def test_tree_search_types(self):
data = np.matrix(np.random.normal(size=3000)).reshape(1000, 3)
query = np.matrix(np.random.normal(size=3000)).reshape(1000, 3)
n_standard = pyann.nn2(data, query, k=5, searchtype='standard')
n_priority = pyann.nn2(data, query, k=5, searchtype='priority')
n_bd_standard = pyann.nn2(data,
query,
k=5,
searchtype='standard',
treetype='bd')
n_bd_priority = pyann.nn2(data,
query,
k=5,
searchtype='priority',
treetype='bd')
self.assertTrue(
np.array_equal(n_standard.to_array(), n_priority.to_array()))
self.assertTrue(
np.array_equal(n_standard.to_array(), n_bd_standard.to_array()))
self.assertTrue(
np.array_equal(n_standard.to_array(), n_bd_priority.to_array()))
def test_wrong_input_type_treetype(self):
with self.assertRaises(
TypeError, msg="treetype must be str; detected <class 'int'>"):
pyann.nn2(data=np.random.normal(size=10),
query=np.random.normal(size=5),
treetype=1,
k=1)
def test_wrong_str_phrase_searchtype(self):
with self.assertRaises(
ValueError,
msg=
"unrecognized option 'random' for searchtype; must be one of ['standard', 'priority', 'radius']"
):
pyann.nn2(data=np.random.normal(size=10),
query=np.random.normal(size=5),
searchtype='random',
k=1)
def test_wrong_str_phrase_treetype(self):
with self.assertRaises(
ValueError,
msg=
"unrecognized option 'random' for treetype; must be one of ['kd', 'bd']"
):
pyann.nn2(data=np.random.normal(size=10),
query=np.random.normal(size=5),
treetype='random',
k=1)
def test_mix_mat_vector(self):
query = np.random.normal(size=10)
data = np.matrix(query).reshape(10, 1)
s = pyann.nn2(data, query, k=1)
s_inv = pyann.nn2(query, data, k=1)
s_ind = pyann.nn2(query, query, k=1)
self.assertTrue(np.array_equal(s.to_array(), s_inv.to_array()))
self.assertTrue(np.array_equal(s.to_array(), s_ind.to_array()))
def test_fixed_large_radius(self):
data = np.matrix(np.random.normal(size=3000)).reshape(1000, 3)
query = np.matrix(np.random.normal(size=3000)).reshape(1000, 3)
n_standard = pyann.nn2(data, query, k=5, searchtype='standard')
n_rad = pyann.nn2(data, query, k=5, searchtype='radius', radius=20.0)
n_bd_rad = pyann.nn2(data,
query,
k=5,
searchtype='radius',
radius=20.0,
treetype='bd')
self.assertTrue(np.array_equal(n_standard.to_array(),
n_rad.to_array()))
self.assertTrue(
np.array_equal(n_standard.to_array(), n_bd_rad.to_array()))
def test_basic_use_1(self):
s = pyann.nn2(np.matrix([[1, 0], [2, 0]]),
np.matrix([[1.01, 0], [3, 0], [4.0, 0]]),
k=1)
nn_match = np.array_equal(s.nn_idx, np.matrix([1, 2, 2]).reshape(3, 1))
dist_match = np.array_equal(
s.nn_dists,
np.matrix([0.010000000000000009, 1, 2],
dtype=np.double).reshape(3, 1))
self.assertTrue(
np.array_equal(np.array([nn_match, dist_match]),
np.array([True, True])))
def test_nn2_with_identical_point(self):
data = np.matrix([[1, 0], [2, 0]])
query = np.matrix([[1, 0], [3, 0], [4, 0]])
s = pyann.nn2(data, query, k=1)
nn_dists = np.matrix([[0], [1], [2]], dtype=np.double).reshape(3, 1)
nn_idx = np.matrix([[1], [2], [2]]).reshape(3, 1)
nn_match = np.array_equal(s.nn_idx, nn_idx)
dist_match = np.array_equal(s.nn_dists.round(8), nn_dists.round(8))
self.assertTrue(
np.array_equal(np.array([nn_match, dist_match]),
np.array([True, True])))
def test_basic_use_2(self):
data = np.matrix(
[[-0.626453810742332, 1.51178116845085, 0.918977371608218],
