[DOC] clarify distance mode in graph creation functions.

[ixjlyons]

The documentation for the mode argument of radius_neighbors_graph and kneighbors_graph states that if 'distance' is specified, the matrix returned contains the Euclidean distances between neighbors, but actually the distances calculated according to metric are returned.

Example:

from sklearn import neighbors

X = [[0, 0], [1, 1], [2, 2]]
w1 = neighbors.radius_neighbors_graph(X, 2, mode='distance', metric='manhattan')
w2 = neighbors.kneighbors_graph(X, 1, mode='distance', metric='manhattan')

print(w1.toarray())
print(w2.toarray())

Output:

[[ 0.  2.  0.]
 [ 2.  0.  2.]
 [ 0.  2.  0.]]
[[ 0.  2.  0.]
 [ 2.  0.  0.]
 [ 0.  2.  0.]]

[jnothman]

thanks!