Auto-export consolidated excel workbook (results.xlsx: Summary/Perf/Categories/Samples) on every --out-dir run; per-bench artifact .md instead of .txt; xlsxwriter joins core deps

Co-Authored-By: Claude <noreply@anthropic.com>

evalharness/cli.py

@@ -449,6 +449,7 @@ def _cmd_eval_run(args) -> int:
         (Path(out_dir) / 'reports').mkdir(exist_ok=True)
 
     rows = []
+    all_reports = []
     run_started = _time.time()
     total_runs = len(args.datasets)
     model_spec = _compose_model_spec(args)
@@ -524,14 +525,15 @@ def _cmd_eval_run(args) -> int:
                 _print_phase(console, i + 1, total_runs, name,
                              'writing report and visualization files')
                 report.save(f'{out_dir}/reports/{name}.report.json')
-                with open(f'{out_dir}/viz/{name}.txt', 'w', encoding='utf-8') as f:
+                with open(f'{out_dir}/viz/{name}.md', 'w', encoding='utf-8') as f:
-                    f.write(render(report, style='text'))
+                    f.write(render(report, style='md'))
             if args.verbose:
                 if not _print_result_panel(console, report, _time.time() - t0):
                     print(render(report, style=args.style))
             elif len(args.datasets) == 1:
                 if console is None:
                     print(render(report, style=args.style))
+            all_reports.append(report)
             primary = next(iter(report.metrics), '')
             secs_total = sum(float((s.usage or {}).get('latency_s', 0) or 0)
                              for s in report.samples)
@@ -576,6 +578,16 @@ def _cmd_eval_run(args) -> int:
             _print_benchmark_result(console, i + 1, total_runs, name,
                                     'failed', _time.time() - t0)
 
+    if all_reports and out_dir:
+        try:
+            from evalharness.viz import render as _render
+
+            xb = _render(all_reports, style='excel',
+                         out=f'{out_dir}/viz/results.xlsx')
+            print(f'excel  -> {xb}', flush=True)
+        except Exception as e:
+            print(f'excel export skipped: {type(e).__name__}: {str(e)[:80]}',
+                  file=sys.stderr)
     if rows:
         _print_summary_table(console, rows)
         if out_dir:
@@ -637,10 +649,12 @@ def _cmd_eval_run(args) -> int:
                 f'\n{mark} [bold]运行结束[/bold] · {ok_n}/{len(rows)} benchmarks ok\n'
                 f'  结果 {ap}\n'
                 f'  ├─ reports/<bench>.report.json\n'
+                f'  ├─ viz/results.xlsx (excel 打开)\n'
                 f'  └─ viz/summary.md (+ summary.csv)')
         else:
             print(f'\n运行结束 · {ok_n}/{len(rows)} benchmarks ok\n'
-                  f'  结果 {ap}\n  ├─ reports/<bench>.report.json\n  └─ viz/summary.md')
+                  f'  结果 {ap}\n  ├─ reports/<bench>.report.json\n'
+                  f'  ├─ viz/results.xlsx\n  └─ viz/summary.md')
     elif rows and rows[0]['ok'] and args.out:
         _notice('运行结束 · 结果已保存', args.out)
         print(f'{ok_n}/{len(rows)} benchmarks ok')

pyproject.toml

@@ -16,6 +16,7 @@ dependencies = [
     "numpy",          # official DROP aligner
     "scipy",
     "rich",
+    "xlsxwriter",     # excel result workbook
     "tree_sitter>=0.21",        # vendored BFCL official AST checker (python)
     "tree-sitter-java>=0.21",   # bfcl java categories
     "tree-sitter-javascript>=0.21",  # bfcl javascript categories