Support PEP 515 underscores in string-to-float astype

[fallintoplace]

What

• Support PEP 515 underscores when casting strings to float dtypes.
• Keep invalid underscore placements rejected.

Why

Integer casts already support this syntax, but float casts still reject valid values such as 1_2.3_4.

Implementation

• Reuse the existing two-pass underscore cleanup for integer and floating-point casts.
• Add float32 and float64 coverage for valid and invalid forms.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 37644c2d-86e0-4911-b090-7baa0c6a67ee

📥 Commits

Reviewing files that changed from the base of the PR and between 9a601e2 and 31eac4a.

📒 Files selected for processing (2)

• python/cudf/cudf/core/column/string.py
• python/cudf/cudf/tests/series/methods/test_astype.py

Included review availability: Your plan provides up to 12 included reviews per hour; 10 remain after this review.

---

📝 Summary

Summary by CodeRabbit

• Bug Fixes

  • Improved conversion of strings containing valid digit-separated underscores to integer and floating-point values.
  • Underscores in valid numeric positions are now handled consistently with pandas.
  • Invalid underscore placement continues to be rejected, with floating-point validation correctly identifying malformed values.

• Tests

  • Added coverage for valid and invalid underscore placement during floating-point string conversion.

Walkthrough

The numeric string conversion path now accepts valid digit-separated underscores for integer and floating-point targets. Floating-point validation uses normalized strings. Tests cover valid and invalid floating-point underscore placement against pandas.

Changes

Numeric string casting

Layer / File(s)	Summary
Normalize and validate numeric strings python/cudf/cudf/core/column/string.py, python/cudf/cudf/tests/series/methods/test_astype.py	as_numerical_column removes valid digit-separated underscores before numeric conversion. Floating-point validation uses the normalized data. Tests cover valid decimal, exponent, and integer-like strings, plus invalid underscore placements, for float32 and float64.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Merge Risk: ⚪ Minimal · up to 31eac

The float underscore conversion change has no remaining actionable issue from the reviewed behavior and is ready to merge.

🚥 Pre-merge checks | ✅ 4 | ❌ 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%. Docstring coverage is scoped to functions touched by this diff. Analyzed 4 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: support for PEP 515 underscores in string-to-float astype conversions.
Description check	✅ Passed	The description directly explains the float-casting change, invalid underscore handling, implementation, and test coverage.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

• Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Create a new PR

---

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

[mroeschke]

/ok to test 31eac4a