LogisticRegression's predict_proba(X) returns (n,) instead of (n,2) for binary class

Just like [sklearn's precit_proba(X)](http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression.predict_proba), [Dask's documentation](https://ml.dask.org/modules/generated/dask_ml.linear_model.LogisticRegression.html#dask_ml.linear_model.LogisticRegression.predict_proba) says it returns an "array-like, shape = [n_samples, n_classes]".
However, it returns an "array-like, shape = [n_samples, 1]", for a binary classifier.
E.g. run this:
```
import numpy as np
from sklearn.linear_model import LogisticRegression
from dask_ml.linear_model import LogisticRegression as DaskLogit

# Data
X = np.random.rand(200,20)
y = np.random.rand(200).round(0) # random binary labels

# Dask
lrdask = DaskLogit()
lrdask.fit(X, y)
print("Logit Reg 'predict_proba' shape Dask\t", lrdask.predict_proba(X).shape)

# SKLearn
lrsklr = LogisticRegression()
lrsklr.fit(X, y)
print("Logit Reg 'predict_proba' shape SKLearn\t", lrsklr.predict_proba(X).shape)
```

Of course, no information is lost, but this *does* break stuff when used in other code.
E.g. in GridSearchCV's scorer I get:
```
C:\ProgramData\Anaconda3\lib\site-packages\sklearn\metrics\scorer.py in __call__(self, clf, X, y, sample_weight)
184
185 if y_type == "binary":
--> 186 y_pred = y_pred[:, 1]
187 elif isinstance(y_pred, list):
188 y_pred = np.vstack([p[:, -1] for p in y_pred]).T

IndexError: too many indices for array
```

predict_proba(X)'s return value is indeed implemented differently in sklearn and dask-ml:
[sklearn](https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/linear_model/base.py):
```
def _predict_proba_lr(self, X):
...
return np.vstack([1 - prob, prob]).T
```

[dask-ml](https://github.com/dask/dask-ml/blob/master/dask_ml/linear_model/glm.py):
```
def predict_proba(self, X):
...
return sigmoid(dot(X_, self._coef))
```

Is this intentional and would changing this break code that currently expects Dask to return (n,) and *not* (n,2)? Or is this an oversight, and should it return (n,2) as per documentation?