Update URL path setting

backend/app/main.py

@@ -8,7 +8,7 @@
 from starlette.middleware.sessions import SessionMiddleware
 from fastapi.middleware.cors import CORSMiddleware
 from contextlib import asynccontextmanager
-from backend.app.core.init_settings import args, global_settings
+from backend.app.core.init_settings import args
 from backend.app.core.constants import API_BASE_URL
 from backend.app.api.v1.endpoints import (
     user,
@@ -54,7 +54,7 @@ async def lifespan(app: FastAPI):
 # Set Middleware
 # Define the allowed origins
 origins = [
-    global_settings.API_BASE_URL,
+    API_BASE_URL,
     "http://localhost",
     "http://localhost:5000",
 ]

frontend/gradio/text/event_listeners.py

@@ -4,6 +4,7 @@
 import time
 import gradio as gr
 import json
+from urllib.parse import urljoin
 from typing import List, Optional, Tuple
 from backend.app.core.constants import API_BASE_URL, MEMORY_WINDOW_SIZE
 from frontend.gradio.utils import add_message_to_db, start_chat
@@ -60,7 +61,7 @@ def enable_vote_buttons_arena():
 def add_rating_to_db(state: gr.State, rating_type: str):
     chat_id = state.value['current_chat_id']
     if chat_id:
-        rating_api_endpoint = API_BASE_URL + "api/v1/ratings/"
+        rating_api_endpoint = urljoin(API_BASE_URL, "api/v1/ratings/")
         rating_data = {
             "chat_id": chat_id,
             "rating_type": rating_type
@@ -106,7 +107,7 @@ def llm_text_completion_stream(state: gr.State, model_name: str, history: List[T
     model_endpoint = state.value['model_map'][model_name]
 
     # Call the LLM model with streaming
-    llm_api_endpoint = API_BASE_URL + "api/v1/llm/generation/stream/text"
+    llm_api_endpoint = urljoin(API_BASE_URL, "api/v1/llm/generation/stream/text")
     llm_data = {
         "user_id": user_id,
         "api_endpoint": model_endpoint,
@@ -139,7 +140,7 @@ def llm_text_completion_stream(state: gr.State, model_name: str, history: List[T
                 time.sleep(0.01)  # Control the streaming speed for the interface
 
     # Add Message data(user) into DB
-    message_api_endpoint = API_BASE_URL + "api/v1/messages/"
+    message_api_endpoint = urljoin(API_BASE_URL, "api/v1/messages/")
     message_data_user = {
         "user_id": user_id,
         "content": user_message,
@@ -182,7 +183,7 @@ async def llm_text_completion_stream_async(
     state.value['current_chat_id'] = chat_id
 
     # Call the LLM model with streaming
-    llm_api_endpoint = API_BASE_URL + "api/v1/llm/generation/stream/text"
+    llm_api_endpoint = urljoin(API_BASE_URL, "api/v1/llm/generation/stream/text")
     llm_params = {
         "system_prompt": system_prompt,
         "temperature": temperature,
@@ -243,7 +244,7 @@ async def llm_text_completion_memory_stream_async(
     state.value['current_chat_id'] = chat_id
 
     # Call the LLM model with streaming
-    llm_api_endpoint = API_BASE_URL + "api/v1/llm/generation/stream/text/memory"
+    llm_api_endpoint = urljoin(API_BASE_URL, "api/v1/llm/generation/stream/text/memory")
     llm_params = {
         "system_prompt": system_prompt,
         "temperature": temperature,
@@ -310,7 +311,7 @@ async def llm_text_completion_stream_arena(
     state.value['current_chat_id'] = chat_id
 
     # Call the LLM model with streaming
-    llm_api_endpoint = API_BASE_URL + "api/v1/llm/generation/stream/text"
+    llm_api_endpoint = urljoin(API_BASE_URL, "api/v1/llm/generation/stream/text")
     llm_params_1 = {
         "system_prompt": system_prompt_1,
         "temperature": temperature_1,
@@ -401,7 +402,7 @@ async def llm_text_completion_memory_stream_arena(
     state.value['current_chat_id'] = chat_id
 
     # Call the LLM model with streaming
-    llm_api_endpoint = API_BASE_URL + "api/v1/llm/generation/stream/text/memory"
+    llm_api_endpoint = urljoin(API_BASE_URL, "api/v1/llm/generation/stream/text/memory")
     llm_params_1 = {
         "system_prompt": system_prompt_1,
         "temperature": temperature_1,

frontend/gradio/utils.py

@@ -1,9 +1,8 @@
-# Reference: https://github.com/lm-sys/FastChat/fastchat/serve/gradio_web_server.py
 from typing import List, Dict, Optional, Tuple, Any
 import aiohttp
 import requests
 import gradio as gr
-from backend.app.core.constants import API_BASE_URL
+from urllib.parse import urljoin
 from backend.app.core.constants import (
     API_BASE_URL,
     TEXT_API_QUOTA_LIMIT
@@ -23,7 +22,8 @@ def get_client_ip(request: gr.Request) -> str:
 def fetch_available_models(llm_type: str) -> Dict[str, str]:
     """Fetch available large language models from the base API URL."""
     models = {}
-    response = requests.get(API_BASE_URL + f"api/v1/llms/type/{llm_type}")
+    url = urljoin(API_BASE_URL, f"api/v1/llms/type/{llm_type}")
+    response = requests.get(url)
     assert response.status_code == 200
     model_info = response.json()
 
