refactor: optimize benchmark retrieval in create_submit_tab function to enhance clarity and error messaging

app.py

@@ -2,8 +2,6 @@ import json
 import re
 import sys
 import threading
-from collections import namedtuple
-from functools import partial
 from textwrap import dedent
 
 import gradio as gr
@@ -452,6 +450,7 @@ def create_submit_tab(tab_id: int, demo: gr.Blocks):
                     table.select(fn=on_select, inputs=table, outputs=model_name_textbox)
 
             # ========== Section 2: Steps 8-10 (Benchmark Evaluation Results) ==========
+            benchmarks_list = get_benchmarks()
             with gr.Accordion("📊 Benchmark Evaluation Results (Steps 8-10)", open=True):
                 gr.Markdown(
                     "**8-9. Select benchmarks and fill in evaluation result values**\n\n"
@@ -459,9 +458,8 @@ def create_submit_tab(tab_id: int, demo: gr.Blocks):
                 )
 
                 # Simple form for benchmark results
-                benchmarks = get_benchmarks()
                 benchmark_results_form: list = []
-                for benchmark in benchmarks:
+                for benchmark in benchmarks_list:
                     with gr.Row():
                         benchmark_checkbox = gr.Checkbox(
                             label=f"{benchmark.title} ({benchmark.key})",
@@ -555,16 +553,19 @@ def create_submit_tab(tab_id: int, demo: gr.Blocks):
 
             # Build results: {benchmark_key: {metric: value}}
             results = {}
-            benchmarks_list = get_benchmarks()
             for benchmark, checkbox_checked, result_value in zip(
-                benchmarks_list, benchmark_checkbox_values, benchmark_result_values, strict=True
+                benchmarks_list,
+                benchmark_checkbox_values,
+                benchmark_result_values,
+                strict=True,
             ):
                 if checkbox_checked and result_value is not None:
                     # Use "acc" as the default metric (can be extended)
                     results[benchmark.key] = {"acc": float(result_value)}
 
             if not results:
-                raise ValueError("At least one benchmark result is required")
+                msg = f"At least one benchmark result is required. benchmarks: {[b.title for b in benchmarks_list]!r}"
+                raise ValueError(msg)
 
             return json.dumps({"config": config, "results": results}, indent=2, ensure_ascii=False)