Update

app.py

@@ -44,8 +44,14 @@ except Exception:
 dataset = load_dataset("dtcxzyw/llvm-apr-benchmark")
 total_issues = dataset.num_rows["test"]
 bug_id_to_time = dict()
+bug_id_by_cat = {
+    "crash": [],
+    "miscompilation": [],
+    "hang": [],
+}
 for issue in dataset["test"]:
     bug_id_to_time[issue["bug_id"]] = pd.to_datetime(issue["knowledge_cutoff"])
+    bug_id_by_cat[issue["bug_type"]].append(issue["bug_id"])
 timeline_xs = []
 timeline_ys = []
 timeline_cols = []
@@ -56,6 +62,14 @@ for bug_id, time in bug_id_to_time.items():
     timeline_ys.append(0)
     timeline_cols.append("All")
     timeline_bugids.append(bug_id)
+cat_cnt = 4
+for cat, bug_ids in bug_id_by_cat.items():
+    cat_cnt -= 1
+    for bug_id in bug_ids:
+        timeline_xs.append(bug_id_to_time[bug_id])
+        timeline_ys.append(cat_cnt)
+        timeline_cols.append(str(cat).capitalize())
+        timeline_bugids.append(bug_id)
 LEADERBOARD_DF = get_leaderboard_df(EVAL_REQUESTS_PATH, total_issues)
 for row in LEADERBOARD_DF.itertuples():
     print(row)
@@ -113,6 +127,7 @@ with demo:
                 y_label="Model",
                 title="Timeline",
                 y_lim=(-model_cnt - 1, 1),
+                tooltip=["bug_id", "method_name", "time"],
             )
 
         with gr.TabItem("🚀 Submission", elem_id="llm-benchmark-tab-table", id=1):