Fix how build.sh determines CUDA version

[KyleFromNVIDIA]

Description

In #23472, we incorrectly used RAPIDS_CUDA_VERSION in build.sh. This variable comes from CI only and should not be required to build locally. If RAPIDS_CUDA_VERSION is not available, determine the CUDA version in a similar manner to rapids-build-backend: run nvcc --version and parse the output.

Fixes #23587

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• Chores
  • Improved build configuration to automatically identify the installed CUDA version.
  • Package installation now uses the detected CUDA environment, improving reliability across supported setups.
  • Added an error message when the CUDA version cannot be detected.

Walkthrough

build.sh now detects CUDA_VERSION from nvcc, exits when detection fails, and uses the detected value for Python package installation matrix configuration.

Changes

CUDA version configuration

Layer / File(s)	Summary
Detect and apply CUDA version build.sh	build.sh extracts CUDA_VERSION from nvcc --version, exits when no valid version is found, and uses it instead of RAPIDS_CUDA_VERSION for RAPIDS matrix configuration.

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

Suggested reviewers: msarahan

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The change addresses issue #23587 by removing the local build dependency on RAPIDS_CUDA_VERSION.
Out of Scope Changes check	✅ Passed	The changes are limited to CUDA version detection in build.sh and match the linked issue objectives.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Title check	✅ Passed	The title clearly and concisely describes the fix to CUDA version detection in build.sh.
Description check	✅ Passed	The description explains the build.sh fix, its local-build impact, and the use of nvcc for CUDA version detection.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@build.sh`:
- Around line 66-67: Validate CUDA_VERSION immediately after the nvcc parsing
assignment: detect nvcc command failure and an empty parsed value, then
terminate the build with a clear error before the RAPIDS matrix configuration
uses it. Preserve the existing version extraction behavior for valid nvcc
output.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 89a2a2e3-12d7-47d3-9a89-b1d6698409dc

📥 Commits

Reviewing files that changed from the base of the PR and between d8fdbcc and ee3f5f0.

📒 Files selected for processing (1)

• build.sh

[KyleFromNVIDIA]

/merge