Fix issue #5222: [Refactor]: Refactor the evaluation directory

[openhands-agent]

This PR fixes #5222 by reorganizing the evaluation directory structure to improve clarity and maintainability.

Changes

• Created evaluation/benchmarks/ directory to house all ML literature benchmarks
• Kept utility directories (utils, integration_tests, regression, static) directly under evaluation/
• Updated paths in documentation and GitHub workflows to reflect the new structure
• Added missing benchmarks to evaluation/README.md:
  • Commit0 and DiscoveryBench under Software Engineering
  • Browsing Delegation under Web Browsing
  • ScienceAgentBench under Misc. Assistance

Testing

• All pre-commit hooks pass (ruff, mypy, etc.)
• All unit tests pass (377 tests)

Review Notes

Key files to review:

• .github/workflows/eval-runner.yml - Updated paths for integration tests and benchmarks
• evaluation/README.md - Added missing benchmarks and updated paths
• Documentation files - Updated references to benchmark paths

---

To run this PR locally, use the following command:

docker run -it --rm   -p 3000:3000   -v /var/run/docker.sock:/var/run/docker.sock   --add-host host.docker.internal:host-gateway   -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:a759dd8-nikolaik   --name openhands-app-a759dd8   docker.all-hands.dev/all-hands-ai/openhands:a759dd8

[neubig]

Just noting that I have confirmed the code and it looks good to me, but I'd like a second review.