def test_excessive_aggregate(self):
print('\n\n *** test_excessive_aggregate ***\n')
a = dstats()
a.__set_state__({"n":6,"moments":np.zeros((5,1),np.float64)})
b = dstats()
b.__set_state__({"n":2,"moments":np.zeros((5,1),np.float64)})
a.aggregate(b)
def test_state(self):
print('\n\n *** test_state ***\n')
source = dstats(4,2)
depth = source.moments.shape[0]
width = source.moments.shape[1]
state = source.__get_state__()
source_stats = source.statistics()
print('source stats:\n%s' % source_stats)
discard = [ nt.assert_equals(source_stats[i][j],0)
for i in range(0,depth)
for j in range(0,width) ]
dest = dstats(1,1)
dest.__set_state__(state)
dest_stats = dest.statistics()
print('dest stats:\n%s' % dest_stats)
source.add([1,2])
source.add([2,3])
source.add([3.4])
state = source.__get_state__()
source_stats = source.statistics()
print('source stats:\n%s' % source_stats)
dest.__set_state__(state)
dest_stats = dest.statistics()
print('dest stats:\n%s' % dest_stats)
discard = [ nt.assert_equals(source_stats[i][j],dest_stats[i][j])
for i in range(0,depth)
for j in range(0,width) ]
def test_aggregation_all(self):
print('\n\n *** test_aggregation_all ***\n')
depth = 4
cols = 3
lhs = dstats(depth,cols)
lhs.add(np.array([10,20,30]))
lhs.add(np.array([10,20,30]))
lhs.add(np.array([10,20,30]))
lhs.add(np.array([10,20,30]))
lhs.add(np.array([10,20,30]))
print('lhs is: %s' % lhs.statistics(True))
rhs = dstats(depth,cols)
rhs.add(np.array([20,40,60]))
rhs.add(np.array([20,40,60]))
rhs.add(np.array([20,40,60]))
rhs.add(np.array([20,40,60]))
rhs.add(np.array([20,40,60]))
print('rhs is: %s' % rhs.statistics(True))
lhs.aggregate(rhs)
stats = lhs.statistics(True)
for moment in range(0,depth):
for col in range(0,cols):
print('col: %d, moment: %d, value: %22.15g' %
(col,moment+1,stats[moment,col]))
full = dstats(depth,cols)
full.add(np.array([10,20,30]))
full.add(np.array([10,20,30]))
full.add(np.array([10,20,30]))
full.add(np.array([10,20,30]))
full.add(np.array([10,20,30]))
full.add(np.array([20,40,60]))
full.add(np.array([20,40,60]))
full.add(np.array([20,40,60]))
full.add(np.array([20,40,60]))
full.add(np.array([20,40,60]))
print('FULL is: %s' % full.statistics(True))
nt.assert_equal(stats.shape[0], depth)
nt.assert_equal(stats.shape[1], cols)
nt.assert_almost_equal(stats[0,0], 15, places = 14)
nt.assert_almost_equal(stats[0,1], 30, places = 14)
nt.assert_almost_equal(stats[0,2], 45, places = 14)
nt.assert_almost_equal(stats[1,0], 5, places = 14)
nt.assert_almost_equal(stats[1,1], 10, places = 14)
nt.assert_almost_equal(stats[1,2], 15, places = 14)
nt.assert_almost_equal(stats[2,0], 0, places = 14)
nt.assert_almost_equal(stats[2,1], 0, places = 14)
nt.assert_almost_equal(stats[2,2], 0, places = 14)
nt.assert_almost_equal(stats[3,0], -3, places = 14)
nt.assert_almost_equal(stats[3,1], -3, places = 14)
nt.assert_almost_equal(stats[3,2], -3, places = 14)
def test_str(self):
print('\n\n *** test_str ***\n')
a = dstats()
print("STR: %s" % a)
nt.assert_equal(str(a), "dstats: 4 moments, 1 columns, 0 rows\n[[ 0.]\n [ 0.]\n [ 0.]\n [ 0.]]")
b = dstats(2,4)
print("STR: %s" % b)
nt.assert_equal(str(b), "dstats: 2 moments, 4 columns, 0 rows\n[[ 0. 0. 0. 0.]\n [ 0. 0. 0. 0.]]")
