def save(data, outputdir, outputfile, outputformat):
"""
Save data to a variety of formats
Automatically determines whether data is an array
or an RDD and handles appropriately
For RDDs, data are sorted and reshaped based on the keys
:param data: RDD of key value pairs or array
:param outputdir: Location to save data to
:param outputfile: file name to save data to
:param outputformat: format for data ("matlab", "text", or "image")
"""
filename = os.path.join(outputdir, outputfile)
if (outputformat == "matlab") | (outputformat == "text"):
if isrdd(data):
dims = getdims(data)
data = subtoind(data, dims.max)
keys = data.map(lambda (k, _): int(k)).collect()
nout = size(data.first()[1])
if nout > 1:
for iout in range(0, nout):
result = data.map(lambda (_, v): float16(v[iout])).collect()
result = array([v for (k, v) in sorted(zip(keys, result), key=lambda (k, v): k)])
if outputformat == "matlab":
savemat(filename+"-"+str(iout)+".mat",
mdict={outputfile+str(iout): squeeze(transpose(reshape(result, dims.num[::-1])))},
oned_as='column', do_compression='true')
if outputformat == "text":
savetxt(filename+"-"+str(iout)+".txt", result, fmt="%.6f")
else:
result = data.map(lambda (_, v): float16(v)).collect()
result = array([v for (k, v) in sorted(zip(keys, result), key=lambda (k, v): k)])
def testBrokenTypes(self):
with self.assertRaisesWithPredicateMatch(TypeError, "Categorical"):
distributions_py.Mixture(None, [])
cat = distributions_py.Categorical([0.3, 0.2])
# components must be a list of distributions
with self.assertRaisesWithPredicateMatch(
TypeError, "all .* must be Distribution instances"):
distributions_py.Mixture(cat, [None])
with self.assertRaisesWithPredicateMatch(TypeError, "same dtype"):
distributions_py.Mixture(
cat, [
distributions_py.Normal(
mu=[1.0], sigma=[2.0]), distributions_py.Normal(
mu=[np.float16(1.0)], sigma=[np.float16(2.0)])
])
with self.assertRaisesWithPredicateMatch(ValueError, "non-empty list"):
distributions_py.Mixture(distributions_py.Categorical([0.3, 0.2]), None)
with self.assertRaisesWithPredicateMatch(TypeError,
"either be continuous or not"):
distributions_py.Mixture(
cat, [
distributions_py.Normal(
mu=[1.0], sigma=[2.0]), distributions_py.Bernoulli(
dtype=dtypes.float32, logits=[1.0])
])
def testReprWorksCorrectlyScalar(self):
normal = tfd.Normal(loc=np.float16(0), scale=np.float16(1))
self.assertEqual(
("<tf.distributions.Normal"
" 'Normal'"
" batch_shape=()"
" event_shape=()"
" dtype=float16>"), # Got the dtype right.
repr(normal))
chi2 = tfd.Chi2(df=np.float32([1., 2.]), name="silly")
self.assertEqual(
("<tf.distributions.Chi2"
" 'silly'" # What a silly name that is!
" batch_shape=(2,)"
" event_shape=()"
" dtype=float32>"),
repr(chi2))
exp = tfd.Exponential(rate=array_ops.placeholder(dtype=dtypes.float32))
self.assertEqual(
("<tf.distributions.Exponential"
" 'Exponential'"
" batch_shape=<unknown>"
" event_shape=()"
" dtype=float32>"),
repr(exp))
def testStrWorksCorrectlyScalar(self):
normal = tfd.Normal(loc=np.float16(0), scale=np.float16(1))
self.assertEqual(
("tf.distributions.Normal("
"\"Normal\", "
"batch_shape=(), "
"event_shape=(), "
"dtype=float16)"), # Got the dtype right.
str(normal))
chi2 = tfd.Chi2(df=np.float32([1., 2.]), name="silly")
self.assertEqual(
("tf.distributions.Chi2("
"\"silly\", " # What a silly name that is!
"batch_shape=(2,), "
"event_shape=(), "
"dtype=float32)"),
str(chi2))
exp = tfd.Exponential(rate=array_ops.placeholder(dtype=dtypes.float32))
self.assertEqual(
("tf.distributions.Exponential(\"Exponential\", "
# No batch shape.
"event_shape=(), "
"dtype=float32)"),
str(exp))
def __init__(self,
Agent,
alpha,
gamma,
epsilon):
'''
Fill all values of Q based on a given optimistic value.
'''
# Set the agent for this QLearning session
self.Agent = Agent
self.alpha = alpha
self.gamma = gamma
self.epsilon = epsilon
self.policy = dict()
self.Q = dict()
self.V = dict()
S = set( [ (i,j) for i in range(-5,6) for j in range(-5,6)] )
for s in S:
self.V[s] = numpy.float16( 0 )
self.Q[s] = dict()
self.policy[s] = dict()
for a in self.Agent.actions:
self.policy[s][a] = numpy.float16( 1.0 / len( self.Agent.actions ) )
for o in self.Agent.actions:
self.Q[s][(a,o)] = numpy.float16( 0.0 )
def callback_1(self, data):
# print "difference_cal receives position_1 update. Processing...."
self._position_1_x = data.x
self._position_1_y = data.y
self._position_1_ID = data.ID
# xmin = self._position_1_x + 0.1
# xmax = self._position_1_x + 0.6
# ymin = self._position_1_y - 0.3
# ymax = self._position_1_y + 0.3
# if self._position_0_x > xmin and self._position_0_x < xmax and self._position_0_y > ymin and self._position_0_y < ymax:
# w1 = 1
# else:
# w1 = 0
# xmin = self._position_1_x - 0.3
# xmax = self._position_1_x + 0.3
# ymin = self._position_1_y + 0.1
# ymax = self._position_1_y + 0.6
# if self._position_0_x > xmin and self._position_0_x < xmax and self._position_0_y > ymin and self._position_0_y < ymax:
# w2 = 1
# else:
# w2 = 0
# xmin = self._position_1_x - 0.6
# xmax = self._position_1_x - 0.1
# ymin = self._position_1_y - 0.3
# ymax = self._position_1_y + 0.3
# if self._position_0_x > xmin and self._position_0_x < xmax and self._position_0_y > ymin and self._position_0_y < ymax:
# w3 = 1
# else:
# w3 = 0
# xmin = self._position_1_x - 0.3
# xmax = self._position_1_x + 0.3
# ymin = self._position_1_y - 0.6
# ymax = self._position_1_y - 0.1
# if self._position_0_x > xmin and self._position_0_x < xmax and self._position_0_y > ymin and self._position_0_y < ymax:
# w4 = 1
# else:
# w4 = 0
if not self._position_0_x:
self._position_0_x = [-1]*len(self._position_1_x)
self._position_0_y = [-1]*len(self._position_1_x)
w1 = list(np.float16(np.array(self._position_0_x)) - np.float16(np.array(self._position_1_x)))
w2 = list(np.float16(np.array(self._position_0_y)) - np.float16(np.array(self._position_1_y)))
if abs(w1[0])<0.3 and abs(w2[0])<0.3:
rospy.loginfo("The two car are too close with distance of x and y: %s, %s", str(w1[0]), str(w1[2]))
w3 = list()
for i in range(len(w1)):
if abs(w1[i])<1 and abs(w2[i])<1:
w3.append(1)
else:
w3.append(0)
# print w3
w1 = [0]*len(w1)
w2 = [0]*len(w1)
w4 = [0]*len(w1)
# print list(self._position_0_x)
# print list(self._position_0_y)
self.pub_2.publish(w1, w2, w3, w4, list(self._position_0_x), list(self._position_0_y), self._position_1_ID)
# print "disturbance_signal_1 sent with w3 = %s, and ID = %s."%(str(w3), str(self._position_1_ID))
print "disturbance_signal_1 ID = %s."%(str(self._position_1_ID))
def callback_1(self, data):
# print "difference_cal receives position_1 update. Processing...."
self._position_1_x = data.x
self._position_1_y = data.y
self._position_1_ID = data.ID
if not self._position_0_x:
self._position_0_x = [-1]*len(self._position_1_x)
self._position_0_y = [-1]*len(self._position_1_x)
w1 = list(np.float16(np.array(self._position_0_x)) - np.float16(np.array(self._position_1_x)))
w2 = list(np.float16(np.array(self._position_0_y)) - np.float16(np.array(self._position_1_y)))
if self._position_1_ID < self._horizon:
if abs(w1[self._position_1_ID])<0.3 and abs(w2[self._position_1_ID])<0.3:
rospy.loginfo("The two car are getting close.")
else:
if abs(w1[self._horizon])<0.3 and abs(w2[self._horizon])<0.3:
rospy.loginfo("The two car are getting close.")
