def grayscale(im, algorithm=CIELab_gray):
if len(im.shape) > 2:
if im.shape[2] == 4:
return algorithm(premultiply(im))
else:
return algorithm(im)
else:
return im
def binarize(im, algorithm=adaptive_otsu, gray=CIELab_gray, resize=1.0):
if (im + 1 < 2).all(): # black and white
return im
else:
if resize < 0.99 or resize > 1.01:
im = cv2.resize(im, (0, 0), None, resize, resize)
return algorithm(grayscale(im, algorithm=gray))
def __init__(self, simulator, channel, algorithm, max_length,
frame_probability):
self.id = uuid.uuid4()
self.simulator = simulator
self.channel = channel
self.channel.client = self
self.algorithm = algorithm(self.channel, max_length)
self.max_length = max_length
self.frame_probability = frame_probability
self.frame = None
self.next_part = 0
logging.debug(
"Client %s : using algorithm %s on channel %s" %
(self.id, self.algorithm.__class__.__name__, self.channel.id))
def main():
try:
args = docopt(__doc__, version=__version__)
output_dir = None
input_dir, output_dir = get_dirs(args)
filenames = [
('alternatives.xml', False),
('outranking.xml', False),
('nonoutranking.xml', True),
('method_parameters.xml', False),
]
params = [
'alternatives',
'outranking',
'nonoutranking',
'crisp_outranking'
]
d = get_input_data(input_dir, filenames, params)
alternativesId = d.alternatives
outranking = d.outranking
nonoutranking = d.nonoutranking
crisp_outranking = d.crisp_outranking
alg = algorithm(alternativesId, outranking, nonoutranking, crisp_outranking)
result = alg.Run()
type = 'real'
if crisp_outranking == "true": type = 'integer'
xmcda = ranks_to_xmcda(result[0], type, None)
write_xmcda(xmcda, os.path.join(output_dir, 'nfs.xml'))
xmcda = ranks_to_xmcda(result[1], type, None)
write_xmcda(xmcda, os.path.join(output_dir, 'strength.xml'))
xmcda = ranks_to_xmcda(result[2], type, None)
write_xmcda(xmcda, os.path.join(output_dir, 'weakness.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
return 0
except Exception, err:
err_msg = get_error_message(err)
log_msg = traceback.format_exc()
print(log_msg.strip())
create_messages_file((err_msg, ), (log_msg, ), output_dir)
return 1
def __init__(self, graph, algorithm, routing_algo, jobs, \
num_mappers, num_reducers, num_path, paths, cpu=0, mem=0):
self.seed = randrange(100)
self.graph = graph
self.algorithm = algorithm(self.graph, routing_algo, num_mappers, num_reducers, paths)
self.jobs = jobs
self.num_mappers = num_mappers
self.num_reducers = num_reducers
self.cpu = cpu
self.mem = mem
self.f = open(Manager.LOG_NAME, 'a')
self.t = float(0) # Virtual time in the datacenter
self._write("%s %s %d %d\n" % (graph.get_name() + '_' + str(num_path), algorithm.get_name(), \
len(self.graph.get_hosts()), num_mappers))
def test_algorithm(problem, algorithm, algorithm_name):
""" Solve the given problem using the given algorithm """
# Extract problem parameters
n_products, n_dividers, costs = problem
# Set start time
start_time = time.time()
# Execute algorithm
answer = algorithm(n_products, n_dividers, costs)
# Compute end time
end_time = time.time()
# If VERBOSE, print the result and how long it took to compute
if VERBOSE:
print(
f"{algorithm_name}: {answer} ({end_time-start_time:.3f} seconds)")
# Return the answer
return answer
def __init__(self, graph, algorithm, routing_algo, jobs, \
num_mappers, num_reducers, num_path, paths, cpu=0, mem=0):
