Add bias to scores pandas api

[wuxx66]

No description provided.

[John-Sharples]

I'm not sure a general definition of Bias is needed here. Suggest deleting the first sentence to make this consistent with other docstrings in this file.

Suggested change

	Bias is commonly defined as the mean signed difference between forecast and observed values. A detailed explanation is on https://scores.readthedocs.io/en/stable/tutorials/Additive_and_multiplicative_bias.html
	A detailed explanation is on https://scores.readthedocs.io/en/stable/tutorials/Additive_and_multiplicative_bias.html

[wuxx66]

Removed first line definition

[John-Sharples]

Since you're expanding the input types for additive_bias you need to test for xarray inputs as well.

For 100% test coverage you'll need to check the exception is correctly raised.

Something like

 with pytest.raises(ValueError):
    scores.continuous.additive_bias(fcst=fcst, obs=obs, weights=xr.DataArray([1,2,3]))

[wuxx66]

Done the test

[tennlee]

@John-Sharples why do we need this in the pandas-specific API? Is the idea it's xarray plus pandas, or just pandas only?

[John-Sharples]

My mistake, in putting it here it probably belongs in tests/continuous/test_standard.py

[John-Sharples]

We still need a test checking we get the expected answer when the inputs are xarray like objects. Something like:

def test_additive_bias_xarray():

    fcst =xr.DataArray([1, 3, 1, 3, 2, 2, 2, 1, 1, 2, 3])
    obs = xr.DataArray([1, 1, 1, 2, 1, 2, 1, 1, 1, 3, 1])

    expected = 0.5455

    result = scores.continuous.additive_bias(fcst=fcst, obs=obs)

    assert isinstance(result, numbers.Real)
    assert round(float(result), PRECISION) == expected
    
    # And possibly with weights and reduce_dims
    
    result = scores.continuous.additive_bias(fcst=fcst, obs=obs, reduce_dims = ...., weights = ....)
    
    assert result == ....
    

[tennlee]

Do we? This is for the pandas API, where xarraylike objects aren't expected. Do we need the pandas API to also handle xarraylikes?

[John-Sharples]

Do we? This is for the pandas API, where xarraylike objects aren't expected. Do we need the pandas API to also handle xarraylikes?

Apologies - I was mistaken about which function we were testing here. Current test looks good.

Happy to approve once confilcts are resolved.

[John-Sharples]

@tennlee
@wuxx66 and I had a call to sort out the confusion I created. We now have appropriate tests in the right place, maintaining 100% coverage.