[ompatil1906]

Summary

math.isfinite() raises an unlabelled OverflowError when given integers or Fractions that cannot be converted to float (e.g. 10**400). Public entry points such as TransformConfig(gop_seconds=...) and plan_batches(..., stagger_interval_s=...) therefore leak OverflowError instead of the field-named ValueError used for every other invalid numeric input.

Root cause

In require_finite_float and require_non_negative_real, the call to math.isfinite is not protected against overflow during the float conversion that CPython performs internally.

Fix

Catch OverflowError around math.isfinite and re-raise a ValueError that names the field, matching the existing error style. No change to accepted value sets or to the behaviour of values that fit in a float.

Verification

• Manual reproduction before/after: TransformConfig(gop_seconds=10**400) and plan_batches({'item': 1}, batch_count=1, stagger_interval_s=Fraction(10**400, 1)) now raise ValueError containing the field name.
• New unit tests in tests/test_field_guards.py cover large int for the float guards and both large int/Fraction for the Real guard.
• Extended parametrized cases in tests/test_batching.py and tests/test_transform.py.
• PYTHONPATH=src python -m pytest tests/test_field_guards.py -q → 9 passed.
• Full suite could not be run in this environment (uv dependency resolution hits an exclude-newer / private-index interaction for duckdb); the change is isolated to pure-Python guards and the added tests do not require the rest of the package.

Fixes #739

---

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[github-actions]

👋 Hi @ompatil1906 — thank you so much for your first contribution to HFlow!

A maintainer will review your pull request as soon as possible. In the meantime:

• ⭐ Star the repo: https://github.com/Hebbian-Robotics/hflow
• 💬 Join our Discord: https://discord.gg/vacepQvjmg
• 📖 Check CONTRIBUTING.md for the validation commands CI will run
• 📣 Share the project with your network

💡 Tip: one open pull request per contributor at a time. Before starting an issue, check its sidebar for an assignee or a linked pull request. Either one means somebody is already on it; everything else is fair game.

We are excited to have you here and appreciate your help making the project better! 🙌

[greptile-apps]

Confidence Score: 4/5

[Medium impact] The PR changes error reporting across shared numeric validation paths.

The PR appears safe to merge, with a non-blocking gap in error messages for extremely large values.

Findings

1.  Larger values hide the field ▶

Summary

The PR turns float-conversion overflow into field-named errors.

• Oversized numeric settings now raise a field-named ValueError.

_{Reviews (1) · Last reviewed commit: "fix(guards): raise ValueError for values..." · Reviewed by Greptile}

[greptile-apps]

Larger values hide the field

Both new overflow handlers format the rejected number before raising the field-named error. On Python 3.11 with the default digit limit, 10**5000 and Fraction(10**5000, 1) make that formatting raise its own ValueError, which does not name the bad setting. This leaves the fix incomplete for larger inputs. Avoid formatting the full value in both handlers, and add tests beyond the digit limit.

[kstonekuan]

Thanks for the contribution. #739 is already assigned to @yoveneim, so I’m keeping it reserved for them. You’re welcome to pick an unassigned good first issue.