w3 = list()
for i in range(len(w1)):
if abs(w1[i])<1 and abs(w2[i])<1:
w3.append(1)
else:
w3.append(0)
# print w3
# w1 = [0]*len(w1)
# w2 = [0]*len(w1)
# w4 = [0]*len(w1)
# print list(self._position_0_x)
# print list(self._position_0_y)
self.pub_2.publish(w3, list(self._position_0_x), list(self._position_0_y), self._position_1_ID)
# print "disturbance_signal_1 sent with w3 = %s, and ID = %s."%(str(w3), str(self._position_1_ID))
if self._position_1_ID < self._horizon:
print "disturbance_signal_1 ID = %s. pos is %s, %s."%(str(self._position_1_ID),str(self._position_1_x[self._position_1_ID]), str(self._position_1_y[self._position_1_ID]))
else:
print "disturbance_signal_1 ID = %s. pos is %s, %s."%(str(self._position_1_ID),str(self._position_1_x[self._horizon]), str(self._position_1_y[self._horizon]))
def estimate(self, iterations=100, tolerance=1e-5):
last_rmse = None
# the algorithm will converge, but really slow
# use MF's initialize latent parameter will be better
for iteration in xrange(iterations):
# update item & user parameter
self._update_item_params()
self._update_user_params()
# update item & user_features
self._udpate_item_features()
self._update_user_features()
# compute RMSE
# train errors
train_preds = self.predict(self.train)
train_rmse = RMSE(train_preds, np.float16(self.train[:, 2]))
# validation errors
validation_preds = self.predict(self.validation)
validation_rmse = RMSE(
validation_preds, np.float16(self.validation[:, 2]))
self.train_errors.append(train_rmse)
self.validation_erros.append(validation_rmse)
print "iterations: %3d, train RMSE: %.6f, validation RMSE: %.6f " % (iteration + 1, train_rmse, validation_rmse)
# stop if converge
if last_rmse:
if abs(train_rmse - last_rmse) < tolerance:
break
last_rmse = train_rmse
def test_invalid(self):
prop = bcpp.Int()
assert not prop.is_valid(0.0)
assert not prop.is_valid(1.0)
assert not prop.is_valid(1.0+1.0j)
assert not prop.is_valid("")
assert not prop.is_valid(())
assert not prop.is_valid([])
assert not prop.is_valid({})
assert not prop.is_valid(_TestHasProps())
assert not prop.is_valid(_TestModel())
assert not prop.is_valid(np.bool8(False))
assert not prop.is_valid(np.bool8(True))
assert not prop.is_valid(np.float16(0))
assert not prop.is_valid(np.float16(1))
assert not prop.is_valid(np.float32(0))
assert not prop.is_valid(np.float32(1))
assert not prop.is_valid(np.float64(0))
assert not prop.is_valid(np.float64(1))
assert not prop.is_valid(np.complex64(1.0+1.0j))
assert not prop.is_valid(np.complex128(1.0+1.0j))
if hasattr(np, "complex256"):
assert not prop.is_valid(np.complex256(1.0+1.0j))
def _costFunction(self,ep1,ep2,method):
# Endpoint Atribute (1)
x1 = numpy.array(ep1.Position,dtype=numpy.float16)
o1 = numpy.array(ep1.Orientation,dtype=numpy.float16)
t1 = numpy.float16(ep1.Thickness)
# Endpoint Atribute (2)
x2 = numpy.array(ep2.Position,dtype=numpy.float16)
o2 = numpy.array(ep2.Orientation,dtype=numpy.float16)
t2 = numpy.float16(ep2.Thickness)
# Pair Features
c = (x1+x2)/2 # centre
d = numpy.linalg.norm(x1-x2) # distance
k1= (x1-c) / numpy.linalg.norm(x1-c) # vector pointing to center (1)
k2= (x2-c) / numpy.linalg.norm(x2-c) # vector pointing to center (2)
# Meassure if orientation of endpoints is alligned [0: <- -> , 1: -> <-]
A=1-(2+numpy.dot(k1,o1)+numpy.dot(k2,o2))/4 #[0,1]
if d>=self.maxDist:return -1
if A<self.minOrie:return -1
if method == "dist":
return d
def test_nans_infs(self):
oldsettings = np.seterr(all='ignore')
try:
# Check some of the ufuncs
assert_equal(np.isnan(self.all_f16), np.isnan(self.all_f32))
assert_equal(np.isinf(self.all_f16), np.isinf(self.all_f32))
assert_equal(np.isfinite(self.all_f16), np.isfinite(self.all_f32))
assert_equal(np.signbit(self.all_f16), np.signbit(self.all_f32))
assert_equal(np.spacing(float16(65504)), np.inf)
# Check comparisons of all values with NaN
nan = float16(np.nan)
assert_(not (self.all_f16 == nan).any())
assert_(not (nan == self.all_f16).any())
assert_((self.all_f16 != nan).all())
assert_((nan != self.all_f16).all())
assert_(not (self.all_f16 < nan).any())
assert_(not (nan < self.all_f16).any())
assert_(not (self.all_f16 <= nan).any())
assert_(not (nan <= self.all_f16).any())
assert_(not (self.all_f16 > nan).any())
assert_(not (nan > self.all_f16).any())
assert_(not (self.all_f16 >= nan).any())
assert_(not (nan >= self.all_f16).any())
finally:
np.seterr(**oldsettings)
def dist_vec(training, test):
n1, d = training.shape
n2, d1 = test.shape
assert n1 != 0, 'Training set is empty'
assert n2 != 0, 'Test set is empty'
assert d==d1, 'Images in training and test sets have different size'
tstart = time.time()
train_squared = np.sum(np.square(training), axis = 1)
test_squared = np.sum(np.square(test), axis = 1)
A = np.tile(train_squared, (n2,1)) # n2xn1 matrix
A = A.transpose((1,0)) # n1xn2 matrix
B = np.tile(test_squared, (n1,1) ) # n2xn2 matrix
a = np.tile(training, (1,1,1)) # 1xn1x64 matrix
a = a.transpose((1,0,2)) # n1x1x64 matrix
b = np.tile(test, (1,1,1) ) # 1xn2x64 matrix
C = np.tensordot(a,b, [[1,2],[0,2]])
dist = A + B - C - C
dist = np.sqrt(dist)
np.float16(dist)
tstop = time.time()
return dist, tstop-tstart
def Scale_Image(image): #Scales the input image to the range:(0,255) min = np.amin(image) max = np.amax(image) image -= min image = np.float16(image) * (np.float16(255) / np.float16(max-min)) image = np.uint8(image) return image
def npHalfArrayToOIIOFloatPixels(width, height, channels, npPixels):
if oiio.VERSION < 10800:
# Read half-float pixels into a numpy float pixel array
oiioFloatsArray = np.frombuffer(np.getbuffer(np.float16(npPixels)), dtype=np.float32)
else:
# Read half-float pixels into a numpy float pixel array
oiioFloatsArray = np.frombuffer(np.getbuffer(np.float16(npPixels)), dtype=np.uint16)
return oiioFloatsArray
def __init__(self,timeSeries=None,
lenSeries=2**18,
numChannels=1,
fMin=400,fMax=800,
sampTime=None,
noiseRMS=0.1):
""" Initializes the AmplitudeTimeSeries instance.
If a array is not passed, then a random whitenoise dataset is generated.
Inputs:
Len -- Number of time data points (usually a power of 2) 2^38 gives about 65 seconds
of 400 MHz sampled data
The time binning is decided by the bandwidth
fMin -- lowest frequency (MHz)
fMax -- highest frequency (MHz)
noiseRMS -- RMS value of noise (TBD)
noiseAlpha -- spectral slope (default is white noise) (TBD)
ONLY GENERATES WHITE NOISE RIGHT NOW!
"""
self.shape = (np.uint(numChannels),np.uint(lenSeries))
self.fMax = fMax
self.fMin = fMin
if sampTime is None:
self.sampTime = np.uint(numChannels)*1E-6/(fMax-fMin)
else:
self.sampTime = sampTime
if timeSeries is None:
# then use the rest of the data to generate a random timeseries
if VERBOSE:
print "AmplitudeTimeSeries __init__ did not get new data, generating white noise data"
self.timeSeries = np.complex64(noiseRMS*(np.float16(random.standard_normal(self.shape))
+np.float16(random.standard_normal(self.shape))*1j)/np.sqrt(2))
else:
if VERBOSE:
print "AmplitudeTimeSeries __init__ got new data, making sure it is reasonable."
if len(timeSeries.shape) == 1:
self.shape = (1,timeSeries.shape[0])
else:
self.shape = timeSeries.shape
self.timeSeries = np.reshape(np.complex64(timeSeries),self.shape)
self.fMin = fMin
self.fMax = fMax
if sampTime is None:
self.sampTime = numChannels*1E-6/(fMax-fMin)
else:
self.sampTime = sampTime
return None
def compute_grad_img(image, height, width):
image_Sobel_x = Sobel(image, CV_64F, 1, 0, ksize=5)
image_Sobel_y = Sobel(image, CV_64F, 0, 1, ksize=5)
img_grad = zeros((height, width), dtype=tuple)
for row in range(height):
for col in range(width):
img_grad[row, col] = (float16(sqrt(image_Sobel_x[row, col][0] ** 2 + image_Sobel_y[row, col][0] ** 2)), \
float16(sqrt(image_Sobel_x[row, col][1] ** 2 + image_Sobel_y[row, col][1] ** 2)), \
float16(sqrt(image_Sobel_x[row, col][2] ** 2 + image_Sobel_y[row, col][2] ** 2)))
return img_grad
def testFloat(self):
num = np.float(256.2013)
self.assertEqual(np.float(ujson.decode(ujson.encode(num))), num)
num = np.float16(256.2013)
self.assertEqual(np.float16(ujson.decode(ujson.encode(num))), num)
num = np.float32(256.2013)
self.assertEqual(np.float32(ujson.decode(ujson.encode(num))), num)
num = np.float64(256.2013)
self.assertEqual(np.float64(ujson.decode(ujson.encode(num))), num)
def testFloatMax(self):
num = np.float(np.finfo(np.float).max / 10)
assert_approx_equal(np.float(ujson.decode(ujson.encode(num))), num, 15)
num = np.float16(np.finfo(np.float16).max / 10)
assert_approx_equal(np.float16(ujson.decode(ujson.encode(num))), num, 15)
num = np.float32(np.finfo(np.float32).max / 10)
assert_approx_equal(np.float32(ujson.decode(ujson.encode(num))), num, 15)
num = np.float64(np.finfo(np.float64).max / 10)
assert_approx_equal(np.float64(ujson.decode(ujson.encode(num))), num, 15)
def get_resource_data(group_serv, DFR_RAM, DFR_CPU):
D=pd.DataFrame()
for G in DFR_RAM.Group.unique():
if group_serv in G:
print(G)
df=pd.DataFrame()
fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(20, 5))
ax0, ax1= axes.flat
d=DFR_RAM[DFR_RAM.Group==G]
#print(d)
D=D.append(d)
X, Y=d.shape
ind = np.arange(X)
width = 0.5
#d=d[['server','Used', 'Available']]
p1 = ax0.bar(ind, d.Used.values, color='red', label='Used_memory_Gb')
p2 = ax0.bar(ind, d.Available.values, color='green', bottom=d.Used.values, label='Available')
ax0.set_xticks(range(d.shape[0]))
ax0.set_xticklabels([str(x) for x in d.server.values], rotation=90)
ax0.set_title(G+" Memory usage Gb")
ax0.get_legend()
ax0.set_xlabel("KB")
df=df.append(pd.DataFrame({"Group":G, "Number":d.shape[0],"Resource":"Memory","Sum_used":d.Used.sum(),"Sum_capacity":d.Capacity.sum() }, index=['RAM']))
d=DFR_CPU[DFR_CPU.Group==G]
D=D.append(d)
X, Y=d.shape
ind = np.arange(X)
width = 0.5
#d=d[['server','Used', 'Available']]
d['Used']=np.float16(d['Used'].str.replace(" %", ""))
p1 = ax1.bar(ind, d.Used.values, color='red', label='Used_CPU_%')
p2 = ax1.bar(ind, d.Available.values, color='green', bottom=d.Used.values, label='Available')
ax1.set_xticks(range(d.shape[0]))
ax1.set_xticklabels([str(x) for x in d.server.values], rotation=90)
ax1.set_title(G+" CPU usage %")
ax1.set_xlabel("%")
#fig.text(G)
fig.set_label(G)
fig.show
df=df.append(pd.DataFrame({"Group":G, "Number":d.shape[0], "Resource":"CPU","Sum_used":d.Used.sum(),"Sum_capacity":100*d.Capacity.sum() }, index=['CPU']), ignore_index=True)
display(df)
print("Summary CPU ")
print("Used "+str(d.Used.sum()/d.Capacity.sum())+" %")
print("Summary RAM ")
print("Used "+str(100*np.float16(df['Sum_used'][df.Resource=='Memory'].values/df.Sum_capacity[df.Resource=='Memory'].values)[0])+" %")
return D
def pack(data, ind=None, dims=None, sorting=False, axes=None):
"""Pack an RDD into a dense local array, with options for
sorting, reshaping, and projecting based on keys
Parameters
----------
data : RDD of (tuple, array) pairs
The data to pack into a local array
ind : int, optional, default = None
An index, if each record has multiple entries
dims : Dimensions, optional, default = None
Dimensions of the keys, for use with sorting and reshaping
sorting : Boolean, optional, default = False
Whether to sort the RDD before packing
axes : int, optional, default = None
Which axis to do maximum projection along
Returns
-------
result : array
A local numpy array with the RDD contents
"""
if dims is None:
dims = getdims(data)
if axes is not None:
nkeys = len(data.first()[0])
data = data.map(lambda (k, v): (tuple(array(k)[arange(0, nkeys) != axes]), v)).reduceByKey(maximum)
dims.min = list(array(dims.min)[arange(0, nkeys) != axes])
dims.max = list(array(dims.max)[arange(0, nkeys) != axes])
sorting = True # will always need to sort because reduceByKey changes order
if ind is None:
result = data.map(lambda (_, v): float16(v)).collect()
nout = size(result[0])
else:
result = data.map(lambda (_, v): float16(v[ind])).collect()
nout = size(ind)
if sorting is True:
data = subtoind(data, dims.max)
keys = data.map(lambda (k, _): int(k)).collect()
result = array([v for (k, v) in sorted(zip(keys, result), key=lambda (k, v): k)])
return squeeze(transpose(reshape(result, ((nout,) + dims.count())[::-1])))
def Load_images(file_dir, file_name):
if (os.path.exists(file_dir + file_name + '/' + file_name + '_images.npy')==False):
img = Image.open(file_dir + file_name + '/' + file_name+'.tif')
counter=0
while True:
try:
img.seek(counter)
except EOFError:
break
counter+=1
#Default pic sizes
n_pixels = 256
#Initialize 3D image array
n_frames = counter
images_raw = np.zeros((n_frames, n_pixels, n_pixels), dtype = np.float16)
print "n_frames: ", n_frames
for i in range(n_frames):
try:
img.seek(i)
print "Loading frame: ", i
images_raw [i] = np.float16(img)
if i>9640:# and i%10==0: #
im = plt.imshow(images_raw[i])
plt.title("Frame: " +str(i))
plt.show()
except EOFError:
break
images_start= 0
images_end = 9642
images_raw=images_raw[images_start:images_end]
print "Re-saving imaging array..."
