Rename eval_runner.py to experiment_runner.py

[cvolkcvolk]

Summary

Align runner naming with Experiments. Renames eval_runner.py to experiment_runner.py because its unit of work is one ArenaExperiment: an ordered list[ArenaRunCfg].

Detailed description

• Uses Experiment Runner for the already-namespaced component while retaining the Arena prefix on public domain types.
• Updates Python APIs, helpers, tests, and documentation without retaining a compatibility wrapper or changing execution behavior.

[greptile-apps]

Greptile Summary

This PR renames the batch evaluation runner to match Arena Experiment terminology. The main changes are:

• eval_runner.py and related helper modules renamed to arena_experiment_runner.py names.
• Runner imports, CLI parser names, tests, and subprocess helpers updated to the new names.
• Documentation examples and Sphinx references updated to use arena_experiment_runner.py.

Confidence Score: 5/5

This looks safe to merge.

• No blocking issues found in the changed code.

Important Files Changed

Filename	Overview
isaaclab_arena/evaluation/arena_experiment_runner.py	Renamed the main runner module and updated imports for the renamed CLI and legacy helpers.
isaaclab_arena/evaluation/arena_experiment_runner_cli.py	Renamed the runner-specific CLI helpers while keeping the argument behavior unchanged.
isaaclab_arena/evaluation/legacy_arena_experiment_runner.py	Renamed the legacy JSON chunking helper and updated its runner-facing log text.
isaaclab_arena/tests/test_arena_experiment_runner.py	Updated runner tests and helpers to execute the renamed script path.
docs/pages/concepts/policy/concept_evaluation_types.rst	Updated the concept docs, runner filename references, and batch-runner Sphinx label.

_{Reviews (1): Last reviewed commit: "Rename Eval Runner to Arena Experiment R..." | Re-trigger Greptile}

[arena-review-bot]

🤖 Isaac Lab-Arena Review Bot

Summary

This PR renames eval_runner.py → arena_experiment_runner.py (plus its _cli and legacy_ helpers, their functions, tests, and docs) to align naming with the Arena Experiment concept — one experiment being an ordered list[ArenaRunCfg]. It is a pure rename with no behavior change.

I fetched the full tree at the PR head and verified the rename is complete and consistent: every old symbol (parse_eval_runner_args, add_eval_runner_arguments, run_eval_runner, legacy_eval_runner, test_eval_runner) and the old evaluation/eval_runner.py script path is gone, with no dangling references anywhere in code, docs, CI, or Docker. The renamed RST cross-reference anchor (sequential-batch-eval-runner → sequential-batch-arena-experiment-runner) is updated at both its definition and its single :ref:, so the docs build stays intact. The unrelated legacy_eval_config.py / legacy_environment_cli_args.py modules are correctly left untouched.

Findings

None. The renamed files are git renames of pre-existing files, so their 2025-2026 copyright headers are correct.

Test Coverage

Tests were renamed alongside the module and keep their @pytest.mark.with_subprocess markers; test_all_environments.py imports the renamed run_arena_experiment_runner / helpers correctly, and test_arena_experiment_config_loader.py monkeypatches the renamed arena_experiment_runner module. No new logic is introduced, so no new tests are warranted.

Verdict

Ship it.

[alexmillane]

Makes a ton of sense. LGTM