[0.183643324222082, 0.389843236411431, 0.782136300731067],
[-0.835628612410047, -0.621240580541804, 0.0745649833651906],
[1.59528080213779, -2.2146998871775, -1.98935169586337],
[0.329507771815361, 1.12493091814311, 0.61982574789471],
[-0.820468384118015, -0.0449336090152309, -0.0561287395290008],
[0.487429052428485, -0.0161902630989461, -0.155795506705329],
[0.738324705129217, 0.9438362106853, -1.47075238389927],
[0.575781351653492, 0.821221195098089, -0.47815005510862],
[-0.305388387156356, 0.593901321217509, 0.417941560199702]])
query = np.matrix(
[[1.35867955152904, -0.164523596253587, 0.398105880367068],
[-0.102787727342996, -0.253361680136508, -0.612026393250771],
[0.387671611559369, 0.696963375404737, 0.341119691424425],
[-0.0538050405829051, 0.556663198673657, -1.12936309608079],
[-1.37705955682861, -0.68875569454952, 1.43302370170104],
[-0.41499456329968, -0.70749515696212, 1.98039989850586],
[-0.394289953710349, 0.36458196213683, -0.367221476466509],
[-0.0593133967111857, 0.768532924515416, -1.04413462631653],
[1.10002537198388, -0.112346212150228, 0.569719627442413],
[0.763175748457544, 0.881107726454215, -0.135054603880824]])
nn_idx = np.matrix([[7, 2, 9, 5, 10], [7, 6, 3, 9,
10], [5, 2, 10, 9, 7],
[9, 8, 7, 6, 10], [3, 6, 10, 2,
1], [2, 3, 10, 6, 5],
[6, 10, 7, 9, 3], [9, 8, 7, 6,
10], [7, 2, 5, 9, 10],
[9, 5, 7, 2, 10]]).reshape(10, 5)
nn_dists = np.matrix(
[[
1.04301819567468, 1.35481071903339, 1.53376883835267,
1.66464626768377, 1.82885863701822
],
[
0.782785251535478, 0.931412898425681, 1.06948509416105,
1.27793083718641, 1.3489757618116
],
[
0.514019762075842, 0.574844301957827, 0.704879811283639,
0.849722423887674, 0.874908208472254
],
[
0.943635726058299, 0.945472961596115, 1.25256907321914,
1.44966319723756, 1.56806650285981
],
[
1.46393847125353, 1.71518967014802, 1.95552547157996,
2.00569777023645, 2.38117935086023
],
[
1.73157573316917, 1.95360681279363, 2.03639936314308,
2.179643484087, 2.40067304459157
],
[
0.66791447696244, 0.822782991518318, 0.983420985126367,
1.07789743302845, 1.16696237017602
],
[
0.852326186446708, 0.921390522381496, 1.30532148465522,
1.4890400761415, 1.49288843286122
],
[
0.954406845715277, 1.06633538007411, 1.45844529983923,
1.49813576432115, 1.58019302443234
],
[
0.395496580572611, 0.904080753178077, 0.938940949941509,
1.19098185772077, 1.23698083599426
]],
dtype=np.double).reshape(10, 5)
s = pyann.nn2(data, query, k=5)
nn_match = np.array_equal(s.nn_idx, nn_idx)
dist_match = np.array_equal(s.nn_dists.round(8), nn_dists.round(8))
self.assertTrue(
np.array_equal(np.array([nn_match, dist_match]),
np.array([True, True])))
def test_input_diff_dim(self):
data = np.matrix(np.random.normal(size=20)).reshape(10, 2)
query = np.asarray(data[:, 1])
with self.assertRaises(pyann.DimensionError):
pyann.nn2(data, query, k=1)
def test_vector_input_length(self):
s = pyann.nn2(data=np.random.normal(size=10),
query=np.random.normal(size=5),
k=1)
self.assertTrue(len(s.nn_idx) == 5)
def test_all_na(self):
data = np.matrix(np.random.normal(size=10)).reshape(5, 2)
query = np.matrix([np.nan] * 10).reshape(5, 2)
with self.assertRaises(ValueError):
pyann.nn2(data, query, k=1)
def test_zero_columns(self):
data = np.matrix([]).reshape(90, 0)
with self.assertRaises(ValueError):
pyann.nn2(data)