np.save(file_dir + file_name + '/' + file_name + '_images', images_raw)
np.savetxt(file_dir + file_name + '/' + file_name + '_images_start_'+str(images_start)+'_end_'+
str(images_end), [images_start, images_end])
quit()
else:
images_raw = np.load(file_dir + file_name + '/' + file_name + '_images.npy')
images_raw = np.float16(images_raw)
return images_raw
def test_is_float():
# is float
assert isinstance(1.0, float) is True
assert isinstance(np.float(1.0), float) is True
assert isinstance(np.float16(1.0), float) is False
assert isinstance(np.float32(1.0), float) is False
assert isinstance(np.float64(1.0), float) is True
# is np.float
assert isinstance(1.0, np.float) is True
assert isinstance(np.float(1.0), np.float) is True
assert isinstance(np.float16(1.0), np.float) is False
assert isinstance(np.float32(1.0), np.float) is False
assert isinstance(np.float64(1.0), np.float) is True
def test_is_float(self):
self.assertIsInstance(1.0, float) # Yes
self.assertIsInstance(np.float(1.0), float) # Yes
self.assertIsInstance(np.float64(1.0), float) # Yes
self.assertNotIsInstance(np.float16(1.0), float) # No
self.assertNotIsInstance(np.float32(1.0), float) # No
self.assertIsInstance(1.0, np.float) # Yes
self.assertIsInstance(np.float64(1.0), np.float) # Yes
self.assertNotIsInstance(np.float16(1.0), np.float) # No
self.assertNotIsInstance(np.float32(1.0), np.float) # No
def convert_R11G11B10_FLOAT(buf): # Extract the channels with shifts and masks, add the implicit 1 bit to # the mantissas and unbias the exponents: rm = np.int8 (((buf >> 0) & 0x3f) | 0x40) re = np.float16(((buf >> 6) & 0x1f)) - 15 gm = np.int8 (((buf >> 11) & 0x3f) | 0x40) ge = np.float16(((buf >> 17) & 0x1f)) - 15 bm = np.int8 (((buf >> 22) & 0x1f) | 0x20) be = np.float16(((buf >> 27) & 0x1f)) - 15 # Calculate floating point values and scale: r = scale_float(rm * (2**re)) g = scale_float(gm * (2**ge)) b = scale_float(bm * (2**be)) return np.uint8(np.column_stack((r, g, b)))
def test_Complex(self):
prop = Complex()
self.assertTrue(prop.is_valid(None))
# TODO: self.assertFalse(prop.is_valid(False))
# TODO: self.assertFalse(prop.is_valid(True))
self.assertTrue(prop.is_valid(0))
self.assertTrue(prop.is_valid(1))
self.assertTrue(prop.is_valid(0.0))
self.assertTrue(prop.is_valid(1.0))
self.assertTrue(prop.is_valid(1.0 + 1.0j))
self.assertFalse(prop.is_valid(""))
self.assertFalse(prop.is_valid(()))
self.assertFalse(prop.is_valid([]))
self.assertFalse(prop.is_valid({}))
self.assertFalse(prop.is_valid(Foo()))
try:
import numpy as np
# TODO: self.assertFalse(prop.is_valid(np.bool8(False)))
# TODO: self.assertFalse(prop.is_valid(np.bool8(True)))
self.assertTrue(prop.is_valid(np.int8(0)))
self.assertTrue(prop.is_valid(np.int8(1)))
self.assertTrue(prop.is_valid(np.int16(0)))
self.assertTrue(prop.is_valid(np.int16(1)))
self.assertTrue(prop.is_valid(np.int32(0)))
self.assertTrue(prop.is_valid(np.int32(1)))
self.assertTrue(prop.is_valid(np.int64(0)))
self.assertTrue(prop.is_valid(np.int64(1)))
self.assertTrue(prop.is_valid(np.uint8(0)))
self.assertTrue(prop.is_valid(np.uint8(1)))
self.assertTrue(prop.is_valid(np.uint16(0)))
self.assertTrue(prop.is_valid(np.uint16(1)))
self.assertTrue(prop.is_valid(np.uint32(0)))
self.assertTrue(prop.is_valid(np.uint32(1)))
self.assertTrue(prop.is_valid(np.uint64(0)))
self.assertTrue(prop.is_valid(np.uint64(1)))
self.assertTrue(prop.is_valid(np.float16(0)))
self.assertTrue(prop.is_valid(np.float16(1)))
self.assertTrue(prop.is_valid(np.float32(0)))
self.assertTrue(prop.is_valid(np.float32(1)))
self.assertTrue(prop.is_valid(np.float64(0)))
self.assertTrue(prop.is_valid(np.float64(1)))
self.assertTrue(prop.is_valid(np.complex64(1.0 + 1.0j)))
self.assertTrue(prop.is_valid(np.complex128(1.0 + 1.0j)))
self.assertTrue(prop.is_valid(np.complex256(1.0 + 1.0j)))
except ImportError:
pass
def _remap(self, dependency_dims, derived_dims):
if derived_dims:
ndim = max(derived_dims) + 1
else:
ndim = 1
nd_points_by_key = {}
for key, coord in self.dependencies.iteritems():
if coord:
# Get the points as consistent with the Cube.
nd_points = self._nd_points(coord, dependency_dims[key], ndim)
# Restrict to just the dimensions relevant to the derived coord.
# NB. These are always in Cube-order, so no transpose is needed.
shape = []
for dim in derived_dims:
shape.append(nd_points.shape[dim])
# Ensure the array always has at least one dimension to be
# compatible with normal coordinates.
if not derived_dims:
shape.append(1)
nd_points.shape = shape
else:
# If no coord, treat value as zero.
# Use a float16 to provide `shape` attribute and avoid
# promoting other arguments to a higher precision.
nd_points = np.float16(0)
nd_points_by_key[key] = nd_points
return nd_points_by_key
def test_convert_np_float(self):
u = Unit('mile')
v = Unit('meter')
self.assertEqual(u.convert(np.float(1.0), v), 1609.344)
self.assertEqual(u.convert(np.float16(1.0), v), 1609.344)
self.assertEqual(u.convert(np.float32(1.0), v), 1609.344)
self.assertEqual(u.convert(np.float64(1.0), v), 1609.344)
def predict(mdl, img, patch_size, patch_step, batch_size, dim_img):
"""
the cnn model for image transformation
Parameters
----------
img : array
The image need to be calculated
patch_size : (int, int)
The patches dimension
dim_img : int
The input image dimension
Returns
-------
img_rec
Description.
"""
img = np.float16(utils.nor_data(img))
img_y, img_x = img.shape
x_img = utils.extract_patches(img, patch_size, patch_step)
x_img = np.reshape(x_img, (len(x_img), 1, dim_img, dim_img))
y_img = mdl.predict(x_img, batch_size=batch_size)
del x_img
y_img = np.reshape(y_img, (len(y_img), dim_img, dim_img))
img_rec = utils.reconstruct_patches(y_img, (img_y, img_x), patch_step)
return img_rec
def test_half_coercion(self):
"""Test that half gets coerced properly with the other types"""
a16 = np.array((1,),dtype=float16)
a32 = np.array((1,),dtype=float32)
b16 = float16(1)
b32 = float32(1)
assert_equal(np.power(a16,2).dtype, float16)
assert_equal(np.power(a16,2.0).dtype, float16)
assert_equal(np.power(a16,b16).dtype, float16)
assert_equal(np.power(a16,b32).dtype, float16)
assert_equal(np.power(a16,a16).dtype, float16)
assert_equal(np.power(a16,a32).dtype, float32)
assert_equal(np.power(b16,2).dtype, float64)
assert_equal(np.power(b16,2.0).dtype, float64)
assert_equal(np.power(b16,b16).dtype, float16)
assert_equal(np.power(b16,b32).dtype, float32)
assert_equal(np.power(b16,a16).dtype, float16)
assert_equal(np.power(b16,a32).dtype, float32)
assert_equal(np.power(a32,a16).dtype, float32)
assert_equal(np.power(a32,b16).dtype, float32)
assert_equal(np.power(b32,a16).dtype, float16)
assert_equal(np.power(b32,b16).dtype, float32)
def test_numpy(self):
"""NumPy objects get serialized to readable JSON."""
