- Added summary that could be directly copy pasted into sheets

app.py

@@ -558,7 +558,8 @@ def run_evaluation(
     markdown_summary = generate_compact_summary_markdown(
         questions, results, summary_stats
     )
-    csv_summary = generate_csv_summary(questions, results, summary_stats)
+    # csv_summary = generate_csv_summary(questions, results, summary_stats)
+    csv_summary = generate_excel_summary(questions, results, summary_stats)
     slurm_id = os.environ.get("SLURM_JOB_ID", "")
     if save_summary and slurm_id:
         file_name = f"summaries/{slurm_id}_summary_{time()}.md"
@@ -1027,6 +1028,41 @@ def generate_csv_summary(questions, results, summary_stats):
     return "\n".join(lines)
 
 
+def generate_excel_summary(questions, results, summary_stats):
+    """Generate Excel format summary"""
+    # TODO: add Excel file download if necessary
+    if not summary_stats or not questions or not results:
+        return "No data available"
+
+    lines = []
+
+    # Header for detailed results
+    model_names = list(results.keys())
+    header = "\t".join(model_names)
+    lines.append(header)
+
+    # Detailed results
+    for q_idx, question in enumerate(questions):
+        # row = f'{q_idx + 1},"{question["question"]}",{question["correct_answer"]}'
+        row = ""
+
+        for model in model_names:
+            if q_idx < len(results[model]) and "error" not in results[model][q_idx]:
+                result = results[model][q_idx]
+                predicted = result.get("predicted", "N/A")
+                is_correct = result.get("correct", False)
+                confidence = result.get("confidence", 0)
+                # row += is_correct
+                status_emoji = "✅" if is_correct else "❌"
+                row += f"{status_emoji} {predicted} ({confidence:.2f})\t"
+            else:
+                row += "ERROR\t"
+
+        lines.append(row)
+
+    return "\n".join(lines)
+
+
 # Sample datasets for quick testing
 SAMPLE_DATASETS = {
     "Custom (enter below)": "",
@@ -1168,7 +1204,7 @@ What is the capital of France?,Paris,London,Berlin,Paris""",
                 )
                 prefix = gr.Textbox(
                     label="Filename Prefix",
-                    placeholder="SLURM_JOB_ID will be used  by default",
+                    placeholder=f"The file will be saved at summaries/{slurm_id}_{prefix}_TIME.md will be used  by default",
                     value="",
                     interactive=True,
                     visible=True,
@@ -1250,9 +1286,11 @@ bigscience/bloom-560m""",
                 )
 
             with gr.Column():
-                gr.Markdown("### 📊 CSV Format")
+                gr.Markdown("### 📊 Excel Format (Tab separated)")
+                # gr.Markdown("### 📊 CSV Format")
                 csv_summary_output = gr.Textbox(
-                    label="CSV Summary (Copy & Paste Ready)",
+                    label="Excel Summary (Copy & Paste Ready, No additional formatting)",
+                    # label="CSV Summary (Copy & Paste Ready)",
                     lines=15,
                     max_lines=25,
                     show_copy_button=True,