from triager import Triager
from tablemaker import make_table
from mailer import send_mail
if __name__ == "__main__":
triager = Triager()
print("Gathering issues for triage...")
issues = triager.triage()
if issues:
table = make_table(issues)
print(table)
print("Mailing table to maintainers...")
send_mail(
content=table,
sender=triager.sender,
receivers=triager.maintainers,
)
metric = random.choice(metrics)
return N, K, distribution, xi, metric
def trial(trial_number, metric, K, N, distribution, xi):
X = point.PointCloud.generate_with_classes(N, K, distribution, [True, False])
Y = point.PointCloud.generate(N, M, distribution)
for y in Y:
closest_point = y.nearest_point(X, metric)
y.clazz = closest_point.clazz
perturbed_Y = deepcopy(Y).perturb_points(xi)
for y in perturbed_Y:
closest_point = y.nearest_point(X, metric)
y.clazz = closest_point.clazz
confusion_matrix = ConfusionMatrix.generate_from_point_clouds(Y, perturbed_Y)
result = [trial_number, metric, K, N, xi.sigma] + confusion_matrix.row()
print(result)
return result
results = []
for trial_number in range(trials):
N, K, distribution, xi, metric = generate_params(N_Range, K_Range, metrics, distributions, sigma_range)
results.append(trial(trial_number, metric, K, N, distribution, xi))
tablemaker.make_table('experiment1-2(1)', ['Trial', 'Metric', 'K', 'N', 'sigma'] + ConfusionMatrix.headers(), results)
import point
from helpertools import Metric
import random
from numpy import inf
from statistics import mean, stdev
import tablemaker
def trial(trial_number, metric, K, N, distribution):
point_cloud = point.PointCloud.generate(N, K, distribution)
point_distances = (point.distances(point_cloud.points, metric) for point in point_cloud.points)
ratios = [min(distances)/max(distances) for distances in point_distances]
result = trial_number, metric, K, N, mean(ratios), stdev(ratios)
print(result)
return result
N_Values = list(range(1, 11)) + list(range(20, 101, 10)) # The dimension of space
K_Values = [5, 50, 500] # The number of point values
metrics = [Metric(ord) for ord in [1, 2, inf]] # Different norms to use
trials = 25
distribution = (lambda: random.uniform(0, 1)) # Point Distribution
results = [trial(trial_number, metric, K, N, distribution) for metric in metrics for N in N_Values for K in K_Values for trial_number in range(trials)]
tablemaker.make_table('experiment1-1', ['TrialNumber', 'Metric', 'K', 'N', 'Mean', 'StandardDeviation'], results)
