Fix types in beat tracking code

[emilazy]

Reference Issue

What does this implement/fix? Explain your changes.

The development version of Numba 0.62.0 is unhappy about range being applied to floating point values here.

Any other comments?

[bmcfee]

Thanks for this - it appears to be the result of the fix for numba/numba#9909

I'd like to hold off on doing anything with this until 0.62 is final, or we somehow get #1528 off the ground, as we can't otherwise verify that it works properly in CI.

I think the solution implemented here looks fine, though the explicit cast to int for frames_per_beat may be premature since it gets cast back to float in the loop line that triggered the problem here.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.72%. Comparing base (e403272) to head (902bcca).
⚠️ Report is 54 commits behind head on 1.0.0dev.

Additional details and impacted files

@@            Coverage Diff            @@
##           1.0.0dev    #1977   +/-   ##
=========================================
  Coverage     98.72%   98.72%           
=========================================
  Files            35       35           
  Lines          4628     4628           
=========================================
  Hits           4569     4569           
  Misses           59       59           

Flag	Coverage Δ
unittests	98.72% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[emilazy]

Very reasonable, both on waiting on CI and on the cast. I’ve pushed a much simpler fix to preserve the previous behaviour without changing the types.

[bmcfee]

Thanks @emilazy - hitting lint errors because the main branch is a bit stale right now. I'm rebasing to the 1.0dev branch just to get past that.

[emilazy]

FWIW, it looks like Numba plan to roll back the change due to it also causing issues with integers of mixed signedness. I assume the check for floats will return at some point though (this code doesn’t work in vanilla Python either), so this will still be relevant then.

[bmcfee]

Ok, thanks for the heads up @emilazy.

I guess this is now a fairly innocuous change, but I see no reason to delay merging. Is there anything else to do here for now?

[emilazy]

It should be good to go as far as I’m concerned, if you want to hit the button :) Thanks for the review!

[bmcfee]

Done - thanks again!