[sdk] Python SDK topological_sort does not detect cyclic task dependencies

[Riddhima8]

Environment

• KFP version:
  N/A (local Python SDK usage)

• KFP SDK version:
  main branch (commit-based, cloned from https://github.com/kubeflow/pipelines)

Description

The local graph utility used for pipeline task ordering does not explicitly detect
cyclic dependencies.

Steps to reproduce

Minimal reproducible example:

from kfp.local import graph_utils

graph = {
    'A': ['B'],
    'B': ['A'],
}

graph_utils.topological_sort(graph)

Expected result

The function should fail fast with a clear and user-friendly error
(e.g. ValueError) indicating that a cyclic dependency was detected.

Since topological sorting is only defined for DAGs, cyclic task
dependencies should be treated as a user error during local execution.

Materials and Reference

• Source code under discussion:
  sdk/python/kfp/local/graph_utils.py

• Background reference:
  Topological sorting is only defined for directed acyclic graphs (DAGs).
  Cyclic task dependencies represent a user error and should be detected
  explicitly during local pipeline execution.

---

Impacted by this bug? Give it a 👍.

[github-actions]

🎉 Welcome to the Kubeflow Pipelines repo! 🎉

Thanks for opening your first issue! We're happy to have you as part of our community 🚀

Next steps:

• Check out the Contributing Guide for repo-specific guidelines and the Kubeflow Contributor Guide for general community standards
• Join our #kubeflow-pipelines Slack channel
• Attend the Kubeflow Pipelines WG meetings

Feel free to ask questions in the comments.

[AbhiRam-beep]

Greetings, I would like to work on this issue.
/assign

[AbhiRam-beep]

The earlier PR updated graph_utils_test.py, which was incorrect.
The correct change is the PR updating graph_utils.py.

[Riddhima8]

Hi, thanks for picking this up!

I opened this issue and would be happy to work on it if it’s available.
I can add the unit test and work on a fix for cycle detection in
kfp/local/graph_utils.py.

Please let me know if it’s okay for me to proceed or if there’s any
guidance you’d like me to follow. Thanks!