@@ -33,25 +33,24 @@ def fetch_available_models(llm_type: str) -> Dict[str, str]:
     return models
 
 async def start_chat(user_id: str, mode: str) -> str:
-    chat_api_endpoint = API_BASE_URL + "api/v1/chats/async"
+    url = urljoin(API_BASE_URL, "api/v1/chats/async")
     chat_data = {"user_id": user_id, "mode": mode}
 
     async with aiohttp.ClientSession() as session:
-        async with session.post(chat_api_endpoint, json=chat_data) as response:
+        async with session.post(url, json=chat_data) as response:
             assert response.status == 200
             data = await response.json()
             return data["id"]
 
-
 async def add_message_to_db(session: aiohttp.ClientSession, chat_id: str, content: str, message_type: str, origin: str):
-    message_api_endpoint = API_BASE_URL + "api/v1/messages/async"
+    url = urljoin(API_BASE_URL, "api/v1/messages/async")
     message_data = {
         "chat_id": chat_id,
         "content": content,
         "message_type": message_type,
         "origin": origin
     }
-    async with session.post(message_api_endpoint, json=message_data) as message_response:
+    async with session.post(url, json=message_data) as message_response:
         assert message_response.status == 200
 
 def load_terms_of_use_js() -> str:
@@ -118,21 +117,21 @@ def load_demo(url_params: gr.JSON, request: gr.Request) -> Tuple[Optional[Dict[s
     client_ip = get_client_ip(request)
 
     # Check if the client is first time user
-    user_ip_api_endpoint = base_api_url + f"api/v1/users/ip/{client_ip}"
+    user_ip_api_endpoint = urljoin(base_api_url, f"api/v1/users/ip/{client_ip}")
     user_ip_response = requests.get(user_ip_api_endpoint)
     first_time_user = user_ip_response.status_code != 200
 
     if first_time_user:
         # Create User Profile based on IP address
-        user_api_endpoint = base_api_url + "api/v1/users/"
+        user_api_endpoint = urljoin(base_api_url, "api/v1/users/")
         user_data = {"ip_address": client_ip}
         user_response = requests.post(user_api_endpoint, json=user_data)
         user_response.raise_for_status()
         user_id = user_response.json()["id"]
         user_name = user_response.json()["name"]
 
         # Create a new Quota(Text and Image API) for a the new user
-        quota_api_endpoint = base_api_url + "api/v1/quotas/"
+        quota_api_endpoint = urljoin(base_api_url, "api/v1/quotas/")
         for llm_model_type in api_quota_map:
             quota_limit, resource, _, _ = api_quota_map[llm_model_type]
             quota_data = {"users": [{"id": user_id}], "quota_limit": quota_limit, "resource": resource}
@@ -162,7 +161,7 @@ def load_demo(url_params: gr.JSON, request: gr.Request) -> Tuple[Optional[Dict[s
         * See more in the [Gradio](https://www.gradio.app/) 📖
 
         ## 👇 Choose any models for {task}:
-    """.format(user_name = user_name, task = task)
+    """.format(user_name=user_name, task=task)
 
     markdown_update = gr.Markdown(notice_markdown)
     state = gr.State({
@@ -174,10 +173,9 @@ def load_demo(url_params: gr.JSON, request: gr.Request) -> Tuple[Optional[Dict[s
         "history": []
     })
     if is_arena_ui:
-        model_1_dropdown_update = gr.Dropdown(label= 'LLM', choices=model_names, value=selected_model_1, visible=True, show_label=False)
-        model_2_dropdown_update = gr.Dropdown(label= 'LLM', choices=model_names, value=selected_model_2, visible=True, show_label=False)
+        model_1_dropdown_update = gr.Dropdown(label='LLM', choices=model_names, value=selected_model_1, visible=True, show_label=False)
+        model_2_dropdown_update = gr.Dropdown(label='LLM', choices=model_names, value=selected_model_2, visible=True, show_label=False)
         return state, markdown_update, model_1_dropdown_update, model_2_dropdown_update
     else:
-        model_dropdown_update = gr.Dropdown(label= 'LLM', choices=model_names, value=selected_model_1, visible=True)
+        model_dropdown_update = gr.Dropdown(label='LLM', choices=model_names, value=selected_model_1, visible=True)
         return state, markdown_update, model_dropdown_update
-