test: extend CUDA test suite for PyTorch packages and CUDA_HOME

[smoparth]

Add tests to validate the packages from #70 and env var from #69:

• CUDA_HOME env var equals /usr/local/cuda
• libcupti.so present (cuda-cupti package)
• libcusparseLt.so present (libcusparselt0 package)
• libcudss.so present (libcudss0-cuda package)

Closes #74

Description

How Has This Been Tested?

Merge criteria:

• The commits are squashed in a cohesive manner and have meaningful messages.
• Testing instructions have been added in the PR body (for PRs involving changes that are not immediately obvious).
• The developer has manually tested the changes and verified that the changes work

Summary by CodeRabbit

• Tests
  • Added validation tests for CUDA environment: verifies CUDA_HOME location and expands library checks to include libcupti, libcusparseLt and libcudss, with clear guidance if libraries are missing.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

---

📝 Walkthrough

Walkthrough

Adds four tests to tests/test_cuda_image.py validating the CUDA_HOME environment variable and presence of three CUDA/PyTorch libraries (libcupti, libcusparseLt, libcudss) by querying the system linker cache (ldconfig).

Changes

Cohort / File(s)	Summary
CUDA environment & CUDA/PyTorch library tests tests/test_cuda_image.py	Added four tests: test_cuda_home(cuda_container) asserts CUDA_HOME == /usr/local/cuda; test_libcupti_present(cuda_container), test_libcusparselt_present(cuda_container), and test_libcudss_present(cuda_container) check for libcupti, libcusparseLt, and libcudss in ldconfig -p with guidance messages on failures.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Poem

🐰 I hopped through bins and paths tonight,
Searching for cupti, sparseLt, cudss bright,
I sniffed CUDA_HOME at /usr/local/cuda too,
Tests now hop on, checking what's true — woo! ✨

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The PR title clearly and concisely summarizes the main change: extending the CUDA test suite to include validation for PyTorch packages and the CUDA_HOME environment variable.
Linked Issues check	✅ Passed	The PR fulfills all coding requirements from issue #74: adds tests for CUDA_HOME validation, libcupti, libcusparseLt, and libcudss presence via ldconfig checks.
Out of Scope Changes check	✅ Passed	All changes are within scope—only four new test functions were added to tests/test_cuda_image.py directly addressing the requirements from issues #69 and #70.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@tests/test_cuda_image.py`:
- Around line 108-109: The assertion message in tests/test_cuda_image.py (the
assert on variable result from cuda_container.run("ldconfig -p | grep
libcusparseLt")) exceeds the 100-char line limit; shorten or wrap the message so
the line length is <=100 characters (for example, shorten the message text or
move part of it to a variable) while preserving the assertion and meaning that
libcusparseLt.so may be missing if returncode != 0.