KDTree example in docs produces DeprecationWarning

[weatherfrog]

The basic KDTree example from the docs produces a DeprecatioinWarning (tested with 0.17.1):

X = np.random.random((10, 3))  # 10 points in 3 dimensions
tree = KDTree(X, leaf_size=2)
dist, ind = tree.query(X[0], k=3)

==>

[...]/lib/python3.4/site-packages/sklearn/utils/validation.py:386: DeprecationWarning:
Passing 1d arrays as data is deprecated in 0.17 and willraise ValueError in 0.19.
Reshape your data either using X.reshape(-1, 1) if your data has a single feature or
X.reshape(1, -1) if it contains a single sample.
  DeprecationWarning)

I'm not (yet) very familiar with scikit-learn. But doesn't it seem "wrong" to say "Passing 1d arrays as data is deprecated" when the X passed is actually 2d?

[weatherfrog]

Oh, data refers to the first argument of query(), not to the input points. That makes sense, of course. I think that argument – which, according to docs, is an array of points to query – should be wrapped in a list then. That makes the DeprecationWarning go away:

dist, ind = tree.query([X[0]], k=3)

[jnothman]

yes, wrapping in a list being much like X.reshape(1, -1)

[weatherfrog]

My point was: the docs should probably be updated. :-)

[jnothman]

You're right. Sorry :/

[jnothman]

Feel free to submit a PR fixing sklearn/neighbors/binary_tree.pxi

[jnothman]

It might also be interesting to try creating doctest fixtures that raise warnings as exceptions.