def runScript(bestRankersFile, frK, tK): #"bestRanker.p" sys.argv[1]
fromK = int(frK)
toK = int(tK) + 1
global bestKClusterGroup, queryRankerList, queryRankerDict
#commented out part is for test purposes
#data = np.vstack((rand(150,2) + np.array([.5,.5]),rand(150,2), rand(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
bestKClusterGroup1 = get_best_clusters(
getData(bestRankersFile), fromK,
toK) #list > list(cluster#) > np.array,np.array etc...
bestKClusterGroup2 = []
#converting list > list(cluster#) > np.array (ranker),np.array etc... to list > list(cluster#-->index of list) > normal list(ranker),list etc...
for i in range(len(bestKClusterGroup1)):
bestKClusterGroup2.append([])
for j in range(len(bestKClusterGroup1[i])):
bestKClusterGroup2[i].append(bestKClusterGroup1[i][j].tolist())
clusterDataObject = clusterData()
for i in range(len(bestKClusterGroup2)):
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
clusterDataObject.clusterToRanker[i] = bestKClusterGroup2[i]
print type(clusterDataObject.clusterToRanker[i]), len(
clusterDataObject.clusterToRanker[i])
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker:
for j in clusterDataObject.clusterToRanker[i]:
for k in queryRankerDict:
if type(queryRankerDict[k]) == list:
for l in queryRankerDict[k]:
if l.tolist() == j:
if k in clusterDataObject.queryToCluster.keys():
clusterDataObject.queryToCluster[k].append(i)
else:
clusterDataObject.queryToCluster[k] = [i]
elif queryRankerDict[k].tolist() == j:
clusterDataObject.queryToCluster[k] = i
'''for i in clusterDataObject.queryToCluster:
print i, clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, clusterDataObject.clusterToRanker[i]'''
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
paths = bestRankersFile.split('/')
name = paths[len(paths) - 1]
parts = name.split('.')
name = parts[0]
pickle.dump(clusterDataObject, open("ClusterData/" + name + ".data", "wb"))
'''print '-----------------Print output of one of the object files-----------------------'
loadedFile = pickle.load( open( "../../../ClusterData/clusterToRankerDict.data", "rb" ) )
for i in loadedFile:
print i
#sys.exit()
for j in loadedFile[i]:
print j'''
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker
def runScript(self):#"bestRanker.p" sys.argv[1]
#commented out part is for test purposes
#data = np.vstack((random(150,2) + np.array([.5,.5]),random(150,2), random(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
data = self.getData()
self.bestKClusterGroup1 = self.get_best_clusters(data) #list > list(cluster#) > np.array,np.array etc...
self.bestKClusterGroup2 = []
#converting list > list(cluster#) > np.array (ranker),np.array etc... to list > list(cluster#-->index of list) > normal list(ranker),list etc...
for i in range(len(self.bestKClusterGroup1)):
self.bestKClusterGroup2.append([])
for j in range(len(self.bestKClusterGroup1[i])):
self.bestKClusterGroup2[i].append(self.bestKClusterGroup1[i][j].tolist())
clusterDataObject = clusterData()
for i in range(len(self.bestKClusterGroup2)):
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
clusterDataObject.clusterToRanker[i] = self.bestKClusterGroup2[i]
print type(clusterDataObject.clusterToRanker[i]), len(clusterDataObject.clusterToRanker[i])
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker:
for j in clusterDataObject.clusterToRanker[i]:
for k in self.queryRankerDict:
if type(self.queryRankerDict[k]) == list:
for l in self.queryRankerDict[k]:
if l.tolist() == j:
if k in clusterDataObject.queryToCluster.keys():
clusterDataObject.queryToCluster[k].append(i)
else:
clusterDataObject.queryToCluster[k] = [i]
elif self.queryRankerDict[k].tolist() == j:
clusterDataObject.queryToCluster[k] = i
'''for i in clusterDataObject.queryToCluster:
print i, clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, clusterDataObject.clusterToRanker[i]'''
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
pickle.dump(clusterDataObject, open("ClusterData/"+self.dataset+str(self.iterations)+'.data', "wb"))
#pickle.dump(clusterDataObject, open("ClusterData/"+self.dataset+" k"+self.bestK+".data", "wb"))
#pickle.dump(clusterDataObject.queryToCluster, open( "ClusterData/queryToClusterDict.data", "wb" ) )
#pickle.dump(clusterDataObject.clusterToRanker, open( "ClusterData/clusterToRankerDict.data", "wb" ) )
'''print '-----------------Print output of one of the object files-----------------------'
loadedFile = pickle.load( open( "ClusterData/clusterToRankerDict.data", "rb" ) )
for i in loadedFile:
print i
#sys.exit()
for j in loadedFile[i]:
print j'''
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker
def runScript(self): #"bestRanker.p" sys.argv[1]
#commented out part is for test purposes
#data = np.vstack((random(150,2) + np.array([.5,.5]),random(150,2), random(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
data = self.getData()
dataToClusters = self.getClusters(
data) #list > list(cluster#) > np.array,np.array etc...
