deps: use pandas-gbq to determine schema in load_table_from_dataframe

[tswast]

Thank you for opening a Pull Request! Before submitting your PR, there are a few things you can do to make sure it goes smoothly:

• Make sure to open an issue as a bug/issue before writing your code! That way we can discuss the change, evaluate designs, and agree on the general idea
• Ensure the tests and linter pass
• Code coverage does not decrease (if any source code was changed)
• Appropriate docs were updated (if necessary)

Towards go/pandas-gbq-and-bigframes-redundancy
🦕

[tswast]

tests/unit/test_client.py::TestClientUpload::test_load_table_from_dataframe_w_higher_scale_decimal128_datatype looks like a real failure and might need a change in pandas-gbq.

[tswast]

I've tested locally with googleapis/python-bigquery-pandas#844 and that PR should fix the remaining system test.

(venv) ➜  python-bigquery git:(b323176126-pandas-gbq) pytest tests/unit/test_client.py::TestClientUpload::test_load_table_from_dataframe_w_higher_scale_decimal128_datatype
==================================================== test session starts ====================================================
platform linux -- Python 3.12.6, pytest-8.3.4, pluggy-1.5.0
rootdir: /home/swast/src/github.com/googleapis/python-bigquery
configfile: pyproject.toml
collected 1 item                                                                                                            

tests/unit/test_client.py .                                                                                           [100%]

===================================================== 1 passed in 0.08s =====================================================

[tswast]

test_load_table_from_dataframe_w_partial_schema_extra_types

Looks like a real failure (the warning message changed slightly in pandas-gbq). I'll try to update the test to be a bit more robust.

[tswast]

@Linchin @chalmerlowe I believe this is finally ready for review. The failing checks are for infrastructure / Kokoro problems for a new setup, possibly?

[tswast]

Not sure is we should have defined what happens when you mix and match naive datetime (DATETIME in BQ) with datetime + timezone (TIMESTAMP in BQ). Since the test expects TIMESTAMP, I update this to be a datetime + timezone.

We still mix/match timezones, so I think that part is important to make sure we end up returning UTC values.

[chalmerlowe]

By and large, this looks good.
I have a question that might be a potential blocker.

[tswast]

Note: this means we are testing without pandas-gbq on Python 3.7.