c = dstats(4,12)
print("STR: %s" % c)
nt.assert_equal(str(c), "dstats: 4 moments, 12 columns, 0 rows\n[[ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]]")
a.add(24)
a.add(42)
a.add(24)
a.add(42)
print("STD: %s" % a)
nt.assert_equal(str(a), "dstats: 4 moments, 1 columns, 4 rows\n[[ 33.]\n [ 324.]\n [ 0.]\n [ 26244.]]")
def test_repr(self):
print('\n\n *** test_repr ***\n')
a = dstats()
print("REPR: %s" % repr(a))
nt.assert_equal(repr(a), "dstats(n=0,moments=\n[[ 0.]\n [ 0.]\n [ 0.]\n [ 0.]])")
b = dstats(2,4)
print("REPR: %s" % repr(b))
nt.assert_equal(repr(b), "dstats(n=0,moments=\n[[ 0. 0. 0. 0.]\n [ 0. 0. 0. 0.]])")
c = dstats(4,12)
print("REPR: %s" % repr(c))
nt.assert_equal(repr(c), "dstats(n=0,moments=\n[[ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]])")
a.add(24)
a.add(42)
a.add(24)
a.add(42)
print("REPR: %s" % repr(a))
nt.assert_equal(repr(a), "dstats(n=4,moments=\n[[ 33.]\n [ 324.]\n [ 0.]\n [ 26244.]])")
def test_no_variance_stats(self):
print('\n\n *** test_no_variance_stats ***\n')
depth = 4
cols = 3
pstats = dstats(depth,cols)
pstats.add(np.array([10,20,30]))
pstats.add(np.array([10,20,30]))
pstats.add(np.array([10,20,30]))
pstats.add(np.array([10,20,30]))
pstats.add(np.array([10,20,30]))
pstats.add(np.array([20,20,60]))
pstats.add(np.array([20,20,60]))
pstats.add(np.array([20,20,60]))
pstats.add(np.array([20,20,60]))
pstats.add(np.array([20,20,60]))
stats = pstats.statistics(True)
for moment in range(0,depth):
for col in range(0,cols):
print('col: %d, moment: %d, value: %22.15g' %
(col,moment+1,stats[moment,col]))
nt.assert_equal(stats.shape[0], depth)
nt.assert_equal(stats.shape[1], cols)
nt.assert_almost_equal(stats[0,0], 15, places = 14)
nt.assert_almost_equal(stats[0,1], 20, places = 14)
nt.assert_almost_equal(stats[0,2], 45, places = 14)
nt.assert_almost_equal(stats[1,0], 5, places = 14)
nt.assert_almost_equal(stats[1,1], 0, places = 14)
nt.assert_almost_equal(stats[1,2], 15, places = 14)
nt.assert_almost_equal(stats[2,0], 1.27105748646260e-15, places = 14)
nt.assert_true(np.isnan(stats[2,1]))
nt.assert_almost_equal(stats[2,2], -1.56553681485797e-15, places = 14)
nt.assert_almost_equal(stats[3,0], -2, places = 14)
nt.assert_true(np.isnan(stats[3,1]))
nt.assert_almost_equal(stats[3,2], -2, places = 14)
def test_serdes(self):
print('\n\n *** test_serdes ***\n')
ds = dstats(4,9)
dstr = pickle.dumps(ds)
nt.assert_greater(len(dstr),100)
print('\nserialized empty dstats to length: %s' % len(dstr))
ds2 = pickle.loads(dstr)
depth = ds2.moments.shape[0]
width = ds2.moments.shape[1]
