Implement cftime vectorization as discussed in PR #8322

[antscloud]

As discussed in #8322, here is the test for implementing the vectorization

Only this test seems to fail in test_coding_times.py :

xarray/xarray/tests/test_coding_times.py

Lines 1061 to 1071 in f895dc1

	@requires_cftime
	@pytest.mark.parametrize("freq", ["U", "L", "S", "T", "H", "D"])
	def test_encode_decode_roundtrip_cftime(freq) -> None:
	initial_time = cftime_range("0001", periods=1)
	times = initial_time.append(
	cftime_range("0001", periods=2, freq=freq) + timedelta(days=291000 * 365)
	)
	variable = Variable(["time"], times)
	encoded = conventions.encode_cf_variable(variable)
	decoded = conventions.decode_cf_variable("time", encoded, use_cftime=True)
	assert_equal(variable, decoded)

I don't really understand why though if you have an idea

[kmuehlbauer]

@antscloud Nice catch! I have two suggestion to get the test suite running.

[headtr1ck]

@antscloud are you willing to continue this? It would be a good addition.

[antscloud]

Hi, sorry i completely forgot about this issue, i've implemented the suggestions and all the tests appear to pass, I hope it works in the CI 🤞

[headtr1ck]

BENCHMARKS NOT SIGNIFICANTLY CHANGED.

I didn't check in detail if we have a test that might cover this already and if yes, what the changes are.

[spencerkclark]

I don't think we currently do, but I think this new benchmark should cover it: #9262.

[spencerkclark]

Thanks @antscloud! I think this looks good to me now. Could you add a what's new entry under a new Performance heading?

[headtr1ck]

| Change | Before [91a52dc] | After [3512bee] | Ratio | Benchmark (Parameter) |
|----------|----------------------|---------------------|---------|-----------------------------------------------------------|
| - | 139±0.9ms | 42.3±0.7ms | 0.3 | coding.EncodeCFDatetime.time_encode_cf_datetime('noleap') |

3 times faster, not bad

[headtr1ck]

Looks good to me!

[headtr1ck]

mypy complains here.
Not sure if this is an issue of wrong typing of numpy, or if this is not a recommended way of calling equals. (I think the None is the problem).