Fix mean of endpoints

app.py

@@ -13,21 +13,22 @@ def make_user_clickable(name):
     link =f'https://huggingface.co/{name}'
     return f'<a target="_blank" href="{link}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">{name}</a>'
 def make_tag_clickable(tag):
-    return f'<a target="_blank" href="{tag}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">Report</a>'
+    return f'<a target="_blank" href="{tag}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">link</a>'
 
 def get_leaderboard(endpoint):
     dset = load_dataset(results_repo, split='train', download_mode="force_redownload")
     full_df = dset.to_pandas()
     to_show = full_df.copy(deep=True)
     to_show = to_show[to_show['endpoint'] == endpoint]
+    to_show = to_show[to_show['user'] != 'test'] 
     to_show['user'] = to_show['user'].apply(lambda x: make_user_clickable(x)).astype(str)
+    to_show['model details'] = to_show['model_report'].apply(lambda x: make_tag_clickable(x)).astype(str)
     # Get the most recent submission
     to_show["submission_time"] = pd.to_datetime(to_show["submission_time"])
     latest_per_user = to_show.loc[to_show.groupby("user")["submission_time"].idxmax()].reset_index(drop=True)
     latest_per_user.rename(columns={"submission_time": "submission time"}, inplace=True)
     # The columns to display publicly
-    df = latest_per_user[["user", "MAE", "R2", "Spearman R", "Kendall's Tau", "submission time"]]
-    # TODO: Also display the column with report link and make it clickable with make_tag_clickable 
+    df = latest_per_user[["user", "MAE", "R2", "Spearman R", "Kendall's Tau", "submission time", "model details"]]
 
     return df
 
@@ -102,29 +103,48 @@ def gradio_interface():
         """
 
         # --- Gradio Interface ---
+        gr.HTML("""
+                <style>
+                /* bold only the "Overall" tab label */
+                    #lb_subtabs [role="tab"][aria-controls="all_tab"] {
+                        font-weight: 700 !important;
+                    }
+                </style>
+                """)
         with gr.Tabs(elem_classes="tab-buttons"):
             lboard_dict = {}
 
             with gr.TabItem("📝About"):
                 gr.Markdown(welcome_md)
 
-            with gr.TabItem("🚀Leaderboard"):
+            with gr.TabItem("🚀Leaderboard", elem_id="lb_subtabs"):
                 gr.Markdown("View the leaderboard for each ADMET endpoint by selecting the appropiate tab.")
                 # Make separate leaderboards in separate tabs  
+                # Aggregated leaderboard
+                with gr.TabItem('OVERALL', elem_id="all_tab"):
+                    lboard_dict['Average'] = Leaderboard(
+                        value=get_leaderboard('Average'),
+                        datatype=['markdown', 'number', 'number', 'number', 'number', 'str', 'markdown'],
+                        select_columns=["user", "MAE", "R2", "Spearman R", "Kendall's Tau", "submission time", "model details"],
+                        search_columns=["user"],
+                        render=True
+                    )
+                # per-endpoint leaderboard
                 for endpoint in ENDPOINTS:
                     with gr.TabItem(endpoint):
                         lboard_dict[endpoint] = Leaderboard(
                             value=get_leaderboard(endpoint),
-                            datatype=['markdown', 'number', 'number', 'number', 'number'],
-                            select_columns=["user", "MAE", "R2", "Spearman R", "Kendall's Tau", "submission time"],
+                            datatype=['markdown', 'number', 'number', 'number', 'number', 'str', 'markdown'],
+                            select_columns=["user", "MAE", "R2", "Spearman R", "Kendall's Tau", "submission time", "model details"],
                             search_columns=["user"],
+                            render=True
                         )
-                # TODO: Make aggregated leaderboard and display on first tab
+
 
             with gr.TabItem("Submit Predictions"):
                 gr.Markdown(
                 """
-                # ADME Endpoints Submission
+                # ADMET Endpoints Submission
                 Upload your prediction files here as a csv file.
                 """
                 )
@@ -195,7 +215,7 @@ def gradio_interface():
                             """
                         )
                     with gr.Column():
-                        predictions_file = gr.File(label="Single file with ADME predictions (.csv)",
+                        predictions_file = gr.File(label="Single file with ADMET predictions (.csv)",
                                                 file_types=[".csv"],
                                                 file_count="single",)
 

evaluate.py

@@ -162,12 +162,14 @@ def evaluate_data(filename: str) -> None:
         meta = SubmissionMetadata(**_meta)
         username = meta.participant.hf_username
         timestamp = meta.submission_time_utc
+        report = meta.participant.model_tag
     except Exception as e:
         raise gr.Error(f"Failed to load metadata file: {e}. No results written to results dataset.")
 
     # Write results to results dataset
     results_df['user'] = username
     results_df['submission_time'] = timestamp
+    results_df['model_report'] = report
     safe_user = _unsafify_username(username)
     destination_path = f"results/{safe_user}_{timestamp}_results.csv"
     tmp_name = None
@@ -212,4 +214,10 @@ def calculate_metrics(
         df_results.loc[i, 'Spearman R'] = spearman
         df_results.loc[i, "Kendall's Tau"] = ktau
 
-    return df_results
+    num_cols = ["MAE", "R2", "Spearman R", "Kendall's Tau"]
+    df_results[num_cols] = df_results[num_cols].apply(pd.to_numeric, errors="coerce")
+    means = df_results[num_cols].mean()
+    avg_row = {"endpoint": "Average", **means.to_dict()}
+    df_with_average = pd.concat([df_results, pd.DataFrame([avg_row])], ignore_index=True)
+
+    return df_with_average