dataToClusters = list(dataToClusters)
clusterDataObject = clusterData()
data = list(data)
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
for i in range(len(dataToClusters)):
if not dataToClusters[i] in clusterDataObject.clusterToRanker.keys(
):
clusterDataObject.clusterToRanker[dataToClusters[i]] = [
list(data[i])
]
else:
clusterDataObject.clusterToRanker[dataToClusters[i]].append(
list(data[i]))
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker: #for each cluster
for j in clusterDataObject.clusterToRanker[
i]: #for each ranker in cluster
for q in self.queryRankerDict: #for each query
for r in self.queryRankerDict[
q]: #for each ranker in query
if list(
r
) == j: #if ranker in query is equal to j (current ranker in cluster)
if q in clusterDataObject.queryToCluster: #if query key exists in dictionary
clusterDataObject.queryToCluster[q].append(i)
else:
clusterDataObject.queryToCluster[q] = [i]
for i in clusterDataObject.queryToCluster:
print i, len(clusterDataObject.queryToCluster[i]
), clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, len(clusterDataObject.clusterToRanker[i]
) #, clusterDataObject.clusterToRanker[i]
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
pickle.dump(clusterDataObject,
open("ClusterData/" + self.dataset + ".data", "wb"))
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker
def runScript(self):#"bestRanker.p" sys.argv[1]
#commented out part is for test purposes
#data = np.vstack((random(150,2) + np.array([.5,.5]),random(150,2), random(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
data = self.getData()
dataToClusters = self.getClusters(data) #list > list(cluster#) > np.array,np.array etc...
dataToClusters = list(dataToClusters)
clusterDataObject = clusterData()
data = list(data)
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
for i in range(len(dataToClusters)):
if not dataToClusters[i] in clusterDataObject.clusterToRanker.keys():
clusterDataObject.clusterToRanker[dataToClusters[i]] = [list(data[i])]
else:
clusterDataObject.clusterToRanker[dataToClusters[i]].append(list(data[i]))
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker:#for each cluster
for j in clusterDataObject.clusterToRanker[i]:#for each ranker in cluster
for q in self.queryRankerDict:#for each query
for r in self.queryRankerDict[q]:#for each ranker in query
if list(r) == j:#if ranker in query is equal to j (current ranker in cluster)
if q in clusterDataObject.queryToCluster:#if query key exists in dictionary
clusterDataObject.queryToCluster[q].append(i)
else:
clusterDataObject.queryToCluster[q] = [i]
for i in clusterDataObject.queryToCluster:
print i, len(clusterDataObject.queryToCluster[i]), clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, len(clusterDataObject.clusterToRanker[i])#, clusterDataObject.clusterToRanker[i]
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
pickle.dump(clusterDataObject, open("ClusterData/"+self.dataset+".data", "wb"))
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker
def runScript(bestRankersFile, frK, tK):#"bestRanker.p" sys.argv[1]
fromK = int(frK)
toK = int(tK)+1
global bestKClusterGroup, queryRankerList, queryRankerDict
#commented out part is for test purposes
#data = np.vstack((rand(150,2) + np.array([.5,.5]),rand(150,2), rand(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
bestKClusterGroup1 = get_best_clusters(getData(bestRankersFile),fromK,toK) #list > list(cluster#) > np.array,np.array etc...
bestKClusterGroup2 = []
#converting list > list(cluster#) > np.array (ranker),np.array etc... to list > list(cluster#-->index of list) > normal list(ranker),list etc...