l = [
np.float32(12.5),
np.float64(2.0),
np.float16(0.5),
np.bool(True),
np.bool(False),
np.bool_(True),
np.unicode_("hello"),
np.byte(12),
np.short(12),
np.intc(-13),
np.int_(0),
np.longlong(100),
np.intp(7),
np.ubyte(12),
np.ushort(12),
np.uintc(13),
np.ulonglong(100),
np.uintp(7),
np.int8(1),
np.int16(3),
np.int32(4),
np.int64(5),
np.uint8(1),
np.uint16(3),
np.uint32(4),
np.uint64(5),
]
l2 = [l, np.array([1, 2, 3])]
roundtripped = loads(dumps(l2, cls=EliotJSONEncoder))
self.assertEqual([l, [1, 2, 3]], roundtripped)
def __init__(self):
self._lock=threading.RLock()
self._running=False
#load calibration parameters
with open('calibration/Sawyer.yaml') as file:
H_Sawyer = np.array(yaml.load(file)['H'],dtype=np.float64)
with open('calibration/UR.yaml') as file:
H_UR = np.array(yaml.load(file)['H'],dtype=np.float64)
with open('calibration/ABB.yaml') as file:
H_ABB = np.array(yaml.load(file)['H'],dtype=np.float64)
with open('calibration/tx60.yaml') as file:
H_tx60 = np.array(yaml.load(file)['H'],dtype=np.float64)
self.H_UR=H42H3(H_UR)
self.H_Sawyer=H42H3(H_Sawyer)
self.H_ABB=H42H3(H_ABB)
self.H_tx60=H42H3(H_tx60)
self.H_robot={'ur':self.H_UR,'sawyer':self.H_Sawyer,'abb':self.H_ABB,'staubli':self.H_tx60}
self.distance_report=RRN.GetStructureType("edu.rpi.robotics.distance.distance_report")
self.dict={'ur':0,'sawyer':1,'abb':2,'staubli':3}
self.distance_report_dict={}
for robot_name,robot_idx in self.dict.items():
self.distance_report_dict[robot_name]=self.distance_report()
#connect to RR gazebo plugin service
server=RRN.ConnectService('rr+tcp://localhost:11346/?service=GazeboServer')
self.w=server.get_worlds(str(server.world_names[0]))
#create RR pose type
pose_dtype = RRN.GetNamedArrayDType("com.robotraconteur.geometry.Pose", server)
self.model_pose = np.zeros((1,), dtype = pose_dtype)
#form H into RR transformation struct
self.transformations={}
self.transformation=RRN.GetStructureType("edu.rpi.robotics.distance.transformation")
transformation1=self.transformation()
transformation1.name="UR"
transformation1.row=len(self.H_UR)
transformation1.column=len(self.H_UR[0])
transformation1.H=np.float16(self.H_UR).flatten().tolist()
self.transformations['ur']=transformation1
transformation2=self.transformation()
transformation2.name="Sawyer"
transformation2.row=len(self.H_Sawyer)
transformation2.column=len(self.H_Sawyer[0])
transformation2.H=np.float16(self.H_Sawyer).flatten().tolist()
self.transformations['sawyer']=transformation2
transformation3=self.transformation()
transformation3.name="ABB"
transformation3.row=len(self.H_ABB)
transformation3.column=len(self.H_ABB[0])
transformation3.H=np.float16(self.H_ABB).flatten().tolist()
self.transformations['abb']=transformation3
transformation4=self.transformation()
transformation4.name="Staubli"
transformation4.row=len(self.H_tx60)
transformation4.column=len(self.H_tx60[0])
transformation4.H=np.float16(self.H_tx60).flatten().tolist()
self.transformations['staubli']=transformation4
#Connect to robot service
with open('client_yaml/client_ur.yaml') as file:
self.url_ur= yaml.load(file)['url']
with open('client_yaml/client_sawyer.yaml') as file:
self.url_sawyer= yaml.load(file)['url']
with open('client_yaml/client_abb.yaml') as file:
self.url_abb= yaml.load(file)['url']
with open('client_yaml/client_staubli.yaml') as file:
self.url_tx60= yaml.load(file)['url']
self.ur_sub=RRN.SubscribeService(self.url_ur)
self.sawyer_sub=RRN.SubscribeService(self.url_sawyer)
self.abb_sub=RRN.SubscribeService(self.url_abb)
self.tx60_sub=RRN.SubscribeService(self.url_tx60)
UR_state=self.ur_sub.SubscribeWire("robot_state")
Sawyer_state=self.sawyer_sub.SubscribeWire("robot_state")
ABB_state=self.abb_sub.SubscribeWire("robot_state")
tx60_state=self.tx60_sub.SubscribeWire("robot_state")
#link and joint names in urdf
Sawyer_joint_names=["right_j0","right_j1","right_j2","right_j3","right_j4","right_j5","right_j6"]
UR_joint_names=["shoulder_pan_joint","shoulder_lift_joint","elbow_joint","wrist_1_joint","wrist_2_joint","wrist_3_joint"]
Sawyer_link_names=["right_l0","right_l1","right_l2","right_l3","right_l4","right_l5","right_l6","right_l1_2","right_l2_2","right_l4_2","right_hand"]
UR_link_names=['UR_base_link',"shoulder_link","upper_arm_link","forearm_link","wrist_1_link","wrist_2_link","wrist_3_link"]
ABB_joint_names=['ABB1200_joint_1','ABB1200_joint_2','ABB1200_joint_3','ABB1200_joint_4','ABB1200_joint_5','ABB1200_joint_6']
ABB_link_names=['ABB1200_base_link','ABB1200_link_1','ABB1200_link_2','ABB1200_link_3','ABB1200_link_4','ABB1200_link_5','ABB1200_link_6']
tx60_joint_names=['tx60_joint_1','tx60_joint_2','tx60_joint_3','tx60_joint_4','tx60_joint_5','tx60_joint_6']
tx60_link_names=['tx60_base_link','tx60_link_1','tx60_link_2','tx60_link_3','tx60_link_4','tx60_link_5','tx60_link_6']
self.robot_state_list=[UR_state,Sawyer_state,ABB_state,tx60_state]
self.robot_link_list=[UR_link_names,Sawyer_link_names,ABB_link_names,tx60_link_names]
self.robot_joint_list=[UR_joint_names,Sawyer_joint_names,ABB_joint_names,tx60_joint_names]
self.num_robot=len(self.robot_state_list)
######tesseract environment setup:
self.t_env = Environment()
urdf_path = FilesystemPath("urdf/combined.urdf")
srdf_path = FilesystemPath("urdf/combined.srdf")
assert self.t_env.init(urdf_path, srdf_path, GazeboModelResourceLocator())
#update robot poses based on calibration file
self.t_env.changeJointOrigin("ur_pose", Isometry3d(H_UR))
self.t_env.changeJointOrigin("sawyer_pose", Isometry3d(H_Sawyer))
self.t_env.changeJointOrigin("abb_pose", Isometry3d(H_ABB))
self.t_env.changeJointOrigin("staubli_pose", Isometry3d(H_tx60))
contact_distance=0.2
monitored_link_names = self.t_env.getLinkNames()
self.manager = self.t_env.getDiscreteContactManager()
self.manager.setActiveCollisionObjects(monitored_link_names)
self.manager.setContactDistanceThreshold(contact_distance)
# viewer update
self.viewer = TesseractViewer()
self.viewer.update_environment(self.t_env, [0,0,0])
self.viewer.start_serve_background()
self.robot_def_dict={}
try:
UR=self.ur_sub.GetDefaultClientWait(1)
self.robot_def_dict['ur']=Robot(np.transpose(np.array(UR.robot_info.chains[0].H.tolist())),np.transpose(np.array(UR.robot_info.chains[0].P.tolist())),np.zeros(len(UR.robot_info.joint_info)))
except:
pass
try:
Sawyer=self.sawyer_sub.GetDefaultClientWait(1)
self.robot_def_dict['sawyer']=Robot(np.transpose(np.array(Sawyer.robot_info.chains[0].H.tolist())),np.transpose(np.array(Sawyer.robot_info.chains[0].P.tolist())),np.zeros(len(Sawyer.robot_info.joint_info)))
except:
pass
try:
ABB=self.abb_sub.GetDefaultClientWait(1)
self.robot_def_dict['abb']=Robot(np.transpose(np.array(ABB.robot_info.chains[0].H.tolist())),np.transpose(np.array(ABB.robot_info.chains[0].P.tolist())),np.zeros(len(ABB.robot_info.joint_info)))
except:
pass
try:
Staubli=self.tx60_sub.GetDefaultClientWait(1)
self.robot_def_dict['staubli']=Robot(np.transpose(np.array(Staubli.robot_info.chains[0].H.tolist())),np.transpose(np.array(Staubli.robot_info.chains[0].P.tolist())),np.zeros(len(Staubli.robot_info.joint_info)))
except:
pass
#trajectories
self.traj_change=False
self.traj_change_name=None
self.steps=300
self.plan_time=0.15
self.execution_delay=0.03
self.trajectory={'ur':np.zeros((self.steps,7)),'sawyer':np.zeros((self.steps,8)),'abb':np.zeros((self.steps,7)),'staubli':np.zeros((self.steps,7))}
self.traj_joint_names={'ur':['shoulder_pan_joint', 'shoulder_lift_joint', 'elbow_joint', 'wrist_1_joint', 'wrist_2_joint', 'wrist_3_joint'],
'sawyer':['right_j0', 'right_j1', 'right_j2', 'right_j3', 'right_j4', 'right_j5', 'right_j6'],
'abb':['joint_1', 'joint_2', 'joint_3', 'joint_4', 'joint_5', 'joint_6'],
'staubli':['joint_1', 'joint_2', 'joint_3', 'joint_4', 'joint_5', 'joint_6']
}
self.time_step=0.03
#initialize static trajectories
for key, value in self.trajectory.items():
for i in range(self.steps):
try:
value[i]=np.append([0],self.robot_state_list[self.dict[key]].InValue.joint_position)
except:
traceback.print_exc()
value[i]=np.append([0],[0,0,0,0,0,0])
self.inv={'ur':inv_ur,'sawyer':inv_sawyer,'abb':inv_abb,'staubli':inv_staubli}
self.joint_names_traj={'ur':inv_ur,'sawyer':inv_sawyer,'abb':inv_abb,'staubli':inv_staubli}
#register service constant
self.JointTrajectoryWaypoint = RRN.GetStructureType("com.robotraconteur.robotics.trajectory.JointTrajectoryWaypoint")
self.JointTrajectory = RRN.GetStructureType("com.robotraconteur.robotics.trajectory.JointTrajectory")
def nearest_by_svlen_overlap(df_source, df_target, max_dist, size_sim, restrict_samples=False):
"""
For each variant in df_source, get the nearest variant in df_target. Both dataframes must contain fields
"#CHROM", "POS", "END", and "SVLEN" (which is the absolute value, no negative SVLEN for DELs). Return a dataframe
with each source variant ("ID") with the best match by distance and then by size overlap ("TARGET_ID") along with
distance ("DISTANCE"), reciprocal overlap if variants intersect ("RO", 0 if no overlap), and size proportion
("TARGET_SIZE_PROP", target-size / source-size).
:param df_source: Source dataframe.
:param df_target: Target dataframe.
:param size_sim: Length proportion of matches.
:param restrict_samples: If `True` and both dataframes contain a `MERGE_SAMPLES` column, then restrict matches to
only those that share samples.
:return: A dataframe with "ID", "TARGET_ID", "DISTANCE", "RO", and "TARGET_SIZE_PROP".