self.seed = randrange(100)
self.graph = graph
self.algorithm = algorithm(self.graph, routing_algo, num_mappers,
num_reducers, paths)
self.jobs = jobs
self.num_mappers = num_mappers
self.num_reducers = num_reducers
self.cpu = cpu
self.mem = mem
self.f = open(Manager.LOG_NAME, 'a')
self.t = float(0) # Virtual time in the datacenter
self._write("%s %s %d %d\n" % (graph.get_name() + '_' + str(num_path), algorithm.get_name(), \
len(self.graph.get_hosts()), num_mappers))
def main():
try:
args = docopt(__doc__, version=__version__)
output_dir = None
input_dir, output_dir = get_dirs(args)
filenames = [
('alternatives.xml', False),
('outranking.xml', False),
('preorder.xml', True),
('method_parameters.xml', True),
]
params = [
'alternatives',
'outranking',
'preorder',
'direction',
]
d = get_input_data(input_dir, filenames, params)
alternatives = d.alternatives
preorder = None
if (d.preorder is not None):
preorder = d.preorder
outranking = d.outranking
direction = d.direction
alg = algorithm(alternatives, outranking, preorder, direction)
result = alg.Run()
xmcda = ranks_to_xmcda(result, 'integer', None)
write_xmcda(xmcda, os.path.join(output_dir, 'ranking.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
return 0
except Exception, err:
err_msg = get_error_message(err)
log_msg = traceback.format_exc()
print(log_msg.strip())
create_messages_file((err_msg, ), (log_msg, ), output_dir)
return 1
def main():
try:
args = docopt(__doc__, version=__version__)
output_dir = None
input_dir, output_dir = get_dirs(args)
filenames = [
('alternatives.xml', False),
('downwards.xml', False),
('upwards.xml', False),
]
params = [
'alternatives',
'downwards',
'upwards',
]
d = get_input_data(input_dir, filenames, params)
alternativesId = d.alternatives
downwards = d.downwards
upwards = d.upwards
alg = algorithm(alternativesId, downwards, upwards)
result = alg.Run()
comparables = (alternativesId, alternativesId)
#xmcda = comparisons_to_xmcda(result[0], comparables)
xmcda = outranking_to_xmcda(result[0])
write_xmcda(xmcda, os.path.join(output_dir, 'intersection.xml'))
xmcda = ranks_to_xmcda(result[1], 'integer', None)
write_xmcda(xmcda, os.path.join(output_dir, 'rank.xml'))
xmcda = ranks_to_xmcda(result[2], 'integer', None)
write_xmcda(xmcda, os.path.join(output_dir, 'median.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
return 0
except Exception, err:
err_msg = get_error_message(err)
log_msg = traceback.format_exc()
print(log_msg.strip())
create_messages_file((err_msg, ), (log_msg, ), output_dir)
return 1
def main():
try:
args = docopt(__doc__, version=__version__)
output_dir = None
input_dir, output_dir = get_dirs(args)
filenames = [
('alternatives.xml', False),
('credibility.xml', False),
('method_parameters.xml', True),
]
params = [
'alternatives',
'credibility',
'direction',
'alpha',
'beta'
]
d = get_input_data(input_dir, filenames, params)
alternativesId = d.alternatives
credibility = d.credibility
direction = d.direction
alpha = d.alpha
beta = d.beta
alg = algorithm(alternativesId, credibility, direction, alpha, beta)
result = alg.Run()
xmcda = ranks_to_xmcda(result, 'integer', None)
write_xmcda(xmcda, os.path.join(output_dir, 'ranking.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
return 0
except Exception, err:
err_msg = get_error_message(err)