for i in range(len(bestKClusterGroup1)):
bestKClusterGroup2.append([])
for j in range(len(bestKClusterGroup1[i])):
bestKClusterGroup2[i].append(bestKClusterGroup1[i][j].tolist())
clusterDataObject = clusterData()
for i in range(len(bestKClusterGroup2)):
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
clusterDataObject.clusterToRanker[i] = bestKClusterGroup2[i]
print type(clusterDataObject.clusterToRanker[i]), len(clusterDataObject.clusterToRanker[i])
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker:
for j in clusterDataObject.clusterToRanker[i]:
for k in queryRankerDict:
if type(queryRankerDict[k]) == list:
for l in queryRankerDict[k]:
if l.tolist() == j:
if k in clusterDataObject.queryToCluster.keys():
clusterDataObject.queryToCluster[k].append(i)
else:
clusterDataObject.queryToCluster[k] = [i]
elif queryRankerDict[k].tolist() == j:
clusterDataObject.queryToCluster[k] = i
'''for i in clusterDataObject.queryToCluster:
print i, clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, clusterDataObject.clusterToRanker[i]'''
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
paths=bestRankersFile.split('/')
name=paths[len(paths)-1]
parts=name.split('.')
name=parts[0]
pickle.dump(clusterDataObject, open( "ClusterData/"+name+".data", "wb" ) )
'''print '-----------------Print output of one of the object files-----------------------'
loadedFile = pickle.load( open( "../../../ClusterData/clusterToRankerDict.data", "rb" ) )
for i in loadedFile:
print i
#sys.exit()
for j in loadedFile[i]:
print j'''
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker
def runScript(self): #"bestRanker.p" sys.argv[1]
#commented out part is for test purposes
#data = np.vstack((random(150,2) + np.array([.5,.5]),random(150,2), random(150,2) + np.array([2.5,2.5]), rand(150,2) + np.array([10.5,10.5])))
data = self.getData()
self.bestKClusterGroup1 = self.get_best_clusters(
data) #list > list(cluster#) > np.array,np.array etc...
self.bestKClusterGroup2 = []
#converting list > list(cluster#) > np.array (ranker),np.array etc... to list > list(cluster#-->index of list) > normal list(ranker),list etc...
for i in range(len(self.bestKClusterGroup1)):
self.bestKClusterGroup2.append([])
for j in range(len(self.bestKClusterGroup1[i])):
self.bestKClusterGroup2[i].append(
self.bestKClusterGroup1[i][j].tolist())
clusterDataObject = clusterData()
for i in range(len(self.bestKClusterGroup2)):
#make object ---> dict[clusterNumber:int] = list of all rankers (where rankers are also lists)
clusterDataObject.clusterToRanker[i] = self.bestKClusterGroup2[i]
print type(clusterDataObject.clusterToRanker[i]), len(
clusterDataObject.clusterToRanker[i])
#make object ---> dict[queryID:string] = list of cluster numbers as ints
for i in clusterDataObject.clusterToRanker:
for j in clusterDataObject.clusterToRanker[i]:
for k in self.queryRankerDict:
if type(self.queryRankerDict[k]) == list:
for l in self.queryRankerDict[k]:
if l.tolist() == j:
if k in clusterDataObject.queryToCluster.keys(
):
clusterDataObject.queryToCluster[k].append(
i)
else:
clusterDataObject.queryToCluster[k] = [i]
elif self.queryRankerDict[k].tolist() == j:
clusterDataObject.queryToCluster[k] = i
'''for i in clusterDataObject.queryToCluster:
print i, clusterDataObject.queryToCluster[i]
for i in clusterDataObject.clusterToRanker:
print i, clusterDataObject.clusterToRanker[i]'''
if not os.path.exists("ClusterData"):
os.makedirs("ClusterData")
pickle.dump(
clusterDataObject,
open(
"ClusterData/" + self.dataset + str(self.iterations) + '.data',
"wb"))
#pickle.dump(clusterDataObject, open("ClusterData/"+self.dataset+" k"+self.bestK+".data", "wb"))
#pickle.dump(clusterDataObject.queryToCluster, open( "ClusterData/queryToClusterDict.data", "wb" ) )
#pickle.dump(clusterDataObject.clusterToRanker, open( "ClusterData/clusterToRankerDict.data", "wb" ) )
'''print '-----------------Print output of one of the object files-----------------------'
loadedFile = pickle.load( open( "ClusterData/clusterToRankerDict.data", "rb" ) )
for i in loadedFile:
print i
#sys.exit()
for j in loadedFile[i]:
print j'''
return clusterDataObject.queryToCluster, clusterDataObject.clusterToRanker