"""
# Check for expected columns
if any(col not in df_source.columns for col in ('#CHROM', 'POS', 'END', 'SVLEN', 'ID')):
raise RuntimeError('Source Dataframe missing at least one column in ("#CHROM", "POS", "END", "SVLEN"): {}')
if any(col not in df_target.columns for col in ('#CHROM', 'POS', 'END', 'SVLEN', 'ID')):
raise RuntimeError('Target Dataframe missing at least one column in ("#CHROM", "POS", "END", "SVLEN"): {}')
# IDs must be unique
if len(set(df_source['ID'])) != df_source.shape[0]:
raise RuntimeError('Source Dataframe IDs are not unique')
if len(set(df_target['ID'])) != df_target.shape[0]:
raise RuntimeError('target Dataframe IDs are not unique')
# Determine if variants are matched on MERGE_SAMPLES
if restrict_samples and 'MERGE_SAMPLES' in df_source.columns and 'MERGE_SAMPLES' in df_target.columns:
restrict_samples = True
else:
restrict_samples = False
# Subset and cast to int16 (default int64 uses more memory and is not needed)
if restrict_samples:
subset_cols = ('#CHROM', 'POS', 'END', 'SVLEN', 'ID', 'MERGE_SAMPLES')
col_map = {
'#CHROM': np.object, 'POS': np.int32, 'END': np.int32, 'SVLEN': np.int32, 'ID': np.object,
'MERGE_SAMPLES': np.object
}
else:
subset_cols = ('#CHROM', 'POS', 'END', 'SVLEN', 'ID')
col_map = {
'#CHROM': np.object, 'POS': np.int32, 'END': np.int32, 'SVLEN': np.int32, 'ID': np.object
}
stats_box = OrderedDict()
stats_box["TP-base"] = 0
stats_box["TP-call"] = 0
stats_box["FP"] = 0
stats_box["FN"] = 0
stats_box["precision"] = 0
stats_box["recall"] = 0
stats_box["f1"] = 0
stats_box["base cnt"] = 0
stats_box["call cnt"] = 0
df_source = df_source.loc[:, subset_cols]
df_target = df_target.loc[:, subset_cols]
# df_source.set_index('ID', inplace=True)
copy_df_targat = df_target.set_index('ID', inplace=False)
matched_calls = defaultdict(bool)
if size_sim <= np.float16(0.0) or size_sim >= np.float16(1.0):
raise RuntimeError('Length proportion must be between 0 and 1 (exclusive): {}'.format(size_sim))
# Nearest by #CHROM
overlap_list = list() # Dataframe for each chrom (to be merged into one dataframe)
for chrom in sorted(set(df_source['#CHROM'])):
df_source_chr = df_source.loc[df_source['#CHROM'] == chrom]
df_target_chr = copy_df_targat.loc[copy_df_targat['#CHROM'] == str(chrom)]
# Target has no values on this chrom, all calls in source file are missed
if df_target_chr.shape[0] == 0:
stats_box["FN"] += df_source_chr.shape[0]
continue
for index, source_row in df_source_chr.iterrows():
stats_box["base cnt"] += 1
pos = source_row['POS']
end = source_row['END']
source_id = source_row['ID']
# print(df_source_chr.columns)
min_len = np.int32(source_row['SVLEN'] * size_sim)
max_len = np.int32(source_row['SVLEN'] * (2 - size_sim))
# Subset target - Skip if no targets within range
df_target_chr_len = df_target_chr.loc[
(df_target_chr['SVLEN'] >= min_len) & (df_target_chr['SVLEN'] <= max_len)
]
# Cannot find targes of similar size to the source
if df_target_chr_len.shape[0] == 0:
stats_box["FN"] += 1
continue
# Get distance to all target records
distance = df_target_chr_len.apply(
lambda row: row['POS'] - end if row['POS'] > end else (
pos - row['END'] if pos > row['END'] else 0
),
axis=1
)
# Assign index and subset minimum
min_distance = np.min(distance)
# Only consider potential match within max distance
if min_distance > max_dist:
stats_box["FN"] += 1
continue
distance = distance.loc[distance == min_distance]
# Multiple match continue or find the best match
if len(distance) > 1:
match_id = np.argmin(np.abs(df_target_chr_len.loc[distance.index, 'SVLEN'] - source_row['SVLEN']))
continue
else:
# Only one best match, get ID
match_id = distance.index[0]
# Save base ID to matched calls
if not matched_calls[source_row["ID"]]:
stats_box["TP-base"] += 1
matched_calls[source_row["ID"]] = True
if not matched_calls[match_id]:
stats_box["TP-call"] += 1
matched_calls[match_id] = True
# Save match record
overlap_list.append((index, source_id, match_id, min_distance, len(distance)))
do_stats_math = True
if stats_box["TP-base"] == 0 and stats_box["FN"] == 0:
logging.warning("No TP or FN calls in base!")
do_stats_math = False
elif stats_box["TP-call"] == 0 and stats_box["FP"] == 0:
logging.warning("No TP or FP calls in base!")
do_stats_math = False
else:
logging.info("Results peek: %d TP-base %d FN %.2f%% Recall", stats_box["TP-base"], stats_box["FN"],
100 * (float(stats_box["TP-base"]) / (stats_box["TP-base"] + stats_box["FN"])))
for index, entry in df_target.iterrows():
# print(entry)
if not matched_calls[entry["ID"]]:
stats_box['FP'] += 1
if do_stats_math:
stats_box["precision"] = float(stats_box["TP-call"]) / (stats_box["TP-call"] + stats_box["FP"])
stats_box["recall"] = float(stats_box["TP-base"]) / (stats_box["TP-base"] + stats_box["FN"])
# f-measure
neum = stats_box["recall"] * stats_box["precision"]
denom = stats_box["recall"] + stats_box["precision"]
if denom != 0:
stats_box["f1"] = 2 * (neum / denom)
else:
stats_box["f1"] = "NaN"
stats_box["call cnt"] = stats_box["TP-base"] + stats_box["FP"]
# Merge records
df = pd.DataFrame(overlap_list, columns=('ID', 'BASE_ID', 'TARGET_ID', 'DISTANCE','TARGET_MATCHES'))
# Annotate overlap and size proportion (RO and TARGET_SIZE_PROP).
df['#CHROM'] = list(df_source.loc[df['ID'], '#CHROM'])
df['POS'] = list(df_source.loc[df['ID'], 'POS'])
df['END'] = list(df_source.loc[df['ID'], 'END'])
df['SVLEN'] = list(df_source.loc[df['ID'], 'SVLEN'])
df['TARGET_POS'] = list(copy_df_targat.loc[df['TARGET_ID'], 'POS'])
df['TARGET_END'] = list(copy_df_targat.loc[df['TARGET_ID'], 'END'])
df['TARGET_SVLEN'] = list(copy_df_targat.loc[df['TARGET_ID'], 'SVLEN'])
df['RO'] = df.apply(lambda row: reciprocal_overlap(*row.loc[['POS', 'END', 'TARGET_POS', 'TARGET_END']]), axis=1)
# # Set TARGET_SIZE_PROP
# df['TARGET_SIZE_PROP'] = round(df['TARGET_SVLEN'] / df['SVLEN'], 3)
df['TARGET_SIZE_SIM'] = df.apply(
lambda row: size_similarity(*row.loc[['TARGET_SVLEN', 'SVLEN']]), axis=1
)
# df['SIM_SCORE'] = df.apply(
# lambda row: sim_score(*row.loc[['RO', 'TARGET_SIZE_SIM', 'DISTANCE', 'SVLEN', 'TARGET_SVLEN']]), axis=1
# )
# Sort
df.sort_values(['#CHROM', 'POS'])
df = df.loc[:, ('TARGET_ID', 'BASE_ID', 'TARGET_MATCHES', 'RO','TARGET_SIZE_SIM')]
# Return dataframe
return df, stats_box
def testFloatTensor(self):
self.assertEqual(dtypes.float64, _create_tensor(np.float64()).dtype) # pylint: disable=no-value-for-parameter
self.assertEqual(dtypes.float32, _create_tensor(np.float32()).dtype) # pylint: disable=no-value-for-parameter
self.assertEqual(dtypes.float16, _create_tensor(np.float16()).dtype) # pylint: disable=no-value-for-parameter
self.assertEqual(dtypes.float32, _create_tensor(0.0).dtype)
def test_stationarity(ts):
dftest = adfuller(ts, autolag='AIC')
return np.float16(dftest[1])
PATCH_WIDTH = 100
PATCH_HEIGHT = 100
PATCH_SIZE = PATCH_WIDTH * PATCH_HEIGHT * 3
batch_size = 5
dslr_ = tf.placeholder(tf.float32, [None, PATCH_SIZE])
dslr_image = tf.reshape(dslr_, [-1, PATCH_HEIGHT, PATCH_WIDTH, 3])
adv_ = tf.placeholder(tf.float32, [None, 1])
dslr_gray = tf.reshape(tf.image.rgb_to_grayscale(dslr_image),
[-1, PATCH_WIDTH * PATCH_HEIGHT])
#adversarial_ = tf.multiply(enhanced_gray, 1 - adv_) + tf.multiply(dslr_gray, adv_)
adversarial_ = tf.multiply(dslr_gray, adv_)
adversarial_image = tf.reshape(adversarial_,
[-1, PATCH_HEIGHT, PATCH_WIDTH, 1])
discrim_target = tf.concat([adv_, 1 - adv_], 1)
with tf.Session() as sess:
image = np.float16(
misc.imread("test_image\\patches\\iphone\\iphone\\1.jpg")) / 255
image_ = np.reshape(image, [1, PATCH_SIZE])
all_zeros = np.reshape(np.zeros((batch_size, 1)), [batch_size, 1])
print("discrim_target")
print(sess.run(discrim_target, feed_dict={dslr_: image_, adv_: all_zeros}))
print(discrim_target)
print("dslr_gray")
print(sess.run(dslr_gray, feed_dict={dslr_: image_, adv_: all_zeros}))
print(dslr_gray)
print('image_')
print(image_.shape)
print('all_zeros')
print(all_zeros.shape)
def get_machine_half_product(self):
return np.float16(ProductSumTest.__MACHINE_NUMBER * self.__times)
def halfToUInt16(halfValue):
return np.frombuffer(np.getbuffer(np.float16(halfValue)), dtype=np.uint16)[0]
import numpy as np
from artiq.protocols import pyon
_pyon_test_object = {
(1, 2): [(3, 4.2), (2, )],
Fraction(3, 4): np.linspace(5, 10, 1),
"a": np.int8(9),
"b": np.int16(-98),
"c": np.int32(42),
"d": np.int64(-5),
"e": np.uint8(8),
"f": np.uint16(5),
"g": np.uint32(4),
"h": np.uint64(9),
"x": np.float16(9.0),
"y": np.float32(9.0),
"z": np.float64(9.0),
}
class PYON(unittest.TestCase):
def test_encdec(self):
for enc in pyon.encode, lambda x: pyon.encode(x, True):
self.assertEqual(pyon.decode(enc(_pyon_test_object)),
_pyon_test_object)
_json_test_object = {
"a": "b",
"x": [1, 2, {}],
def test_experiment_handles_numpy_numbers(self):
nums_to_test = [
("int_", np.int_()),
("intc", np.intc()),
("intp", np.intp()),
("int8", np.int8()),
("int16", np.int16()),
("int32", np.int32()),
("int64", np.int64()),
("uint8", np.uint8()),
("uint16", np.uint16()),
("uint32", np.uint32()),
("uint64", np.uint64()),
("float16", np.float16()),
("float32", np.float32()),
("float64", np.float64()),
]
# Make sure the SDK doesn't choke and JSON serialization works
exp = Experiment("MNIST")
for name, num in nums_to_test:
exp.metric("test_metric_{}".format(name), num)
exp.param("test_param_{}".format(name), num)
exp.end()
# Test params match what is expected
params_messages = []
for msg in server_sdk_messages:
payload = msg["payload"]
if "params" in payload:
params_messages.append(payload)
expected_params = []
for name, num in nums_to_test:
obj = {
"params": {},
"is_internal": False,
}
obj["params"]["test_param_{}".format(name)] = num
obj["is_internal"] = False
expected_params.append(obj)
assert len(expected_params) == len(params_messages)
for i, message in enumerate(params_messages):
print(message)
print(expected_params[i])
assert message == expected_params[i]
# Test metrics match what is expected
metrics_messages = []
