Release a patched version to PyPI for NumPy compatibility

[carlthome]

Noticed that people have started employing the work around of upper bounding NumPy in their usage of mir_eval now (example) which tells me that it's high time to make a new release.

What steps are remaining to take to make mir_eval compatible with the latest NumPy again?

IIRC it's only a matter of running the GitHub Actions release workflow after bumping the version number in the code but I'm also not fully aware of relevant changes since 0.7 was built, so is it an incoming 0.8 or 0.7.1?

Something I could help with to expedite this? I'm available to help.

[bmcfee]

What steps are remaining to take to make mir_eval compatible with the latest NumPy again?

AFAIK it's really just down to replacing uses of deprecated dtype aliases (np.complex, np.float) for python builtins (complex, float). And of course, updating our CI environments to make sure that it's all tested appropriately.