log_msg = traceback.format_exc()
print(log_msg.strip())
create_messages_file((err_msg, ), (log_msg, ), output_dir)
return 1
def index(request): # 主页
df = DataForm(request.POST, request.FILES)
if not request.user.is_authenticated():
massage = '请登陆'
back = 'login/'
return render_to_response('redirect.html', {'massage': massage, 'back': back})
if request.POST:
if request.user.is_authenticated():
user = request.user
if os.path.exists("/root/vina/demo/static/dv/%s/ok" % user):
shutil.rmtree("/root/vina/demo/static/dv/%s/ok" % user)
if os.path.exists("/root/vina/demo/static/dv/%s" % user):
shutil.rmtree("/root/vina/demo/static/dv/%s" % user)
os.mkdir("/root/vina/demo/static/dv/%s" % user)
if os.path.exists("/root/vina/demo/data/%s" % user):
shutil.rmtree("/root/vina/demo/data/%s" % user)
os.mkdir("/root/vina/demo/data/%s" % user)
os.mkdir("/root/vina/demo/static/dv/%s/ok" % user)
algo = request.POST['algo']
sourcefile = request.FILES['sourcefile']
labelname = None
try:
labelfile = request.FILES['labelfile']
labelname = request.FILES['labelfile'].name
except:
labelfile = None
try:
profile = request.user.userprofile
except Userprofile.DoesNotExist:
profile = Userprofile(user=request.user)
profile.user_id = user.id
profile.file = sourcefile
profile.labelfile = labelfile
profile.save()
user.save()
fileroot = request.FILES['sourcefile'].name
parameter = None
algo = int(algo)
if algo == 3:
parameter = request.POST['k-parameter']
if algo == 4:
parameter = request.POST['parameter']
sfile = '/root/vina/demo/data/'+fileroot
file='/root/vina/demo/data/%s/'%user+fileroot
shutil.copy(sfile,file)
os.remove(sfile)
try:
labelfile1 = '/root/vina/demo/data/'+labelname
labelfile='/root/vina/demo/data/%s/'%user+labelname
shutil.copy(labelfile1, labelfile)
os.remove(labelfile1)
except:
labelfile = None
i1 = GetInfo()
nodes = i1.Dianshu(file)
edges = i1.Bianshu(file)
averageweight = i1.PingjunQz(file)
result, q, listresult,g= algorithm(algo, file, parameter, labelfile, user)
if labelfile == None:
labelresult = result
else:
labelresult = numtolabel(labelfile, result)
q = round(q, 3)
averageweight = round(averageweight, 3)
communities = i1.Comnum(result)
net_src = "/static/dv/%s/SNA.png" % user
return render_to_response('SNA-main.html', {
'nodes': nodes,
'edges': edges,
'averageweight': averageweight,
'communities': communities,
'net_url': net_src,
'module': q,
})
return render(request, 'SNA-main.html', {"df": df})
def analyze(graph, runs, steps, algorithm, tasks, outcomes, workers, parameters):
total_ws = []
total_cs = []
cs = []
qs = []
ws = []
#print tasks
for i in range(0, runs):
resetWorkers(workers)
answers = algorithm(tasks, outcomes, workers, parameters)
#print answers
#print tasks
#print answers
#print tasks
#print answers
for i in range(0, len(tasks)):
if i >= len(ws):
ws.append(answers[i][1])
else:
ws[i] += answers[i][1]
if answers[i][0] == tasks[i]:
if i >= len(cs):
cs.append(1)
else:
cs[i] += 1
else:
if i >= len(cs):
cs.append(0)
#print cs
for i in range(0, len(cs)):
avg = cs[i]
count = 1
for j in range(1, steps + 1):
if i - j >= 0:
avg += cs[i - j]
count += 1
if i + j < len(cs):
avg += cs[i + j]
count += 1
qs.append(float(avg) / (float(runs) * float(count)))
cumulative = 0
cumulative_worker = 0
for i in range(0, len(cs)):
cs[i] = cs[i] + cumulative
total_cs.append(cs[i] / runs)
#total_ws.append(k * (i + 1))
