Allow np.object dtypes into virtualfile_from_vectors

[weiji14]

Description of proposed changes

Loosen the check in virtualfile_from_vectors to allow for any string-like dtype (np.str, np.object) by performing the check using pd.api.types.is_string_dtype(). The array is then converted (if needed) to a proper np.str dtype before giving it to put_strings.

Why is this needed ?

This is one step in enabling text input into modules like:

• meca, see Wrap meca #516 (comment), ff164e6
• velo, see Wrap velo #525 (comment)

Those modules rely on pandas.DataFrame inputs, but a 'str' column in pandas is typically stored as an 'object' dtype (see https://stackoverflow.com/questions/21018654/strings-in-a-dataframe-but-dtype-is-object), unless users take due care to store them in the new pandas.StringDtype. Either way, when we convert these pandas.Series objects to a numpy array, their dtype becomes np.object rather than np.str (hence why our code needs to handle np.object too).

After this PR is merged, we can do something like:

            if kind == "matrix":
                if pd.api.types.is_numeric_dtype(data):
                    # all numeric dtypes (e.g. int, float)
                    file_context = lib.virtualfile_from_matrix(data)
                else:
                    # contains a column with object (str) dtype
                    file_context = lib.virtualfile_from_vectors(
                        *[data[column] for column in data]
                    )

Fixes #

Reminders

• Run make format and make check to make sure the code follows the style guide.
• Add tests for new features or tests that would have caught the bug that you're fixing.
• Add new public functions/methods/classes to doc/api/index.rst.
• Write detailed docstrings for all functions/methods.
• If adding new functionality, add an example to docstrings or tutorials.

Notes

• You can write /format in the first line of a comment to lint the code automatically

[seisman]

Great! I believe it solves a long-term headache when I use pandas with PyGMT.