SWE-bench: fix SWE-agent hanging, adjust expected scores

[ludwig-n]

1. Fixes an issue where SWE-agent was hanging forever on 41 Django instances by downgrading to rich==14.2.0 in the SWE-agent environment.
2. Adjusts expected scores of Qwen3-Coder-30B-A3B in the tests and docs after this evaluation harness fix: Fix patches being reversed if repo state in container differs from base_commit Kipok/SWE-bench#3.

Summary by CodeRabbit

Release Notes

• Documentation

  • Updated SWE-bench evaluation results with latest benchmark metrics.

• Chores

  • Added explicit version constraint for a package dependency during setup.

• Tests

  • Updated test baseline metrics to reflect new measurement results.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[greptile-apps]

_{No files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[coderabbitai]

📝 Walkthrough

Walkthrough

Updates numeric example values in evaluation documentation, adds explicit rich package downgrade (v14.2.0) to SWE-agent setup in swebench.py, and updates metric ranges and reference date in test validation thresholds to reflect new measurement data.

Changes

Cohort / File(s)	Summary
Documentation Updates docs/evaluation/code.md	Updated SWE-bench results example values (issues_resolved: 45.0 → 48.4, no_patch: 0.4 → 1.0, patch_cant_apply: 0.8 → 1.6).
Infrastructure Setup nemo_skills/inference/eval/swebench.py	Added explicit downgrade of rich package to version 14.2.0 in SWE-agent setup sequence with explanatory comment; affects dependency resolution during container initialization.
Test Configuration tests/slurm-tests/qwen3coder_30b_swebench/check_results.py	Updated METRIC_RANGES validation thresholds for openhands (41.9-47.9 → 44.3-52.3) and swe_agent (45.5-52.4 → 45.5-53.5); updated reference date to 2026-01-30 and changed averaging methodology to 3-run average.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Possibly related PRs

• Fixes to support SWE-bench Multilingual #1064: Modifies swebench.py installation sequence, overlapping with rich package management changes in this PR.
• SWE-bench: add Slurm tests & other improvements #920: Updates the same check_results.py test file METRIC_RANGES values.

Suggested reviewers

• Kipok
• gwarmstrong

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the two main changes: fixing SWE-agent hanging (via rich downgrade) and adjusting expected scores (reflected in documentation and test updates).
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch ludwig-n/swe-agent-fix-rich

Tip

🧪 Unit Test Generation v2 is now available!

We have significantly improved our unit test generation capabilities.

To enable: Add this to your .coderabbit.yaml configuration:

reviews:
  finishing_touches:
    unit_tests:
      enabled: true

Try it out by using the @coderabbitai generate unit tests command on your code files or under ✨ Finishing Touches on the walkthrough!

Have feedback? Share your thoughts on our Discord thread!

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[Kipok]

Thanks!