cumulative = cs[i]
cumulative_worker += ws[i]
if i == len(cs) - 1:
print cs[i] / runs
print cumulative_worker / runs
cs[i] = float(cs[i]) / (float(runs) * float(i + 1))
ws[i] = float(ws[i]) / float(runs)
#print cs
#print qs
#labels
xs = np.arange(1, len(tasks) + 1, 1)
graph[0].plot(xs, cs, label='cumulative quality')
graph[0].plot(xs, qs, label='quality')
#graph[1].plot(total_ws, total_cs)
graph[1].plot(xs, ws, label='number of hired workers')
graph[0].legend(bbox_to_anchor=(1, 0.3))
graph[1].legend(bbox_to_anchor=(1, 0.8))
graph[0].set_xlabel('tasks')
graph[0].set_ylabel('accuracy')
graph[0].axis([0, 1000, 0.0, 1.0])
graph[0].set_autoscale_on(False)
graph[1].set_xlabel('tasks')
graph[1].set_ylabel('workers')
graph[1].axis([0, 1000, 0.0, 5.0])
graph[1].set_autoscale_on(False)
def run(algorithm, imgList, k):
myCentroids, clustAssing = algorithm(imgList, k)
return myCentroids, clustAssing
from loader import * from algorithm import * I = 12223 #images --> kjører et beskjedent antall blant de 60000, basert på Stochastic Gradient Descent. #Blir ca 12000 bilder maksimalt med kun utvalg på to siffer. d = 784 # 28x28 Y_0, c = get_dataset() Y_Kk, J, omega, my, iterations, Z = algorithm(Y_0, c, I, d, "training") Y0_chunk, chunk = stochastic_gradient_descent(I, Y_0) print(Y_0) plot_cost_function_convergence(iterations, J) plot_progression(Y_Kk, c) r = 1000 #resolution of plot plot_separation(last_function, Y_Kk[-1, :, :], c, r, omega, my)
def main():
try:
args = docopt(__doc__, version=__version__)
output_dir = None
input_dir, output_dir = get_dirs(args)
filenames = [
('method_parameters.xml', False)
]
params = [
'comparison_with'
]
d = get_input_data(input_dir, filenames, params)
comparison_with = d.comparison_with
profilesId = None
if comparison_with == "profiles":
filenames = [
('alternatives.xml', False),
('profiles.xml', False),
('criteria.xml', False),
('weights.xml', False),
('discordance.xml', False),
]
params = [
'alternatives',
'profiles',
'criteria',
'weights',
'discordance',
]
kwargs = {'use_partials': True, 'comparison_with': comparison_with}
d = get_input_data(input_dir, filenames, params, **kwargs)
profilesId = d.profiles
alternativesId = d.alternatives
criteriaId = d.criteria
weights = d.weights
discordance = d.discordance
else:
filenames = [
('alternatives.xml', False),
('criteria.xml', False),
('weights.xml', False),
('discordance.xml', False),
]
params = [
'alternatives',
'criteria',
'weights',
'discordance',
]
kwargs = {'use_partials': True}
d = get_input_data(input_dir, filenames, params, **kwargs)
alternativesId = d.alternatives
criteriaId = d.criteria
weights = d.weights
discordance = d.discordance
alg = algorithm(alternativesId, profilesId, criteriaId, weights, discordance)
result = alg.Run()
if profilesId == None:
comparables = (alternativesId, alternativesId)
xmcda = comparisons_to_xmcda(result, comparables, None)
write_xmcda(xmcda, os.path.join(output_dir, 'discordance.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
else:
comparables = (alternativesId, profilesId)
xmcda = comparisons_to_xmcda(result, comparables, None, with_profile = True)
write_xmcda(xmcda, os.path.join(output_dir, 'discordance.xml'))
create_messages_file(None, ('Everything OK.',), output_dir)
return 0
except Exception, err:
err_msg = get_error_message(err)
