fix(python): support corr() for single-column DataFrames

[Wainberg]

I also took the liberty of adding rowvar=False in np.corrcoef(), to avoid having to transpose the array before correlating.

Before:

>>> pl.DataFrame([1, 2, 3]).corr()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "polars/dataframe/frame.py", line 10223, in corr
    return DataFrame(np.corrcoef(self.to_numpy().T, **kwargs), schema=self.columns)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "polars/dataframe/frame.py", line 430, in __init__
    raise TypeError(msg)
TypeError: DataFrame constructor called with unsupported type 'float64' for the `data` parameter

After:

>>> pl.DataFrame([1, 2, 3]).corr()
shape: (1, 1)
┌──────────┐
│ column_0 │
│ ---      │
│ f64      │
╞══════════╡
│ 1.0      │
└──────────┘

[ritchie46]

Good fix, thanks! Can you add a test?

[Wainberg]

Just added a test :)