chore: Use Python match statement where possible

[mdboom]

Now that Python 3.10 is long in the past, we can rely on a robust implementation of the match statement. I had my agent find candidates and update them -- most of what it found was in test code, but I thought it best to leave that alone.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository: NVIDIA/cuda-python/.coderabbit.yaml
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 90d3373e-336e-47dd-be9d-31bf1c2a643c

📥 Commits

Reviewing files that changed from the base of the PR and between 1f41281 and 1230917.

📒 Files selected for processing: 5

• cuda_core/cuda/core/utils/_program_cache/_keys.py
• samples/cuda_bindings/3_CUDA_Features/globalToShmemAsyncCopy/globalToShmemAsyncCopy.py
• samples/cuda_bindings/4_CUDA_Libraries/nvidiaSmi/nvidiaSmi.py
• samples/cuda_core/glInteropMipmapLod/glInteropMipmapLod.py
• toolshed/check_cython_abi.py

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 9 remain after this review.

📝 Summary

Summary by CodeRabbit

• Refactor
  • Updated the handling of CUDA sample selections, keyboard controls, and NVML compute-mode labels. Existing sample behavior and displayed labels remain unchanged, including fallback handling for unrecognized modes.
  • Updated supporting utilities to process source values and Cython declarations with equivalent results. These changes do not alter the available features or outputs.

The pull request replaces conditional chains with Python match/case in five code paths. The summaries report that the existing launch configurations, mappings, key handling, and parsing results remain unchanged.

Changes: Pattern-Matching Refactors

Layer / File(s)	Summary
CUDA kernel launch dispatch samples/cuda_bindings/3_CUDA_Features/globalToShmemAsyncCopy/globalToShmemAsyncCopy.py	The warmup and timed dispatches use match/case for all eight kernel choices. Their launch configurations, including the shared-state dimensions, remain unchanged.
Sample input and compute-mode dispatch samples/cuda_core/glInteropMipmapLod/glInteropMipmapLod.py, samples/cuda_bindings/4_CUDA_Libraries/nvidiaSmi/nvidiaSmi.py	Keyboard handling uses match/case while retaining bias adjustments and reset behavior. NVML compute-mode labels use match/case, with the existing N/A error fallback retained.
Cache encoding and ABI parsing cuda_core/cuda/core/utils/_program_cache/_keys.py, toolshed/check_cython_abi.py	Extra-source encoding and Cython node classification use match/case. The described encoding, field extraction, and result mappings remain unchanged.

Priority: ⬇️ Low

Change: Refactor

Merge Risk: ⚪ Minimal · up to 12309

No actionable merge-blocking risk remains.

✨ Finishing Touches: 💡 1

🛠️ Fix failing CI checks: 💡

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

[github-actions]

Doc Preview CI
🚀 View preview at https://nvidia.github.io/cuda-python/pr-preview/pr-3114/
https://nvidia.github.io/cuda-python/pr-preview/pr-3114/cuda-core/
https://nvidia.github.io/cuda-python/pr-preview/pr-3114/cuda-bindings/
https://nvidia.github.io/cuda-python/pr-preview/pr-3114/cuda-pathfinder/
Preview will be ready when the GitHub Pages deployment is complete.