fix: use ffprobe_path() instead of hardcoded /usr/bin/ffprobe in test_video.py

[kragent66-glitch]

Fixes #115

Problem

tests/test_video.py reimplements ffprobe resolution with a hardcoded Linux path (/usr/bin/ffprobe) as fallback, missing the HFLOW_FFPROBE override and proper error handling from ffprobe_path().

Solution

Replaced the manual resolution logic in _ffprobe_video_stream_fields() with a single call to ffprobe_path(), following the pattern established in PR #110.

Changes

• Added ffprobe_path to imports from hflow.ffmpeg
• Replaced 3 lines of manual resolution with ffprobe_path()

Verification

• All 9 tests in tests/test_video.py pass
• No hardcoded binary paths remain in tests
• Quality checks pass (ruff check, ruff format)
• Note: ty check has pre-existing errors about missing pyarrow module unrelated to this change

[github-actions]

👋 Hi @kragent66-glitch — thank you so much for your first contribution to HFlow!

A maintainer will review your pull request as soon as possible. In the meantime:

• ⭐ Star the repo: https://github.com/Hebbian-Robotics/hflow
• 💬 Join our Discord: https://discord.gg/vacepQvjmg
• 📖 Check CONTRIBUTING.md for the validation commands CI will run
• 📣 Share the project with your network

💡 Tip: one open pull request per contributor at a time. Issues with an assignee are taken; everything else is fair game.

We are excited to have you here and appreciate your help making the project better! 🙌

[kstonekuan]

Welcome, and LGTM. Merging.

Exactly the fix the issue asked for, and you kept the ffmpeg_path import because line 33 still uses it, which is the easy thing to get wrong when swapping a resolver. The definition-of-done grep comes back empty, so no hardcoded binary path is left in tests/.

Validated against main (2f99071): ruff check, ruff format --check, ty check all clean, 645 passed and 3 skipped.

One correction worth having for next time: the pyarrow errors you saw are not pre-existing, they are a missing extra. uv sync --locked --all-extras installs it and ty check then passes clean. Worth re-syncing with that flag before deciding a failure is upstream, since it also makes pytest collect the arrow tests you were otherwise skipping.

Also: we cap non-collaborators at one open pull request at a time, so a bot will auto-close a second one with a note rather than losing it.