use dashes instead of underscores

[lev-blit]

resolves #15550

[github-actions]

Diff from mypy_primer, showing the effect of this PR on open source code:

pandas (https://github.com/pandas-dev/pandas)
- pandas/core/computation/ops.py:328: error: Need type annotation for "_binary_ops_dict" (hint: "_binary_ops_dict: dict[<type>, <type>] = ...")  [var-annotated]

[lev-blit]

I think the tests got stuck...

[brianschubert]

Did two sanity checks:

• ls -l stubs | wc -l reports that there's 198 stub packages. That, plus the 6 non-METADATA.toml files that were modified, equals the 204 modified files in this PR. So, every METADATA.toml file was touched.
• grep -R --include='METADATA.toml' _ stubs doesn't turn up any missed keys with underscores.