BUG: Arc_KDTree.query(pt, k=1) raises AttributeError

[schmidtc]

Quering an Arc_KDTree with k = 1 results in AttributeError.
k=1 is the default.

>>> import libpysal
>>> pts = [(0,0), (1,0), (0,1), (1,1)]

>>> type(libpysal.cg.KDTree(pts))
<class 'scipy.spatial.ckdtree.cKDTree'>
>>> libpysal.cg.KDTree(pts).query((.5,.5), 1)
(0.7071067811865476, 0)

>>> type(libpysal.cg.KDTree(pts, distance_metric="Arc"))
<class 'libpysal.cg.kdtree.Arc_KDTree'>
>>> libpysal.cg.KDTree(pts, distance_metric="Arc").query((.5,.5))
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/.../site-packages/libpysal/cg/kdtree.py", line 161, in query
    dims = len(d.shape)
AttributeError: 'float' object has no attribute 'shape'

>>> libpysal.cg.KDTree(pts, distance_metric="Arc").query((.5,.5), 2)
(array([78.62319385, 78.62319385]), array([2, 3]))

• Platform information: posix darwin, posix.uname_result(sysname='Darwin', nodename='ZG04075', release='19.6.0', version='Darwin Kernel Version 19.6.0: Sun Jul 5 00:43:10 PDT 2020; root:xnu-6153.141.1~9/RELEASE_X86_64', machine='x86_64')
• Python version: 3.7.7 (default, Mar 26 2020, 10:32:53) [Clang 4.0.1 (tags/RELEASE_401/final)]
• SciPy version: 1.4.1
• NumPy version: 1.19.1
• LibPySAL version: 4.3.0

[sjsrey]

You are holding it wrong ;->

>>> libpysal.cg.KDTree(pts, distance_metric="Arc").query([(.5,.5)])                                                                                                                                                 
(array([78.62319385]), array([2]))

The doc is a bit ambiguous as the query argument is:

 x : array-like, last dimension self.m
            query points are lng/lat.

Definitely a bug as the behavior with 'Arc' differs from the default case.

Thanks for reporting.

[schmidtc]

Thanks, that makes the return value consistent with k>1 too, which makes things easier.

[ljwolf]

I suppose it's really not a problem to throw a numpy.asarray() around input? We're very lax about this, but we should try to use a similar like... _validate_vector on all our inputs?