def createFiles(logFile, kOptions):
log = xes_factory.apply(logFile)
dataVectors, seq = preprocess.dataPreprocess2012(log)
pairWiseData = outlierPairWise.preprocess(log)
#if needed to calculate time to create tree
timeTreeStart = time.time()
tree = outlierPairWise.createRtree(pairWiseData) #returns values orderd
timeTreeEnd = time.time()
with open("tests/data/timeDistancesEvent.txt", "w") as f:
for k in kOptions:
pairs, executionTime = outlierDistanceActivities.main(
dataVectors, seq, k)
f.write(str(k) + "," + str(executionTime) + "\n")
with open('tests/data/timeDistancesRtree.txt', 'w') as f:
for k in kOptions:
startPairWist = time.time()
scores = outlierPairWise.outlierScore(k, tree, pairWiseData)
scoreTimeEnd = time.time()
outliers = []
for s in scores:
pairData = pairWiseData[s[0]]
outliers.append(pairData + [s[1]])
f.write(str(k) + "," + str(scoreTimeEnd - startPairWist) + "\n")
from algorithms import outlierDistanceActivities, preprocess, outlierPairsDistribution
import algorithms.outlierDistanceActivities as activities
import algorithms.outlierPairsDistribution as distribution
kOptions=[250,500,750,1000,1250,1500,1750,2000]
thresholds=[0.0025,0.005,0.0075,0.01,0.0125,0.0150,0.0175,0.02]
logFiles=["../BPI_Challenge_2012.xes","BPI Challenge 2017.xes"]
from pm4py.objects.log.importer.xes import factory as xes_factory
for logFile in logFiles:
print("importing log")
log = xes_factory.apply(logFile)
# [trace,activity index,time]
print("preprocess ...")
if logFile=="../BPI_Challenge_2012.xes":
dataVectors, seq = preprocess.dataPreprocess2012(log)
identifier="2012"
else:
dataVectors, seq = preprocess.dataPreprocess2017(log)
identifier="2017"
with open("tests/data/events-{}-distance.txt".format(identifier),"w") as f:
for k in kOptions:
pairs,time=outlierDistanceActivities.main(dataVectors,seq,k)
f.write(str(k)+","+str(time)+"\n")
with open("tests/data/events-{}-distribution.txt".format(identifier),"w") as f:
for threshold in thresholds:
pairs,time=outlierPairsDistribution.main(log,dataVectors,seq,threshold)
f.write(str(threshold)+","+str(time)+"\n")
os.remove("distributions.txt")
#plot the times
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
from algorithms.outlierPairsCurveFitting import getDistributionsFitting
from pm4py.algo.filtering.log.attributes import attributes_filter as log_attributes_filter
from pm4py.objects.log.importer.xes import factory as xes_factory
from algorithms import preprocess
from statistics import mean
import numpy as np
import warnings, scipy
from sklearn.preprocessing import StandardScaler
#time: per activity [30] => [trace index, activity index, time in seconds]
#sequence: per trace [10.000]=> [[activity index, time]]
logFile = "../BPM Temporal Anomalies/scala_trace_outlier/input/outliers_30_activities_10k_0.005.xes"
logFileResults = "../BPM Temporal Anomalies/scala_trace_outlier/input/results_30_activities_10k_0.005_description"
log = xes_factory.apply(logFile)
time, sequence = preprocess.dataPreprocess2012(log)
times = [[i[2] for i in x] for x in time]
means = [mean(i) for i in times]
distributionsDF = getDistributionsFitting(times, log)
#get the distributions in a array
thresholds = [0.020, 0.01, 0.0075, 0.005, 0.0025, 0.001]
for threshold in thresholds:
warnings.filterwarnings("ignore")
distributions = []
for index in range(len(distributionsDF)):
if distributionsDF.iloc[index]["R2"] >= 0.9:
dist = getattr(scipy.stats,
distributionsDF.iloc[index]["Distribution"])
param = dist.fit(times[index])