# N.B. This call raises a RuntimeWarning when generating the
# expected NaN values.
ds2_stats = ds2.statistics()
print('\ndeserialized empty ds2 statistics:\n %s' % ds2_stats)
discard = [ nt.assert_equals(ds2_stats[i][j],0) for i in range(0,depth) for j in range(0,width) ]
ds2.add([1,2,3,4,10,6,7,8,9])
ds2.add([9,8,7,6, 0,4,3,2,1])
ds2_stats = ds2.statistics()
print('\n2 element dstats statistics:\n %s' % ds2_stats)
discard = [ nt.assert_equals(ds2_stats[0][i],5) for i in range(0,len(ds2_stats[0])) ]
discard = [ nt.assert_equals(ds2_stats[1][i],[16,9,4,1,25,1,4,9,16][i]) for i in range(0,len(ds2_stats[1])) ]
discard = [ nt.assert_equals(ds2_stats[2][i],0) for i in range(0,len(ds2_stats[2])) ]
discard = [ nt.assert_equals(ds2_stats[3][i],-2) for i in range(0,len(ds2_stats[3])) ]
dstr2 = pickle.dumps(ds2)
nt.assert_greater(len(dstr2),100)
print('\nserialized 2 element dstats to length: %s' % len(dstr2))
ds3 = pickle.loads(dstr2)
ds3_stats = ds3.statistics()
print('\ndeserialized 2 element statistics:\n %s' % ds3_stats)
discard = [ nt.assert_equals(ds3_stats[0][i],5) for i in range(0,len(ds3_stats[0])) ]
discard = [ nt.assert_equals(ds3_stats[1][i],[16,9,4,1,25,1,4,9,16][i]) for i in range(0,len(ds3_stats[1])) ]
discard = [ nt.assert_equals(ds3_stats[2][i],0) for i in range(0,len(ds3_stats[2])) ]
discard = [ nt.assert_equals(ds3_stats[3][i],-2) for i in range(0,len(ds3_stats[3])) ]
def test_aggregation_large(self):
print('\n\n *** test_aggregation_large ***\n')
np.random.seed(0)
### Random data size
rows = 10
cols = 100
### Each accumulators size
depth = 4
width = 1
### Test data -- 10 rows of 100 columns each
test_arr = np.random.random((rows,cols))
print('Test data has shape: %d, %d' % test_arr.shape)
### Expected intermediate output
mid_mean = np.mean(test_arr,axis = 1)
mid_var = np.var(test_arr, axis = 1)
mid_skew = stats.skew(test_arr, axis = 1, bias = True)
mid_kurt = stats.kurtosis(test_arr, axis = 1, bias = True)
### Expected final output
final_mean = np.mean(test_arr)
final_var = np.var(test_arr)
final_skew = stats.skew(test_arr,axis=None,bias=True)
final_kurt = stats.kurtosis(test_arr,axis=None,fisher=True,bias=False)
### Create an object for each row and accumulate the data in that row
statsobjects = [ dstats(depth,width) for i in range(0,rows) ]
discard = [ statsobjects[i].add(test_arr[i,j])
for j in range(0,cols)
for i in range(0,rows)]
print('\nIntermediate Results\n')
for i in range(0,rows):
values = statsobjects[i].statistics()
print('Result %d mean: %11g, variance: %11g, skew: %11g, kurtosis: %11g' % (i,values[0],values[1],values[2],values[3]))
print('Expected mean: %11g, variance: %11g, skew: %11g, kurtosis: %11g' %(mid_mean[i],mid_var[i],mid_skew[i],mid_kurt[i]))
nt.assert_almost_equal(values[0], mid_mean[i], places = 14)
nt.assert_almost_equal(values[1], mid_var[i], places = 14)
nt.assert_almost_equal(values[2], mid_skew[i], places = 14)
nt.assert_almost_equal(values[3], mid_kurt[i], places = 14)
### Aggregate just a couple of intermediate results for diagnostic purposes.
diag_mean = np.mean(test_arr[1:3,],axis=None)
diag_var = np.var(test_arr[1:3,],axis=None)
diag_skew = stats.skew(test_arr[1:3,],axis=None,bias=True)
diag_kurt = stats.kurtosis(test_arr[1:3,],axis=None,fisher=True,bias=True)
### Aggregate result into the index 0 accumulator
discard = [ statsobjects[0].aggregate(statsobjects[i]) for i in range(1,rows) ]
values = statsobjects[0].statistics()
print('\nAggregated Results\n')
print('Result mean: %11g, variance: %11g, skew: %11g, kurtosis: %11g' % (values[0],values[1],values[2],values[3]))
print('Expected mean: %11g, variance: %11g, skew: %11g, kurtosis: %11g' % (final_mean,final_var,final_skew,final_kurt))
nt.assert_almost_equal(values[0], final_mean, places = 14)
nt.assert_almost_equal(values[1], final_var, places = 14)
nt.assert_almost_equal(values[2], final_skew, places = 14)
nt.assert_almost_equal(values[3], final_kurt, places = 5)
def test_excessive_statistics(self):
print('\n\n *** test_excessive_statistics ***\n')
a = dstats(5,1)
a.__set_state__({"n":6,"moments":np.zeros((5,1),np.float64)})
a.statistics()
def test_excessive_add(self):
print('\n\n *** test_excessive_add ***\n')
a = dstats(5,1)
a.add(5)
def test_unsupported(self):
print('\n\n *** test_unsupported_method ***\n')
a = dstats()
a.add(5)
a.remove(5)