[SciPy-user] problem with spatial.kdtree.sparse_distance_matrix

[josef.pktd at gmail.com]

I would like to get the distance_matrix of all point in a 2d array,
but it looks like kdtree cannot create a sparse distance matrix with
itself. Is this intentional, a bug, or am I doing something wrong?
Using a small distortion in the data works.
I followed the example in the testfile (BTW: in class
test_sparse_distance_matrix, M is often empty in the examples I tried,
with given r=0.3).

Josef


>>> from scipy import spatial as ssp
>>> r = 1
>>> xs2 = np.random.randn(4,50)
>>> T1 = ssp.KDTree(xs2,leafsize=2)
>>> M = T1.sparse_distance_matrix(T1, r)
Traceback (most recent call last):
  File "<pyshell#73>", line 1, in <module>
    M = T1.sparse_distance_matrix(T1, r)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 628, in sparse_distance_matrix
    other.tree, Rectangle(other.maxes, other.mins))
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 623, in traverse
    traverse(node1.less,less1,node2.less,less2)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 623, in traverse
    traverse(node1.less,less1,node2.less,less2)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 611, in traverse
    result[i,j] = d
  File "\Programs\Python25\Lib\site-packages\scipy\sparse\dok.py",
line 222, in __setitem__
    del self[(i,j)]
KeyError: (0, 0)
>>> T1b = ssp.KDTree(xs2.copy(),leafsize=2)
>>> M = T1.sparse_distance_matrix(T1b, r)
Traceback (most recent call last):
  File "<pyshell#75>", line 1, in <module>
    M = T1.sparse_distance_matrix(T1b, r)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 628, in sparse_distance_matrix
    other.tree, Rectangle(other.maxes, other.mins))
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 623, in traverse
    traverse(node1.less,less1,node2.less,less2)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 623, in traverse
    traverse(node1.less,less1,node2.less,less2)
  File "C:\Josef\_progs\building\scipy\scipy-trunk-new-r5551\dist\scipy-0.8.0.dev5551.win32\Programs\Python25\Lib\site-packages\scipy\spatial\kdtree.py",
line 611, in traverse
    result[i,j] = d
  File "\Programs\Python25\Lib\site-packages\scipy\sparse\dok.py",
line 222, in __setitem__
    del self[(i,j)]
KeyError: (0, 0)


using a small distortion works:

>>> T1b = ssp.KDTree(xs2.copy()+1e-8,leafsize=2)
>>> M = T1.sparse_distance_matrix(T1b, r)
>>> len(M)
110