data = ld.getID2Data()
######If you want to select particular moments in the day
timeShift = -8 # San Francisco
data = ld.dataSelectTime(data, 0, 24, [0, 1, 2, 5, 6], timeShift)
####### Modulo 24 hours
# data = clus.temporalAround24removed(data)
data = clus.temporalAround7Daysremoved(data)
#######Remove time
# data=clus.removeTime(data)
kk = cluster.KMeans(n_clusters=5, random_state=0)
out = kk.fit(data)
clus.plotCluster(data, out.labels_, 5)
print("Cluster center " + str(out.cluster_centers_))
ld.fromDataClusterToCsv(data, out.labels_, "WorkingDays")
###### Select a center and get the 80% closest values to this center removed to far points
partialdata = clus.getThe80percentClosestToACenter(data, out, 0)
print(partialdata)
###### Second cluster on the partial data
"""
kk2 = cluster.KMeans(n_clusters=3,random_state=0)
out2 = kk2.fit(partialdata)
print(out2.cluster_centers_)
data = ld.getID1Data()
data = ld.dataSelectTime(data,10,16,[0,1,2,3,4],1)
data = clus.temporalAround24removed(data)
data=clus.removeTime(data)
kk = cluster.KMeans(n_clusters=10,random_state=0)
out = kk.fit(data)
clus.plotCluster(data,out.labels_,10)
print("Cluster center " + str(out.cluster_centers_))
ld.fromDataClusterToCsv(data,out.labels_,"Work")
partialdata = clus.getThe80percentClosestToACenter(data,out,0)
print(partialdata)
kk2 = cluster.KMeans(n_clusters=3,random_state=0)
out2 = kk2.fit(partialdata)
print(out2.cluster_centers_)
clus.plotCluster(partialdata,out2.labels_,3)
