Fix batch-summary benchmark: use native combine_summaries, drop xfail

[jepegit]

Problem

The benchmark CI check fails with:

missing benchmark 'test_benchmark_batch_summary_collection' in current run

test_benchmark_batch_summary_collection was xfailed pending the Epic C polars migration, still calling the old pandas helpers.concat_summaries — which now raises AttributeError: 'DataFrame' object has no attribute 'index' on the polars summary frames. An xfailed benchmark records no result, so check_baseline.py sees it absent from the current run and fails (it treats a baseline benchmark missing from the run as a regression).

Fix

Epic B/C is done, so this does what the xfail note said to do "then": migrate the benchmark to the native collector path — Batch.combine_summaries() → cellpy.batch.combine_summaries (returns a polars frame) — and remove the xfail. The benchmark runs and records a result again.

Verification

• pytest benchmarks/test_performance.py::test_benchmark_batch_summary_collection --benchmark-only → 1 passed (was 1 xfailed).
• Local mean ≈ 1.19× the committed ubuntu baseline — under the +20% warn line and far under the +100% hard-fail line, so check_baseline.py no longer reports it missing. The committed baseline value can be refreshed on CI when convenient (it was captured against the old pandas path).

🤖 Generated with Claude Code

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.