#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Gradio UI – v2.1 (Leaderboard · Data Viewer · Prompt-to-Leaderboard)
"""
from __future__ import annotations
from pathlib import Path
import gradio as gr
# ---- Tab 组件 ----
from tabs.leaderboard_tab import create_leaderboard_tab
from tabs.data_viewer_tab import create_data_viewer_tab
from tabs.data_viewer_side_by_side_tab import create_data_viewer_side_by_side_tab
# ---------------------------------------------------------------------------
# UI
# ---------------------------------------------------------------------------
with gr.Blocks(title="DeepResearch Bench") as demo:
# ========= 全局 CSS(仅作用于自定义标题 & 简介) =========
gr.HTML("""
""")
# ========= 顶部标题 & 简介(不使用 Markdown 标题语法) =========
gr.HTML("""
DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents
The research aims to comprehensively evaluate the capabilities of Deep Research Agents.
Code |
Website |
Paper |
Eval Dataset |
Total models: 16 | Last Update: 28 May 2025
""")
# ========= 主 Tabs =========
with gr.Tabs():
create_leaderboard_tab() # 🏆 Leaderboard
create_data_viewer_side_by_side_tab()
create_data_viewer_tab() # 🔍 Data Viewer
with gr.Tab("💬Prompt-to-Leaderboard"):
gr.Markdown(
"""
🚧 **Prompt-to-Leaderboard** module not implemented yet.
Planned: inspect how individual prompts affect overall model ranking.
"""
)
# ---------------------------------------------------------------------------
# Entrypoint
# ---------------------------------------------------------------------------
if __name__ == "__main__":
demo.launch()