log_msg = traceback.format_exc()
print(log_msg.strip())
create_messages_file((err_msg, ), (log_msg, ), output_dir)
return 1
def index(request): # 主页
df = DataForm(request.POST, request.FILES)
if not request.user.is_authenticated():
massage = '请登陆'
back = 'login/'
return render_to_response('redirect.html', {
'massage': massage,
'back': back
})
if request.POST:
if request.user.is_authenticated():
user = request.user
if os.path.exists("/root/vina/demo/static/dv/%s/ok" % user):
shutil.rmtree("/root/vina/demo/static/dv/%s/ok" % user)
if os.path.exists("/root/vina/demo/static/dv/%s" % user):
shutil.rmtree("/root/vina/demo/static/dv/%s" % user)
os.mkdir("/root/vina/demo/static/dv/%s" % user)
if os.path.exists("/root/vina/demo/data/%s" % user):
shutil.rmtree("/root/vina/demo/data/%s" % user)
os.mkdir("/root/vina/demo/data/%s" % user)
os.mkdir("/root/vina/demo/static/dv/%s/ok" % user)
algo = request.POST['algo']
sourcefile = request.FILES['sourcefile']
labelname = None
try:
labelfile = request.FILES['labelfile']
labelname = request.FILES['labelfile'].name
except:
labelfile = None
try:
profile = request.user.userprofile
except Userprofile.DoesNotExist:
profile = Userprofile(user=request.user)
profile.user_id = user.id
profile.file = sourcefile
profile.labelfile = labelfile
profile.save()
user.save()
fileroot = request.FILES['sourcefile'].name
parameter = None
algo = int(algo)
if algo == 3:
parameter = request.POST['k-parameter']
if algo == 4:
parameter = request.POST['parameter']
sfile = '/root/vina/demo/data/' + fileroot
file = '/root/vina/demo/data/%s/' % user + fileroot
shutil.copy(sfile, file)
os.remove(sfile)
try:
labelfile1 = '/root/vina/demo/data/' + labelname
labelfile = '/root/vina/demo/data/%s/' % user + labelname
shutil.copy(labelfile1, labelfile)
os.remove(labelfile1)
except:
labelfile = None
i1 = GetInfo()
nodes = i1.Dianshu(file)
edges = i1.Bianshu(file)
averageweight = i1.PingjunQz(file)
result, q, listresult, g = algorithm(algo, file, parameter,
labelfile, user)
if labelfile == None:
labelresult = result
else:
labelresult = numtolabel(labelfile, result)
q = round(q, 3)
averageweight = round(averageweight, 3)
communities = i1.Comnum(result)
net_src = "/static/dv/%s/SNA.png" % user
return render_to_response(
'SNA-main.html', {
'nodes': nodes,
'edges': edges,
'averageweight': averageweight,
'communities': communities,
'net_url': net_src,
'module': q,
})
return render(request, 'SNA-main.html', {"df": df})
import datasets
from algorithm import *
m = 2
# train_points, train_labels, test_points, test_labels = datasets.subset(datasets.cifar10(), 5000, 1000)
# train_points, train_labels, test_points, test_labels = datasets.subset(datasets.cifarX(2), 5000, 1000)
train_points, train_labels, test_points, test_labels = datasets.subset(
datasets.mnist10(), 5000, 1000)
print('Loaded input')
# print('Raw nearest-neighbor accuracy:', nn_accuracy(train_points, train_labels, test_points, test_labels))
train_points_2, test_points_2, graphs = algorithm(
train_points,
m=m,
test_points=test_points,
pca_dim=100,
loss_weights=(1, 1, 1),
)
print(nn_accuracy(train_points_2, train_labels, test_points_2, test_labels))
plt_plot(train_points_2, train_labels)
plt.plot(graphs)
plt.show()