for msg in server_sdk_messages:
payload = msg["payload"]
if "name" in payload:
metrics_messages.append(payload)
expected_metrics = []
for name, num in nums_to_test:
expected_metrics.append({
"name": "test_metric_{}".format(name),
"value": num,
"is_internal": False,
})
assert len(expected_metrics) == len(metrics_messages)
for i, message in enumerate(metrics_messages):
assert message == expected_metrics[i]
def __init__(self):
self._lock = threading.RLock()
self._running = False
#load calibration parameters
with open('calibration/Sawyer.yaml') as file:
H_Sawyer = np.array(yaml.load(file)['H'], dtype=np.float64)
with open('calibration/ABB.yaml') as file:
H_ABB = np.array(yaml.load(file)['H'], dtype=np.float64)
self.H_Sawyer = H42H3(H_Sawyer)
self.H_ABB = H42H3(H_ABB)
self.L2C = ['', '']
#form H into RR transformation struct
transformation = RRN.GetStructureType(
"edu.rpi.robotics.distance.transformation")
transformation1 = transformation()
transformation1.name = "Sawyer"
transformation1.row = len(self.H_Sawyer)
transformation1.column = len(self.H_Sawyer[0])
transformation1.H = np.float16(self.H_Sawyer).flatten().tolist()
transformation2 = transformation()
transformation2.name = "ABB"
transformation2.row = len(self.H_ABB)
transformation2.column = len(self.H_ABB[0])
transformation2.H = np.float16(self.H_ABB).flatten().tolist()
self.transformations = [transformation1, transformation2]
#Connect to robot service
Sawyer = RRN.ConnectService('rr+tcp://localhost:58654?service=robot')
ABB = RRN.ConnectService('rr+tcp://localhost:58655?service=robot')
Sawyer_state = Sawyer.robot_state.Connect()
ABB_state = ABB.robot_state.Connect()
#link and joint names in urdf
Sawyer_joint_names = [
"right_j0", "right_j1", "right_j2", "right_j3", "right_j4",
"right_j5", "right_j6"
]
Sawyer_link_names = [
"right_l0", "right_l1", "right_l2", "right_l3", "right_l4",
"right_l5", "right_l6", "right_l1_2", "right_l2_2", "right_l4_2",
"right_hand"
]
ABB_joint_names = [
'ABB1200_joint_1', 'ABB1200_joint_2', 'ABB1200_joint_3',
'ABB1200_joint_4', 'ABB1200_joint_5', 'ABB1200_joint_6'
]
ABB_link_names = [
'ABB1200_base_link', 'ABB1200_link_1', 'ABB1200_link_2',
'ABB1200_link_3', 'ABB1200_link_4', 'ABB1200_link_5',
'ABB1200_link_6'
]
self.robot_state_list = [Sawyer_state, ABB_state]
self.robot_link_list = [Sawyer_link_names, ABB_link_names]
self.robot_joint_list = [Sawyer_joint_names, ABB_joint_names]
self.num_robot = len(self.robot_state_list)
self.distance_matrix = -np.ones(self.num_robot * self.num_robot)
######tesseract environment setup:
with open("urdf/combined.urdf", 'r') as f:
combined_urdf = f.read()
with open("urdf/combined.srdf", 'r') as f:
combined_srdf = f.read()
t = tesseract.Tesseract()
t.init(combined_urdf, combined_srdf, GazeboModelResourceLocator())
self.t_env = t.getEnvironment()
#update robot poses based on calibration file
self.t_env.changeJointOrigin("sawyer_pose", H_Sawyer)
self.t_env.changeJointOrigin("abb_pose", H_ABB)
contact_distance = 0.1
monitored_link_names = self.t_env.getLinkNames()
self.manager = self.t_env.getDiscreteContactManager()
self.manager.setActiveCollisionObjects(monitored_link_names)
self.manager.setContactDistanceThreshold(contact_distance)
def distance_check(self, robot_idx):
with self._lock:
distance_report = RRN.GetStructureType(
"edu.rpi.robotics.distance.distance_report")
distance_report1 = distance_report()
#update other robot's joints
for i in range(self.num_robot):
robot_joints = self.robot_state_list[i].InValue.joint_position
self.t_env.setState(self.robot_joint_list[i], robot_joints)
env_state = self.t_env.getCurrentState()
self.manager.setCollisionObjectsTransform(
env_state.link_transforms)
contacts = self.manager.contactTest(2)
contact_vector = tesseract.flattenResults(contacts)
distances = np.array([c.distance for c in contact_vector])
nearest_points = np.array(
[c.nearest_points for c in contact_vector])
names = np.array([c.link_names for c in contact_vector])
# nearest_index=np.argmin(distances)
min_distance = 9
min_index = -1
Closest_Pt = [0., 0., 0.]
Closest_Pt_env = [0., 0., 0.]
#initialize
distance_report1.Closest_Pt = Closest_Pt
distance_report1.Closest_Pt_env = Closest_Pt_env
for i in range(len(distances)):
#only 1 in 2 collision "objects"
if (names[i][0] in self.robot_link_list[robot_idx]
or names[i][1] in self.robot_link_list[robot_idx]
) and distances[i] < min_distance and not (
names[i][0] in self.robot_link_list[robot_idx]
and names[i][1] in self.robot_link_list[robot_idx]):
min_distance = distances[i]
min_index = i
J2C = 0
if (min_index != -1):
if names[min_index][0] in self.robot_link_list[
robot_idx] and names[min_index][
1] in self.robot_link_list[robot_idx]:
stop = 1
print("stop")
elif names[min_index][0] in self.robot_link_list[robot_idx]:
J2C = self.robot_link_list[robot_idx].index(
names[min_index][0])
Closest_Pt = nearest_points[min_index][0]
Closest_Pt_env = nearest_points[min_index][1]
print(names[min_index])
print(distances[min_index])
elif names[min_index][1] in self.robot_link_list[robot_idx]:
J2C = self.robot_link_list[robot_idx].index(
names[min_index][1])
Closest_Pt = nearest_points[min_index][1]
Closest_Pt_env = nearest_points[min_index][0]
print(names[min_index])
print(distances[min_index])
if robot_idx == 0:
J2C = self.Sawyer_link(J2C)
distance_report1.Closest_Pt = np.float16(
Closest_Pt).flatten().tolist()
distance_report1.Closest_Pt_env = np.float16(
Closest_Pt_env).flatten().tolist()
distance_report1.min_distance = np.float16(
distances[min_index])
distance_report1.J2C = J2C
return distance_report1
return distance_report1
def poplulate_file_queue(self):
location = self.data_source
image_list = location + "data.txt"
dct = defaultdict(list)
tmp_reference_image = []
tmp_distorted_image = []
tmp_scores = []
with open(location + "data.txt") as f:
file_contents = f.read()
lines = file_contents.split('\n')
lines = [item for item in lines if item != '']
old_content = '-1'
for i in range(len(lines)): # for the full dataset
if i == 0:
continue
line_data = lines[i].split('\t')
line_data
dfilename = location + str(line_data[3]) + '\\' + str(
line_data[0]) + '.' + str(line_data[3]) + '.' + str(
line_data[4]) + '.png'
rfilename = location + str(line_data[0]) + '.png'
score = np.float16(line_data[6])
content = int(line_data[1])
ntype = int(line_data[2])
nlevel = line_data[4]
dct[rfilename].append(i - 1)
tmp_reference_image.append(rfilename)
tmp_distorted_image.append(dfilename)
tmp_scores.append(score)
# print(dfilename)
# Seperate images based on reference image content
self.dist = tmp_distorted_image
self.ref = tmp_reference_image
self.mos = tmp_scores
index_of_contents = []
c_length = len(dct.keys())
for count, key in enumerate(dct.keys()):
index_of_contents.append(dct[key])
[train_content, test_content] = train_test_split(c_length, ratio=0.99)
file_id_list_train = []
for indexs in train_content:
file_id_list_train += index_of_contents[indexs]
file_id_list_test = []
for indexs in test_content:
file_id_list_test += index_of_contents[indexs]
# Global storate of file list
self.file_reference_image_train = []
self.file_distorted_image_train = []
self.MOS_scores_train = []
self.file_reference_image_test = []
self.file_distorted_image_test = []
self.MOS_scores_test = []
for file_idxs in file_id_list_train:
self.file_distorted_image_train.append(
tmp_distorted_image[file_idxs])
self.file_reference_image_train.append(
tmp_reference_image[file_idxs])
self.MOS_scores_train.append(tmp_scores[file_idxs])
for file_idxs in file_id_list_test:
self.file_distorted_image_test.append(
tmp_distorted_image[file_idxs])
self.file_reference_image_test.append(
tmp_reference_image[file_idxs])
self.MOS_scores_test.append(tmp_scores[file_idxs])
self.refill_file_queues()
def run( soltab, tecsoltabOut='tec000', clocksoltabOut='clock000', offsetsoltabOut='phase_offset000', tec3rdsoltabOut='tec3rd000', flagBadChannels=True, flagCut=5., chi2cut=3000., combinePol=False, removePhaseWraps=True, fit3rdorder=False, circular=False, reverse=False, invertOffset=False, nproc=10 ):
"""
Separate phase solutions into Clock and TEC.
The Clock and TEC values are stored in the specified output soltab with type 'clock', 'tec', 'tec3rd'.
Parameters
----------
flagBadChannels : bool, optional
Detect and remove bad channel before fitting, by default True.
flagCut : float, optional
chi2cut : float, optional
combinePol : bool, optional
Find a combined polarization solution, by default False.
removePhaseWraps : bool, optional
Detect and remove phase wraps, by default True.
fit3rdorder : bool, optional
Fit a 3rd order ionospheric ocmponent (usefult <40 MHz). By default False.
circular : bool, optional
Assume circular polarization with FR not removed. By default False.
reverse : bool, optional
Reverse the time axis. By default False.
invertOffset : bool, optional
Invert (reverse the sign of) the phase offsets. By default False. Set to True
if you want to use them with the residuals operation.
"""
import numpy as np
from ._fitClockTEC import doFit
logging.info("Clock/TEC separation on soltab: "+soltab.name)
# some checks
solType = soltab.getType()
if solType != 'phase':
logging.warning("Soltab type of "+soltab.name+" is: "+solType+" should be phase. Ignoring.")
return 1
# Collect station properties
solset = soltab.getSolset()
station_dict = solset.getAnt()
stations = soltab.getAxisValues('ant')
station_positions = np.zeros((len(stations), 3), dtype=np.float)
for i, station_name in enumerate(stations):
station_positions[i, 0] = station_dict[station_name.encode()][0]
station_positions[i, 1] = station_dict[station_name.encode()][1]
station_positions[i, 2] = station_dict[station_name.encode()][2]
returnAxes=['ant','freq','pol','time']
for vals, flags, coord, selection in soltab.getValuesIter(returnAxes=returnAxes,weight=True):
if len(coord['ant']) < 2:
logging.error('Clock/TEC separation needs at least 2 antennas selected.')
