[Chen Yufeiyang (feiiiiii5)]

EvalData accepts awaitable iterators and async data factories, but both Eval and EvalAsync currently pass the resulting coroutine into a synchronous for loop. Evaluation fails with TypeError: 'coroutine' object is not iterable before processing a sample.

Await the data after resolving its factory and before adapting the iterator. This matches the existing workflow_eval.py handling and preserves synchronous iterators and async generators.

Validation on Python 3.13.18 with locked dependencies:

• The new public-API regression has 4 failures and 4 passing synchronous controls on base e26deb0; all 8 cases pass with the fix.
• pytest src/braintrust/test_framework.py src/braintrust/test_framework2.py --vcr-record=none -q: 180 passed, 5 skipped.
• Ruff 0.15.21 lint and format checks pass for both changed files.
• Full pre-commit run --all-files passes (pre-commit 4.6.2, including Ruff and Codespell).
• Targeted Pylint (--errors-only) passes for both changed files.
• Full nox -R -s test_core -- --vcr-record=none -q: 1,127 passed, 114 skipped, 12 xfailed, 23 subtests passed. Git signing was disabled only in this test process for deterministic generated Git fixtures.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-06T14:42:54.539565Z	bb9ea98	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review" or "Codex (@codex) security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.