# train_points_1, test_points_1 = pca(train_points, m, test_points)
# print('PCA nearest-neighbor accuracy:', nn_accuracy(train_points_1, train_labels, test_points_1, test_labels))
#
# for i in range(5):
# train_points_2, test_points_2, graphs = algorithm(
# train_points,
'Nov', 'Dec'
]
# Level 1
# Calculate power usage predictions (and convert to GWh)
predicted_power_usage, mem_history_1 = get_predicted_power_usage(2022)
predicted_power_usage = predicted_power_usage.divide(1000000)
# Display output of power usage (past & future):
#mem_history_2 = visualize_data(predicted_power_usage)
# Output results of predtictions to csv:
predicted_power_usage.columns = [
'Zone 1 (GWh)', 'Zone 2 (GWh)', 'Zone 3 (GWh)', 'Zone 4 (GWh)',
'Zone 5 (GWh)', 'Zone 6 (GWh)', 'Zone 7 (GWh)'
]
predicted_power_usage.index = months
predicted_power_usage = predicted_power_usage.round(2)
predicted_power_usage.to_csv(r"./L1_output.csv")
# Level 2
# Optimize the Costs
costs_output = algorithm(2022)
costs_output.columns = [
'Provincial Cost ($)', 'Total Power Consumed (GWh)',
'Renewable Power Used (%)'
]
costs_output.index = months
costs_output = costs_output.round(2)
costs_output.to_csv(r"./L2_output.csv")
autoRun = False
if sys.argv[1] == '-a':
autoRun = True
while True:
pygame.display.update()
for event in pygame.event.get():
if event.type == QUIT:
trainer.complete()
pygame.quit()
sys.exit()
if event.type == MOUSEBUTTONDOWN:
if nowPlayer.index == 0:
algorithm(player, nowPlayer, map, sys.argv[2], trainer, i)
else:
#myAlgo(player,nowPlayer,map)
algorithm(player, nowPlayer, map, sys.argv[3], trainer, i)
setMap()
infoUpdate(player, i)
pygame.display.update()
if autoRun:
while not player[0].isLose() and not player[1].isLose():
nowPlayer = player[(nowPlayer.index + 1) % 2]
i = i + 1
if nowPlayer.index == 0:
#randomMove(player,nowPlayer,map)
algorithm(player, nowPlayer, map, sys.argv[2], trainer,
def analyze(graph, runs, steps, algorithm, tasks, outcomes, workers, parameters):
total_ws = []
total_cs = []
cs = []
qs = []
ws = []
#print tasks
for i in range(0, runs):
resetWorkers(workers)
answers = algorithm(tasks, outcomes, workers, parameters)
#print answers
#print tasks
#print answers
#print tasks
#print answers
for i in range(0, len(tasks)):
if i >= len(ws):
ws.append(answers[i][1])
else:
ws[i] += answers[i][1]
if answers[i][0] == tasks[i]:
if i >= len(cs):
cs.append(1)
else:
cs[i] += 1
else:
if i >= len(cs):
cs.append(0)
#print cs
for i in range(0, len(cs)):
avg = cs[i]
count = 1
for j in range(1, steps + 1):
if i - j >= 0:
avg += cs[i - j]
count += 1
if i + j < len(cs):
avg += cs[i + j]
count += 1
qs.append(float(avg) / (float(runs) * float(count)))
cumulative = 0
cumulative_worker = 0
for i in range(0, len(cs)):
cs[i] = cs[i] + cumulative
total_cs.append(cs[i] / runs)
#total_ws.append(k * (i + 1))
cumulative = cs[i]
cumulative_worker += ws[i]
if i == len(cs) - 1:
print cs[i] / runs
print cumulative_worker / runs
cs[i] = float(cs[i]) / (float(runs) * float(i + 1))
ws[i] = float(ws[i]) / float(runs)
#print cs
#print qs
#labels
result = {}
result['cs'] = cs
result['qs'] = qs
result['ws'] = ws
global flag
print 'flag', flag
if flag == 0:
with open('rk.json', 'w') as outfile:
json.dump(result, outfile)
elif flag == 1:
with open('tk.json', 'w') as outfile:
json.dump(result, outfile)
elif flag == 2:
with open('dh.json', 'w') as outfile:
json.dump(result, outfile)
flag += 1
xs = np.arange(1, len(tasks) + 1, 1)