return 1
if len(coord['freq']) < 10:
logging.error('Clock/TEC separation needs at least 10 frequency channels, preferably distributed over a wide range')
return 1
freqs=coord['freq']
stations=coord['ant']
times=coord['time']
# get axes index
axes=[i for i in soltab.getAxesNames() if i in returnAxes]
# reverse time axes
if reverse:
vals = np.swapaxes(np.swapaxes(vals, 0, axes.index('time'))[::-1], 0, axes.index('time'))
flags = np.swapaxes(np.swapaxes(flags, 0, axes.index('time'))[::-1], 0, axes.index('time'))
result=doFit(vals,flags==0,freqs,stations,station_positions,axes,\
flagBadChannels=flagBadChannels,flagcut=flagCut,chi2cut=chi2cut,combine_pol=combinePol,removePhaseWraps=removePhaseWraps,fit3rdorder=fit3rdorder,circular=circular,n_proc=nproc)
if fit3rdorder:
clock,tec,offset,tec3rd=result
if reverse:
clock = clock[::-1,:]
tec = tec[::-1,:]
tec3rd = tec3rd[::-1,:]
else:
clock,tec,offset=result
if reverse:
clock = clock[::-1,:]
tec = tec[::-1,:]
if invertOffset:
offset *= -1.0
weights=tec>-5
tec[np.logical_not(weights)]=0
clock[np.logical_not(weights)]=0
weights=np.float16(weights)
if combinePol or not 'pol' in soltab.getAxesNames():
tf_st = solset.makeSoltab('tec', soltabName = tecsoltabOut,
axesNames=['time', 'ant'], axesVals=[times, stations],
vals=tec[:,:,0],
weights=weights[:,:,0])
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
tf_st = solset.makeSoltab('clock', soltabName = clocksoltabOut,
axesNames=['time', 'ant'], axesVals=[times, stations],
vals=clock[:,:,0]*1e-9,
weights=weights[:,:,0])
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
tf_st = solset.makeSoltab('phase', soltabName = offsetsoltabOut,
axesNames=['ant'], axesVals=[stations],
vals=offset[:,0],
weights=np.ones_like(offset[:,0],dtype=np.float16))
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
if fit3rdorder:
tf_st = solset.makeSoltab('tec3rd', soltabName = tec3rdsoltabOut,
axesNames=['time', 'ant'], axesVals=[times, stations],
vals=tec3rd[:,:,0],
weights=weights[:,:,0])
else:
tf_st = solset.makeSoltab('tec', soltabName = tecsoltabOut,
axesNames=['time', 'ant','pol'], axesVals=[times, stations, ['XX','YY']],
vals=tec,
weights=weights)
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
tf_st = solset.makeSoltab('clock', soltabName = clocksoltabOut,
axesNames=['time', 'ant','pol'], axesVals=[times, stations, ['XX','YY']],
vals=clock*1e-9,
weights=weights)
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
tf_st = solset.makeSoltab('phase', soltabName = offsetsoltabOut,
axesNames=['ant','pol'], axesVals=[stations, ['XX','YY']],
vals=offset,
weights=np.ones_like(offset,dtype=np.float16))
tf_st.addHistory('CREATE (by CLOCKTECFIT operation)')
if fit3rdorder:
tf_st = solset.makeSoltab('tec3rd', soltabName = tec3rdsoltabOut,
axesNames=['time', 'ant','pol'], axesVals=[times, stations, ['XX','YY']],
vals=tec3rd,
weights=weights)
return 0
def detect_blobs(frame, thresh, meas_last, meas_now, last_heads, colors):
# Function to find valid fly contours and return centroid coordinates
# Returns new meas_now tuple with error handling to return last valid one if
# no good measurements found
# Generate contours with no discrimination
_, contours, _ = cv2.findContours(thresh.copy(), cv2.RETR_TREE,
cv2.CHAIN_APPROX_SIMPLE)
# Discriminate contours based on area percentage of total frame. Flies are generally between 0.001 and 0.01
good_contours = []
list_of_detections = []
for i, contour in enumerate(contours):
if ((cv2.contourArea(contour) / (thresh.shape[0]**2)) < 0.01) and (
(cv2.contourArea(contour) / (thresh.shape[0]**2)) > 0.0001):
good_contours.append(contour)
# Second round of discrimination based on whether or not centroid is filled and if it is in masked "non-arena" area
better_contours = []
for i, contour in enumerate(good_contours):
M = cv2.moments(contour)
if M['m00'] != 0:
cx = np.float16(M['m10'] / M['m00'])
cy = np.float16(M['m01'] / M['m00'])
center_coords = np.asarray(
[int(frame.shape[0] / 2),
int(frame.shape[1] / 2)])
dist_to_center = np.linalg.norm(center_coords -
np.asarray([cx, cy]))
if (dist_to_center > (frame.shape[0] / 2)):
pass
elif [int(cx), int(cy)] in [[0, 0], [0, frame.shape[0]],
[frame.shape[0], 0], []]:
pass
else:
# Check whether contour is mostly filled or not
area = cv2.contourArea(contour)
edge = int(np.sqrt(area))
h = int(edge / 2)
perim = cv2.arcLength(contour, True)
window = thresh[(int(cy) - h):(int(cy) + h),
(int(cx) - h):(int(cx) + h)]
if np.mean(window) != np.nan:
if (np.mean(window) > 60) and (np.mean(window) < 160):
better_contours.append(contour)
detection = Detection()
detection.add_centroid([cx, cy])
try:
ellipse = cv2.fitEllipse(contour)
MAG = ellipse[1][1] * 0.5
mag = ellipse[1][0] * 0.5
ellipseAngle = math.radians(ellipse[-1])
ellipseAngle = ellipseAngle - (math.pi / 2)
ellipseAngle2 = ellipseAngle - (math.pi)
x_ang = math.degrees(math.cos(ellipseAngle))
y_ang = math.degrees(math.sin(ellipseAngle))
fpx1 = int(cx - MAG * math.radians(x_ang))
fpy1 = int(cy - MAG * math.radians(y_ang))
x_ang2 = math.degrees(math.cos(ellipseAngle2))
y_ang2 = math.degrees(math.sin(ellipseAngle2))
fpx2 = int(cx - MAG * math.radians(x_ang2))
fpy2 = int(cy - MAG * math.radians(y_ang2))
# Need to determine which between [fpx1, fpy1], [fpx2, fpy2]
# is animal head, append along with cx, cy
# Do disoriented patch coords
disoriented_patch = cv2.boundingRect(contour)
# Make disoriented patch a bit bigger so full bigger animal can fit
scale_factor = 3
w = int(disoriented_patch[2] * scale_factor)
h = int(disoriented_patch[3] * scale_factor)
dim = np.max([w, h])
x = int((disoriented_patch[0]) -
(dim - disoriented_patch[2]) / 2)
y = int((disoriented_patch[1]) -
(dim - disoriented_patch[3]) / 2)
# Grab pixels from these dims from color img:
patch = frame[y:(y + h), x:(x + h)]
# Generate minimum bounding rectangle around the larger contour
head_point = get_head_point(
patch, [x, y], [cx, cy],
[[fpx1, fpy1], [fpx2, fpy2]])
detection.add_head(head_point)
detection.add_area(area)
except TypeError:
detection.add_head(None)
else:
pass
else:
pass
try:
if (detection.head is None) and (detection.centroid is None):
pass
else:
list_of_detections.append(detection)
except UnboundLocalError:
pass
return list(set(list_of_detections)), len(better_contours)
def get_real_half_product(self):
return np.float16(
np.float16(ProductSumTest.__REAL_NUMBER) * self.__times)
def test_half_fpe(self):
with np.errstate(all="raise"):
sx16 = np.array((1e-4, ), dtype=float16)
bx16 = np.array((1e4, ), dtype=float16)
sy16 = float16(1e-4)
by16 = float16(1e4)
# Underflow errors
assert_raises_fpe("underflow", lambda a, b: a * b, sx16, sx16)
assert_raises_fpe("underflow", lambda a, b: a * b, sx16, sy16)
assert_raises_fpe("underflow", lambda a, b: a * b, sy16, sx16)
assert_raises_fpe("underflow", lambda a, b: a * b, sy16, sy16)
assert_raises_fpe("underflow", lambda a, b: a / b, sx16, bx16)
assert_raises_fpe("underflow", lambda a, b: a / b, sx16, by16)
assert_raises_fpe("underflow", lambda a, b: a / b, sy16, bx16)
assert_raises_fpe("underflow", lambda a, b: a / b, sy16, by16)
assert_raises_fpe("underflow", lambda a, b: a / b,
float16(2.0**-14), float16(2**11))
assert_raises_fpe(
"underflow",
lambda a, b: a / b,
float16(-(2.0**-14)),
float16(2**11),
)
assert_raises_fpe(
"underflow",
lambda a, b: a / b,
float16(2.0**-14 + 2**-24),
float16(2),
)
assert_raises_fpe(
"underflow",
lambda a, b: a / b,
float16(-(2.0**-14) - 2**-24),
float16(2),
)
assert_raises_fpe(
"underflow",
lambda a, b: a / b,
float16(2.0**-14 + 2**-23),
float16(4),
)
# Overflow errors
assert_raises_fpe("overflow", lambda a, b: a * b, bx16, bx16)
assert_raises_fpe("overflow", lambda a, b: a * b, bx16, by16)
assert_raises_fpe("overflow", lambda a, b: a * b, by16, bx16)
assert_raises_fpe("overflow", lambda a, b: a * b, by16, by16)
assert_raises_fpe("overflow", lambda a, b: a / b, bx16, sx16)
assert_raises_fpe("overflow", lambda a, b: a / b, bx16, sy16)
assert_raises_fpe("overflow", lambda a, b: a / b, by16, sx16)
assert_raises_fpe("overflow", lambda a, b: a / b, by16, sy16)
assert_raises_fpe("overflow", lambda a, b: a + b, float16(65504),
float16(17))
assert_raises_fpe("overflow", lambda a, b: a - b, float16(-65504),
float16(17))
assert_raises_fpe("overflow", np.nextafter, float16(65504),
float16(np.inf))
assert_raises_fpe("overflow", np.nextafter, float16(-65504),
float16(-np.inf))
assert_raises_fpe("overflow", np.spacing, float16(65504))
# Invalid value errors
assert_raises_fpe("invalid", np.divide, float16(np.inf),
float16(np.inf))
assert_raises_fpe("invalid", np.spacing, float16(np.inf))
assert_raises_fpe("invalid", np.spacing, float16(np.nan))
# These should not raise
float16(65472) + float16(32)
float16(2**-13) / float16(2)
float16(2**-14) / float16(2**10)
np.spacing(float16(-65504))
np.nextafter(float16(65504), float16(-np.inf))
np.nextafter(float16(-65504), float16(np.inf))
np.nextafter(float16(np.inf), float16(0))
np.nextafter(float16(-np.inf), float16(0))
np.nextafter(float16(0), float16(np.nan))
np.nextafter(float16(np.nan), float16(0))
float16(2**-14) / float16(2**10)
float16(-(2**-14)) / float16(2**10)
float16(2**-14 + 2**-23) / float16(2)
float16(-(2**-14) - 2**-23) / float16(2)
#use this for LUT-MANT generation for float16 with first 6 bits used to address LUT
import numpy as np
import math
a = np.float64(1)
print("module LUT2(addr, log);")
print(" input [5:0] addr;")
print(" output reg [15:0] log;")
print("")
print(" always @(addr) begin")
print(" case (addr)")
for i in range (64):
temp = np.log(a)
num = bin(np.float16(temp).view('H'))[2:].zfill(16)
print(" 6'b{0:b}".format(i).zfill(5)),
print(" : log = 16'b{0};".format(num))
a += np.exp2(-6)
print(" endcase")
print(" end")
print("endmodule")
def normalize(utterance):
utterance = utterance - np.mean(utterance, axis=0, dtype=np.float64)
return np.float16(utterance)
def SignNumpy(x): return np.float16(2.*np.greater_equal(x,0)-1.) # 将值变为+1或-1
def test_encode_numpy_float(self):
self.assertEqual(
json.dumps(np.float16(3.76953125),
cls=utils_json.AirflowJsonEncoder), '3.76953125')
test_photos = test_photos[0:5]
if phone.endswith("_orig"):
# load pre-trained model
saver = tf.train.Saver()
saver.restore(sess, "models_orig/" + phone)
for photo in test_photos:
# load training image and crop it if necessary
print("Testing original " + phone.replace("_orig", "") +
" model, processing image " + photo)
image = np.float16(
misc.imresize(misc.imread(test_dir + photo),
res_sizes[phone])) / 255
image_crop = utils.extract_crop(image, resolution, phone,
res_sizes)
image_crop_2d = np.reshape(image_crop, [1, IMAGE_SIZE])
# get enhanced image
enhanced_2d = sess.run(enhanced, feed_dict={x_: image_crop_2d})
enhanced_image = np.reshape(enhanced_2d,
[IMAGE_HEIGHT, IMAGE_WIDTH, 3])
enhanced_image = cv2.resize(enhanced_image, (1280, 720))
# before_after = np.hstack((image_crop, enhanced_image))
photo_name = photo.rsplit(".", 1)[0]
import numpy as np
from bitstring import BitArray
def differentPrecision(num, bits):
binum = BitArray(float=num, length=bits)
sign = int(binum[0])
if bits == 32:
exponent = binum[1:9].bin
mantissa = binum[9:].bin
else:
exponent = binum[1:12].bin
mantissa = binum[12:].bin
return sign, exponent, mantissa
def displayPrecision(num):
binum = BitArray(float=num, length=64)
sign = int(binum[0])
exponent = binum[1:12].bin
mantissa = binum[12:].bin
return sign, exponent, mantissa
print("float16:", displayPrecision(np.float16(1 / 3)))
print("float32:", displayPrecision(np.float32(1 / 3)))
print("float64:", displayPrecision(np.float64(1 / 3)))
print('\n')
print("float16 to float32:", displayPrecision(np.float32(np.float16(1 / 3))))
print("float16 to float64:", displayPrecision(np.float64(np.float16(1 / 3))))
print("float32 to float64:", displayPrecision(np.float64(np.float32(1 / 3))))
np.complex128(3j)
np.complex128(C())
np.complex128(None)
np.complex64("1.2")
np.complex128(b"2j")
np.int8(4)
np.int16(3.4)
np.int32(4)
np.int64(-1)
np.uint8(B())
np.uint32()
np.int32("1")
np.int64(b"2")
np.float16(A())
np.float32(16)
np.float64(3.0)
np.float64(None)
np.float32("1")
np.float16(b"2.5")
if sys.version_info >= (3, 8):
np.uint64(D())
np.float32(D())
np.complex64(D())
np.bytes_(b"hello")
np.bytes_("hello", 'utf-8')
np.bytes_("hello", encoding='utf-8')
np.str_("hello")
"hello": "world",
1: 42,
2.5: 45
},
{
"hello": set([2, 3]),
"world": set([42.0]),
"this": None
},
np.int8(3),
np.int32(4),
np.int64(5),
np.uint8(3),
np.uint32(4),
np.uint64(5),
np.float16(1.9),
np.float32(1.9),
np.float64(1.9),
np.zeros([8, 20]),
np.random.normal(size=[17, 10]),
np.array(["hi", 3]),
np.array(["hi", 3], dtype=object),
np.random.normal(size=[15, 13]).T,
]
if sys.version_info >= (3, 0):
PRIMITIVE_OBJECTS += [0, np.array([["hi", u"hi"], [1.3, 1]])]
else:
PRIMITIVE_OBJECTS += [
long(42),
long(1 << 62),
def datasetvormgeven(lijstsequenties):
"""Functie maakt de benodigde lijsten om een zelfde vorm te krijgen
als de moving mnist dataset die als voorbeeld gebruikt wordt.
