Skip to content

Conversation

@juanmichelini
Copy link
Collaborator

Summary

This PR fixes the SWT-Bench build images path issue by:

  • Adding benchmarks/swtbench/build_images.py wrapper script that reuses the SWE-Bench build implementation
  • Updating the GitHub workflow .github/workflows/build-swt-bench-images.yml to use the correct path

Background

This is a focused PR extracted from #180 that contains only the build_image changes needed for SWT-Bench. The wrapper script provides a stable entrypoint for the build workflow while reusing the existing SWE-Bench implementation.

Changes

  1. New file: benchmarks/swtbench/build_images.py - Wrapper script that delegates to benchmarks.swebench.build_images.main
  2. Updated: .github/workflows/build-swt-bench-images.yml - Changed path from benchmarks/swt_bench/build_images.py to benchmarks/swtbench/build_images.py

Testing

The changes are minimal and maintain the same functionality:

  • The wrapper script simply delegates to the existing SWE-Bench implementation
  • The workflow path update ensures the correct script is called

This addresses the build image infrastructure needs for SWT-Bench without introducing any of the broader evaluation output standardization changes from the original PR.

@juanmichelini can click here to continue refining the PR

- Add benchmarks/swtbench/build_images.py wrapper script
- Update GitHub workflow to use correct path
- SWT-Bench reuses SWE-Bench build implementation

Co-authored-by: openhands <[email protected]>
@openhands-ai
Copy link

openhands-ai bot commented Dec 22, 2025

Looks like there are a few issues preventing this PR from being merged!

  • GitHub Actions are failing:
    • Pre-commit checks

If you'd like me to help, just leave a comment, like

@OpenHands please fix the failing actions on PR #187 at branch `build-image-swtbench-fix`

Feel free to include any additional details that might help me get this PR into a better state.

You can manage your notification settings

@juanmichelini
Copy link
Collaborator Author

@OpenHands please fix the failing actions on PR #187 at branch build-image-swtbench-fix

@openhands-ai
Copy link

openhands-ai bot commented Dec 22, 2025

I'm on it! juanmichelini can track my progress at all-hands.dev

@openhands-ai
Copy link

openhands-ai bot commented Dec 22, 2025

I've fixed the failing pre-commit checks on PR #187.

Issue: The file benchmarks/swtbench/build_images.py was missing a newline at the end of the file, which caused the ruff-format pre-commit hook to fail.

Fix: Added the missing newline at the end of the file.

Commit pushed: 5daa965 - "Fix missing newline at end of file"

The changes have been pushed to the build-image-swtbench-fix branch and the pre-commit checks should now pass. You can view the PR at: #187

View full conversation

@simonrosenberg
Copy link
Collaborator

Closing this and reopening #170 instead

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants