Improve statistics.median_grouped on <=3.10

[AlexWaygood]

Providing an int or a float for the interval parameter appears to work, regardless of what weird and whacky inputs you supply for the data parameter:

>>> import statistics
>>> statistics.median_grouped([1, 5, 6, 8, 10], interval=2.0)
6.0
>>> statistics.median_grouped([1, 5, 6, 8, 10], interval=2)
6.0
>>> statistics.median_grouped([1.5, 5, 6, 8, 10], interval=2.0)
6.0
>>> from fractions import Fraction
>>> statistics.median_grouped([1.5, Fraction(5, 5), 6, 8, 10], interval=2.0)
6.0
>>> statistics.median_grouped([Fraction(1, 2), Fraction(4, 5), Fraction(2, 3)], interval=Fraction(1, 2))
0.6666666666666667
>>> from decimal import Decimal
>>> statistics.median_grouped([Decimal("0.5"), Decimal("0.7")], interval=4)
-1.3

(The union return is still required: if you supply a sequence with length exactly 1 to the data parameter, the sole member of the sequence is returned as-is, without any coercion to float.)

[JelleZijlstra]

Seems like we messed something up in mypy-primer.

[AlexWaygood]

Seems like we messed something up in mypy-primer.

The diffs were well formatted in #9629, so I don't think this can be due to any changes we've made over at typeshed. Something else must have changed.

But yeah, the same thing happened in #9635 and #9634.

[JelleZijlstra]

hauntsaninja/mypy_primer@bb842b5 should fix.

[JelleZijlstra]

@hauntsaninja fyi

[github-actions]

According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