Add __array_interface__ to cudf.pandas numpy.ndarray proxy

[mroeschke]

Description

closes #15926

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[Matt711]

Hey @mroeschke, I pulled your changes to try to test them locally. The example in the issue worked for me. Do you know of another example that produces a numpy proxy ndarray other than the example in the issue (which used df.values)?

# file: reproduce.py
import cudf

print(cudf.__version__)         # 24.08.00a101

import pandas as pd
df = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
arr = df.values

print(arr.__cuda_array_interface__) # good
print(arr.__array_interface__)  # AttributeError

[trivialfis]

produces a numpy proxy ndarray other than the example in the issue (which used df.values)

Not familiar with the internal, maybe DataFrame.to_numpy()?

[Matt711]

This looks good to me. I pulled and tested it earlier.

[mroeschke]

/merge