Update app.py

app.py

@@ -39,56 +39,137 @@ def get_leaderboard_df(score_path):
 
 leaderboard_df = get_leaderboard_df("score.json")
 
+import gradio as gr
+import pandas as pd
 
-# Function to update the table based on search query
-def filter_and_search(cols: list[str], search_query: str, agg: str):
-    print("filter")
-    df = leaderboard_df
-    search_terms = "Model"
-    if len(search_query) > 0:
-        search_terms = search_query.split(";")
-        search_terms = [term.strip().lower() for term in search_terms]
-        pattern = "|".join(search_terms)
+# 示例：你已有的 dataframe
+# leaderboard_df = pd.read_csv("your_data.csv")
+
+# 示例任务列字典
+TASKS = {
+    "VQA": ["VQA"],
+    "QA": ["QA"],
+    "VQA Reasoning": ["VQA_Reasoning"],
+    "Reason": ["Reason"],  # 请确保这个列名正确
+    "Embodied Grounding": ["Embodied Grounding"],
+    "GUI Grounding": ["Gui Grounding"],
+}
+
+# 筛选函数：只根据模型名称关键词搜索
+def filter_and_search(search_query: str, task_name: str):
+    df = leaderboard_df.copy()
+    task_cols = TASKS[task_name]
+    score_col = task_cols[0]
+
+    df[score_col] = pd.to_numeric(df[score_col], errors='coerce')
+    df = df.sort_values(by=score_col, ascending=False, na_position='last')
+
+    if search_query.strip():
+        terms = [term.strip().lower() for term in search_query.split(";")]
+        pattern = "|".join(terms)
         df = df[df["Model"].str.lower().str.contains(pattern, regex=True)]
-        # Drop any columns which are all NaN
-        df = df.dropna(how="all", axis=1)
-
-    if len(cols) > 0:
-        index_cols = list(leaderboard_df.columns[:1])
-        new_cols = index_cols + cols
-        df = df.copy()[new_cols]
-        df = df.copy().dropna(how="all", axis=0, subset=[c for c in df.columns if c in cols])
-
-        df[cols] = df[cols].apply(pd.to_numeric, errors='coerce')
-        df = df.sort_values(by=cols, ascending=False, na_position='last')
-        df[cols] = df[cols].astype(str)
-    return df
 
+    return df[["Model"] + task_cols]
+
+# Gradio UI 构建
+with gr.Blocks() as demo:
+    gr.HTML("<h2>Leaderboard</h2>")
+    with gr.Column():
+        gr.Markdown("Search and view results for each task.", elem_classes="markdown-text")
+
+    with gr.Tabs(elem_classes="tabs-buttons") as tabs:
+        for task_name, task_cols in TASKS.items():
+            with gr.TabItem(task_name):
+                # 初始数据：按得分降序
+                sub_df = leaderboard_df[["Model"] + task_cols].copy()
+                sub_df[task_cols[0]] = pd.to_numeric(sub_df[task_cols[0]], errors="coerce")
+                sub_df = sub_df.sort_values(by=task_cols[0], ascending=False, na_position="last")
+
+                with gr.Row():
+                    search_bar = gr.Textbox(placeholder="Search model name...", show_label=False)
+
+                with gr.Group():
+                    table = gr.Dataframe(
+                        value=sub_df,
+                        wrap=True,
+                        column_widths=[400] + [110 for _ in task_cols],
+                    )
+
+                # 绑定搜索逻辑
+                search_bar.submit(
+                    fn=lambda query, t=task_name: filter_and_search(query, t),
+                    inputs=search_bar,
+                    outputs=table,
+                )
+
+        gr.HTML("Threshold corresponding to the values of GUI and Embodied Grounding: <b>20</b>")
+
+demo.launch()
+
+# 筛选函数：只根据模型名称关键词搜索
+def filter_and_search(search_query: str, task_name: str):
+    df = leaderboard_df.copy()
+    task_cols = TASKS[task_name]
+    score_col = task_cols[0]
+
+    df[score_col] = pd.to_numeric(df[score_col], errors='coerce')
+    df = df.sort_values(by=score_col, ascending=False, na_position='last')
+
+    if search_query.strip():
+        terms = [term.strip().lower() for term in search_query.split(";")]
+        pattern = "|".join(terms)
+        df = df[df["Model"].str.lower().str.contains(pattern, regex=True)]
+
+    return df[["Model"] + task_cols]
+
+def get_initial_table(task_name: str):
+    df = leaderboard_df.copy()
+    task_cols = TASKS[task_name]
+    score_col = task_cols[0]
 
-demo = gr.Blocks()
+    df[score_col] = pd.to_numeric(df[score_col], errors='coerce')
+    df = df.sort_values(by=score_col, ascending=False, na_position='last')
+    return df[["Model"] + task_cols]
 
-with demo:
-    gr.HTML(TITLE)
+# Gradio UI 构建
+with gr.Blocks() as demo:
+    gr.HTML("<h2>Leaderboard</h2>")
     with gr.Column():
-        gr.Markdown(DESCRIPTION, elem_classes="markdown-text")
-        with gr.Row():
-            search_bar = gr.Textbox(placeholder="Search for your model...", show_label=False)
-
-        with gr.Row():
-            cols_bar = gr.CheckboxGroup(
-                choices=[c for c in leaderboard_df.columns[1:] if c != "Average"],
-                show_label=False,
-                info="Select columns to display",
-            )
-        with gr.Group():
-            leaderboard_table = gr.Dataframe(
-                value=leaderboard_df,
-                wrap=True,
-                column_widths=[400, 110] + [(260 + len(c)) for c in leaderboard_df.columns[1:]],
-            )
-
-        threshold_text = gr.HTML("Threshold corresponding to the values of gui and embodied: 20")
-    cols_bar.change(filter_and_search, inputs=[cols_bar, search_bar], outputs=[leaderboard_table])
-    search_bar.submit(filter_and_search, inputs=[cols_bar, search_bar], outputs=[leaderboard_table])
+        gr.Markdown("Search and view results for each task.", elem_classes="markdown-text")
+
+    with gr.Tabs(elem_classes="tabs-buttons") as tabs:
+        for task_name, task_cols in TASKS.items():
+            with gr.TabItem(task_name):
+                # 初始数据：按得分降序
+                sub_df = leaderboard_df[["Model"] + task_cols].copy()
+                sub_df[task_cols[0]] = pd.to_numeric(sub_df[task_cols[0]], errors="coerce")
+                sub_df = sub_df.sort_values(by=task_cols[0], ascending=False, na_position="last")
+
+                with gr.Row():
+                    search_bar = gr.Textbox(placeholder="Search model name...", show_label=False)
+
+                refresh_btn = gr.Button("Refresh")
+                with gr.Group():
+                    table = gr.Dataframe(
+                        value=sub_df,
+                        wrap=True,
+                        column_widths=[400] + [110 for _ in task_cols],
+                    )
+
+                # 绑定搜索逻辑
+                search_bar.submit(
+                    fn=lambda query, t=task_name: filter_and_search(query, t),
+                    inputs=search_bar,
+                    outputs=table,
+                )
+                def refresh(task=task_name):
+                    return "", get_initial_table(task)
+
+                refresh_btn.click(
+                    fn=refresh,
+                    outputs=[search_bar, table]
+                )
+
+        gr.HTML("Threshold corresponding to the values of GUI and Embodied Grounding: <b>20</b>")
 
 demo.launch()