def extract(self):
try:
median = float(med(self.flux_data))
except:
self.ex_error(
"EXCEPT in medianextractor() most likely flux_data=[]")
return (median)
def rescale_vertices(self, scale, rescaling_type=RescalingTypeMin):
'''
Rescales the vertex coordinates so that the minimum dimension (X, Y, Z) is exactly min_scale
Params:
scale: (float) scale of the ,esj
rescaling_type: (int) which dimension to scale along; if not absolute then the min,med,max dim is scaled to be exactly scale
Returns:
Nothing. Modified the mesh in place (for now)
'''
vertex_array = np.array(self.mesh_.vertices())
min_vertex_coords = np.min(self.mesh_.vertices(), axis=0)
max_vertex_coords = np.max(self.mesh_.vertices(), axis=0)
vertex_extent = max_vertex_coords - min_vertex_coords
# find minimal dimension
if rescaling_type == MeshCleaner.RescalingTypeMin:
dim = np.where(vertex_extent == np.min(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshCleaner.RescalingTypeMed:
dim = np.where(vertex_extent == np.med(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshCleaner.RescalingTypeMax:
dim = np.where(vertex_extent == np.max(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshCleaner.RescalingTypeAbsolute:
relative_scale = 1.0
# compute scale factor and rescale vertices
scale_factor = scale / relative_scale
vertex_array = scale_factor * vertex_array
self.mesh_.vertices_ = vertex_array.tolist()
def __init__(self, data=[0]):
self.data = data
self.highest = max(data)
self.lowest = min(data)
self.median = med(data)
self.average = mean(data)
for i in self.__dict__:
call = self.__dict__[i]
try:
self.__setattr__(i, round(float(call), 1))
except TypeError:
self.__setattr__(i, list(map(lambda x: round(x,1), call)))
def _rescale_vertices(self, scale, rescaling_type=RescalingTypeMin):
"""
Rescales the vertex coordinates so that the minimum dimension (X, Y, Z) is exactly min_scale
Params:
scale: (float) scale of the mesh
rescaling_type: (int) which dimension to scale along; if not absolute then the min,med,max dim is scaled to be exactly scale
Returns:
Nothing. Modified the mesh in place (for now)
"""
vertex_array = np.array(self.mesh_.vertices())
min_vertex_coords = np.min(self.mesh_.vertices(), axis=0)
max_vertex_coords = np.max(self.mesh_.vertices(), axis=0)
vertex_extent = max_vertex_coords - min_vertex_coords
# find minimal dimension
if rescaling_type == MeshProcessor.RescalingTypeMin:
dim = np.where(vertex_extent == np.min(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshProcessor.RescalingTypeMed:
dim = np.where(vertex_extent == np.med(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshProcessor.RescalingTypeMax:
dim = np.where(vertex_extent == np.max(vertex_extent))[0][0]
relative_scale = vertex_extent[dim]
elif rescaling_type == MeshProcessor.RescalingTypeRelative:
relative_scale = 1.0
elif rescaling_type == MeshProcessor.RescalingTypeDiag:
diag = np.linalg.norm(vertex_extent)
relative_scale = diag / 3.0 # make the gripper size exactly one third of the diagonal
# compute scale factor and rescale vertices
scale_factor = scale / relative_scale
vertex_array = scale_factor * vertex_array
self.mesh_.vertices_ = vertex_array.tolist()
self.mesh_._compute_bb_center()
self.mesh_._compute_centroid()
self.mesh_.set_center_of_mass(self.mesh_.bb_center_)
def medfilt(ipt):
med_len = 31
output = []
for i in range(0, len(ipt)):
#get the values you want the median of
tmp = []
for x in range(i - med_len / 2, i + med_len / 2):
if x < 0: #out of bounds
if ipt[0] < 0:
tmp.append(.5)
else:
tmp.append(ipt[0])
elif x >= len(ipt):
if ipt[len(ipt) - 1] < 0:
tmp.append(.5)
else:
tmp.append(ipt[len(ipt) - 1])
else:
if ipt[x] < 0:
tmp.append(.5)
else:
tmp.append(ipt[x])
output.append(med(tmp))
return output
def extract(self):
try:
median = float(med(self.flux_data))
except:
self.ex_error("EXCEPT in medianextractor() most likely flux_data=[]")
return(median)
def experiment(request, experiment_id):
exp = Experiment.objects.get(id=experiment_id)
tasks = Task.objects.filter(Task_user=exp)
count_g = 0 # посдсчет решенных
count_b = 0 # подсчет не решенных
count_nb = 0 # подсчет счастливых но не ближайших
full = 0
times = []
time_of_task = []
time_of_task_avg = []
diff_of_time = [0]
if exp.Strategy in (1,2,3):
i = 1
for t in tasks:
count = 0
tm = 0.0
for v in Variant.objects.filter(Variant_task = t):
full += 1
count += 1
times.append(v.Time)
tm += v.Time
if v.Check == 'Решено':
count_g += 1
elif v.Check == 'Счастливый, но не ближайший':
count_nb += 1
else:
count_b += 1
time_of_task.append(tm/count)
if len(time_of_task) > 1:
diff_of_time.append(abs(time_of_task[i]-time_of_task[i-1]))
t.Diff = abs(time_of_task[i]-time_of_task[i-1])
i += 1
t.Time = tm
t.save()
elif exp.Strategy in (4,5,7,8):
i = 1
for t in tasks:
count = 0
tm = 0.0
for v in Variant.objects.filter(Variant_task=t):
full += 1
count += 1
times.append(v.Time)
tm += v.Time
if v.Check == 'Решено':
count_g += 1
elif v.Check == 'Счастливый, но не ближайший':
count_nb += 1
else:
count_b += 1
if exp.Strategy == 7 and count != 0:
time_of_task.append(tm/count)
t.Time = tm/count
elif exp.Strategy == 7 and count == 0:
time_of_task.append(0.0)
else:
time_of_task.append(tm)
t.Time = tm
if len(time_of_task) > 1:
diff_of_time.append(abs(time_of_task[i]-time_of_task[i-1]))
t.Diff = abs(time_of_task[i] - time_of_task[i-1])
i += 1
t.save()
Define_lev(t.id)
g_pct = count_g/full * 100
b_pct = count_b/full * 100
nb_pct = count_nb/full * 100
median_all = med(times)
args = {}
args['tasks'] = tasks
args['experiment'] = Experiment.objects.get(id=experiment_id)
args['g_pct'] = g_pct
args['b_pct'] = b_pct
args['nb_pct'] = nb_pct
args['avg_task_time'] = avg(time_of_task)
args['median_all'] = median_all #медиана времени ответа
args['avg_of_diff'] = sum(diff_of_time)/(len(diff_of_time)-1)
args['median_of_diff'] = med(diff_of_time)
return render_to_response('tasks/experiment.html', args)
