this may change in the future
|
|
llm_config = LLMConfigModel(
|
|
model=server.server_llm_config.model,
|
|
model_endpoint=server.server_llm_config.model_endpoint,
|
|
model_endpoint_type=server.server_llm_config.model_endpoint_type,
|
|
model_wrapper=server.server_llm_config.model_wrapper,
|
|
context_window=server.server_llm_config.context_window,
|
|
)
|
|
|
|
return ListModelsResponse(models=[llm_config])
|
|
|
|
return router
|