Make File.is_installed binary-safe

[MaxFreedomPollard]

Summary

Fixes #782.

• File.is_installed compared the installed copy with the source using read_text(), which raises UnicodeDecodeError for files that are not valid UTF-8. It now compares bytes, which matches install(), where the file is copied with shutil.copyfile.
• A file that exists with different contents was reported as "does not exist". That case now has its own message: File <installed path> differs from <src>.

The issue lists the Slurm and Kubernetes installers, but since #885 the check lives in File.is_installed, so that is the only place that needs the change.

Test Plan

• Added test_is_installed_binary_file and test_is_installed_missing_file, and the existing content-mismatch test now checks the message. Without the fix, the binary test fails with UnicodeDecodeError and the mismatch test gets the "does not exist" message.
• uv run --locked --extra dev pytest: 1960 passed, 5 skipped.
• uv run --locked --extra dev pre-commit run --files on the two changed files: all hooks pass.

Additional Notes

read_text() also normalizes line endings, so two files that differed only in CRLF vs LF used to count as the same. With a byte comparison they no longer do, which seems right for checking whether the copy made by copyfile is still identical to the source.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository: NVIDIA/cloudai/.coderabbit.yaml

Review profile: ASSERTIVE

Plan: Enterprise

Run ID: aa7ce660-ac98-4849-9600-d3d8fc8fcb47

📥 Commits

Reviewing files that changed from the base of the PR and between ea9db35 and 7348c5d.

📒 Files selected for processing (2)

• src/cloudai/_core/installables/file.py
• tests/core/installables/test_file.py

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

---

📝 Walkthrough

Walkthrough

Changes

File installation validation

Layer / File(s)	Summary
Byte-safe file validation src/cloudai/_core/installables/file.py, tests/core/installables/test_file.py	File.is_installed now checks for missing destinations, compares source and destination contents as bytes, and reports separate missing-file and content-mismatch errors. Tests cover path-specific errors and binary files with invalid UTF-8 bytes.

Priority: ➖ Normal

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

Change: Bug fix · Severity of issue fixed: Medium

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the main change: making File.is_installed safe for binary files.
Description check	✅ Passed	The description directly explains the binary-safe comparison, distinct error handling, affected scope, and test coverage.
Linked Issues check	✅ Passed	Issue #782 requires binary-safe file comparison, byte-based comparison, and distinct errors for missing files and content mismatches. File.is_installed uses read_bytes() for both files. It reports…
Out of Scope Changes check	✅ Passed	The changes are limited to the shared File.is_installed implementation and its automated tests. These changes directly support Issue #782. No unrelated behavior or files are identified.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create a new PR

---

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

[podkidyshev]

@MaxFreedomPollard thank you for contribution, it's a good change 👍