Afbeeldingen wordne in gelezen in grayscale en gecheckt dat de
breedte == 788 en de hoogte == 525. Als dit niet zo is worden deze
afbeeldingen uit de dataset verwijdert."""
inp = []
gt = []
imdata = []
index = 0
x = 0
for element in lijstsequenties:
tijdelijk = []
errorinseq = False
for model in element:
img = cv2.imread(BASEPATH + '/' + model, cv2.IMREAD_GRAYSCALE)
h, w = img.shape
if 525 != h or 788 != w:
# print("Probleem H of W bij " + model)
errorinseq = True
continue
img = cv2.resize(img, (240, 164), interpolation = cv2.INTER_AREA)
img = np.float32(img)
img = cv2.normalize(img, img, 0, 1, norm_type=cv2.NORM_MINMAX)
img = np.float16(img)
tijdelijk.append([img])
if errorinseq == True:
# print("Sequentie geskipt!")
continue
if args.add_dummy_data != 0:
for i in range(args.add_dummy_data):
dummydata = np.zeros((164, 240), dtype=np.float16)
tijdelijk.append([dummydata])
for t in tijdelijk:
imdata.append(t)
totl = len(tijdelijk)
inpl = SEQGROOTTE - PREDGROOTTE
gtl = PREDGROOTTE + args.add_dummy_data
tijdelijk = []
tijdelijk.append(index)
tijdelijk.append(inpl)
inp.append(tijdelijk)
index += inpl
tijdelijk = []
tijdelijk.append(index)
tijdelijk.append(gtl)
gt.append(tijdelijk)
index += gtl
x += 1
print("Aantal sequenties = " + str(x))
return inp, gt, imdata
def handle_representation():
table = pd.read_csv(process_path, index_col=0, encoding='utf-8')
print('***** data ')
for i in range(table.shape[0]):
try:
if isinstance(table['price'][i], str):
if '-' in table['price'][i]:
a, b = table['price'][i].split('-')
a = np.float16(
a.replace('-',
'').replace(' ',
'').replace('£',
'').replace(',', ''))
b = np.float16(
b.replace('-',
'').replace(' ',
'').replace('£',
'').replace(',', ''))
table['price'][i] = a + b / 2.0
else:
table['price'][i] = np.float16(table['price'][i].replace(
'£', '').replace(' ', '').replace(',', ''))
if isinstance(table['average_review_rating'][i], str):
table['average_review_rating'][i] = np.float16(
table['average_review_rating'][i].replace(
' out of 5 stars', '').replace(' ', ''))
# print('%d finish'%i)
if isinstance(
table['customers_who_bought_this_item_also_bought'][i],
str):
products = re.findall(
regex,
table['customers_who_bought_this_item_also_bought'][i])
p = ''
for k, product in enumerate(products):
product_s = product.split('/')[0]
if k > 0:
p += '|' + product_s
else:
p += product_s
table['customers_who_bought_this_item_also_bought'][i] = p
# print('%d finish'%i)
if isinstance(
table['items_customers_buy_after_viewing_this_item'][i],
str):
products = re.findall(
regex,
table['items_customers_buy_after_viewing_this_item'][i])
p = ''
for k, product in enumerate(products):
product_s = product.split('/')[0]
if k > 0:
p += '|' + product_s
else:
p += product_s
# print(i,' : ',p)
table['items_customers_buy_after_viewing_this_item'][i] = p
# print('%d finish' % i)
if isinstance(table['customer_reviews'][i], str):
s = table['customer_reviews'][i].replace('\n', '')
s = re.sub(r"\s{2,}", " ", s)
a = s.split('//')
p = ''
for k in range(0, len(a), 4):
if k > len(a) - 1 or k + 1 > len(a) - 1:
break
if k > 0:
p += '|%s-%s' % (a[k].replace(
'|', ' '), a[k + 1].replace(' ', ''))
else:
p += '%s-%s' % (a[k].replace(
'|', ' '), a[k + 1].replace(' ', ''))
table['customer_reviews'][i] = p
# print('%d finish' % i)
if isinstance(table['customer_questions_and_answers'][i], str):
s = table['customer_questions_and_answers'][i].replace(
'\n', ' ')
s = re.sub(r"\s{2,}", " ", s).replace('://', '')
p = ''
a = s.split('|')
for k, q_a in enumerate(a):
q, a = q_a.split('//')
if 'see more' in a:
r = re.findall('see more\s*(.*)\s*see less', s)[0]
else:
r = a
if k > 0:
p += '|%s//%s' % (q, r)
else:
p += '%s//%s' % (q, r)
table['customer_questions_and_answers'][i] = p
# print('%d finish' % i)
if isinstance(table['number_of_reviews'][i], str):
table['number_of_reviews'][i] = np.int16(
table['number_of_reviews'][i].replace(',', ''))
# print('%d finish'%i)
print('%d finish' % i)
pass
except Exception as e:
print('str(e):\t\t', str(e))
# print(table['price'][i],' - ',table['average_review_rating'][i])
traceback.print_exc()
break
table.to_csv('D:\HKBU_AI_Classs\IT_Project\process.csv')
def runtest(self):
model = RegressionModel.load(os.path.join(self.modelfile, "linear"), "linear")
betas, stats, resid = model.fit(self.rdd)
result = stats.map(lambda (_, v): float16(v)).collect()
savemat(self.savefile + "tmp.mat", mdict={"tmp": result}, oned_as='column')
def runtest(self):
result = self.rdd.map(lambda (_, v): float16(v[0])).collect()
savemat(self.savefile + "tmp.mat", mdict={"tmp": result}, oned_as='column')
def normalize(arr: np.ndarray) -> np.ndarray:
'''нормализация к1 канального избражения от 0 до 255'''
arr = np.float16(arr)
amin = arr.min()
rng = arr.max() - amin
return ((arr - amin) * 255 / rng).astype(np.uint8)
def distance_check_global(self,robot_name, joints_list):
with self._lock:
robot_idx=self.dict[robot_name]
distance_report=RRN.GetStructureType("edu.rpi.robotics.distance.distance_report")
distance_report1=distance_report()
for i in range(self.num_robot):
robot_joints=joints_list[i]
self.t_env.setState(self.robot_joint_list[i], robot_joints)
env_state = self.t_env.getCurrentState()
self.manager.setCollisionObjectsTransform(env_state.link_transforms)
result = ContactResultMap()
self.manager.contactTest(result, ContactRequest(ContactTestType_ALL))
result_vector = ContactResultVector()
collisionFlattenResults(result,result_vector)
distances = [r.distance for r in result_vector]
nearest_points=[[r.nearest_points[0],r.nearest_points[1]] for r in result_vector]
names = [[r.link_names[0],r.link_names[1]] for r in result_vector]
# nearest_index=np.argmin(distances)
min_distance=9
min_index=-1
Closest_Pt=[0.,0.,0.]
Closest_Pt_env=[0.,0.,0.]
#initialize
distance_report1.Closest_Pt=Closest_Pt
distance_report1.Closest_Pt_env=Closest_Pt_env
for i in range(len(distances)):
#only 1 in 2 collision "objects"
if (names[i][0] in self.robot_link_list[robot_idx] or names[i][1] in self.robot_link_list[robot_idx]) and distances[i]<min_distance and not (names[i][0] in self.robot_link_list[robot_idx] and names[i][1] in self.robot_link_list[robot_idx]):
min_distance=distances[i]
min_index=i
J2C=0
if (min_index!=-1):
if names[min_index][0] in self.robot_link_list[robot_idx] and names[min_index][1] in self.robot_link_list[robot_idx]:
stop=1
print("stop")
elif names[min_index][0] in self.robot_link_list[robot_idx]:
J2C=self.robot_link_list[robot_idx].index(names[min_index][0])-1
Closest_Pt=nearest_points[min_index][0]
Closest_Pt_env=nearest_points[min_index][1]
elif names[min_index][1] in self.robot_link_list[robot_idx]:
J2C=self.robot_link_list[robot_idx].index(names[min_index][1])-1
Closest_Pt=nearest_points[min_index][1]
Closest_Pt_env=nearest_points[min_index][0]
if robot_idx==1:
J2C=self.Sawyer_link(J2C)
distance_report1.Closest_Pt=np.float16(Closest_Pt).flatten().tolist()
distance_report1.Closest_Pt_env=np.float16(Closest_Pt_env).flatten().tolist()
distance_report1.min_distance=np.float16(distances[min_index])
distance_report1.J2C=(J2C if J2C>=0 else 0)
return distance_report1
return distance_report1