# COPYRIGHT 2015 Mohammed AlMalki

# for the used dataset, please refer to http://cvrr.ucsd.edu/bmorris/datasets/dataset_trajectory_clustering.html

from lshash import LSHash

from sklearn.datasets import fetch_mldata, load_iris, load_digits

from sklearn.metrics.cluster import normalized_mutual_info_score

from sklearn.cluster import KMeans

from sklearn import datasets

import numpy as np

import sets

import cPickle

import gzip

import scipy.io

import time

import timeit

import sets

import datetime as dt

import measurements

#------------------------------------------------------------------------------

# Prepration of the output file, initialization of LSH object and parameters

#------------------------------------------------------------------------------

dimensionNumber = 2 # as for now 2 dimensions for the longitude and latitude

numberHFs = 25

numberRadius = 251

usedDataset = 'CVRR_dataset_trajectory_clustering/cross.mat'

fileContainer = open('Pointwise LSH Clustering Experiment -ROUND-100- ('+usedDataset[35:-4]+') HFs_'+ str(numberHFs)+'_R_'+str(numberRadius), 'a')

fileContainer.write('\n')

fileContainer.write('Welcome to our experiment : ')

fileContainer.write(str( '\nSTRATING TIME : '+ time.asctime( time.localtime(time.time()) )))

print str( '\nSTRATING TIME : '+ time.asctime( time.localtime(time.time()) ))

fileContainer.write('\n')

fileContainer.write('The discription needed for each result will be provided accordingly .....')

fileContainer.write('\n')

fileContainer.write('The used Dataset is : '+usedDataset)

fileContainer.write('\n')

print '\nStarting LSH initialization ...'

fileContainer.write(str( '\nTime before LSH initialization : '+ time.asctime( time.localtime(time.time()) )))

newLsh = LSHash(numberHFs, dimensionNumber, num_hashtables = 1)

fileContainer.write(str( '\nTime after LSH initialization : '+ time.asctime( time.localtime(time.time()) )))

print '\nStarting loading the trajectory dataset ...'

fileContainer.write(str( '\nTime before loading the trajectory dataset : '+ time.asctime( time.localtime(time.time()) )))

#------------------------------------------------------------------------------

# The Trajectory dataset - cross

#------------------------------------------------------------------------------

mat = scipy.io.loadmat(usedDataset)

datasetSize = len(mat.values()[0])

trajectoriesContainer = []

for i in range(datasetSize):

trajectoriesContainer.append([(mat.values()[0][i][0][0][j], mat.values()[0][i][0][1][j]) for j in range(len(mat.values()[0][i][0][0]))])

allPoints = []

fileContainer.write(str( '\nTime after loading trajectory dataset : '+ time.asctime( time.localtime(time.time()) )))

#------------------------------------------------------------------------------

# indexing all trajectories

print '\nStarting the indexing procedure ...'

fileContainer.write(str( 'Time before indixing all trajectories points : '+ time.asctime( time.localtime(time.time()) ) ))

queryDictionary = {}

numberOfPoints = 0

for i, trajectory in enumerate(trajectoriesContainer):

for point in trajectory:

hash = newLsh.index(point, loadF=numberRadius)

if queryDictionary.has_key(hash):

queryDictionary[hash].add(i)

else:

queryDictionary[hash] = set()

queryDictionary[hash].add(i)

numberOfPoints += 1

fileContainer.write('\nThe following is the hash table used for querying or clustering ...')

fileContainer.write(str(queryDictionary))

fileContainer.write('\n')

fileContainer.write(str( 'The number of generated buckets is : '+ str(len(queryDictionary.keys()))))

fileContainer.write('\n')

fileContainer.write(str( 'Time after indixing all trajectories points : '+ time.asctime( time.localtime(time.time()) ) ))

fileContainer.write('\n')

fileContainer.write(str( 'The number of point have beenindexed is : '+ str(numberOfPoints)))

bucketingResult = queryDictionary

fileContainer.write('\n')

fileContainer.write(str( 'Clustering time : '+ time.asctime( time.localtime(time.time()) )))

fileContainer.write('\n')

print len(bucketingResult.keys())

fileContainer.write('\n')

fileContainer.write(str(bucketingResult))

print 'Start Constructing new feature space : ', time.asctime( time.localtime(time.time()) )

NFSdimensions = (1900, len(bucketingResult.keys()))

newFeatureSpace = np.zeros(NFSdimensions)

print len(newFeatureSpace)

print len(newFeatureSpace[0])

for i, key in enumerate(bucketingResult):

for trajectory in bucketingResult[key]:

newFeatureSpace[trajectory][i] = 1

print 'End Constructing new feature space : ', time.asctime( time.localtime(time.time()) )

totallG = 0 # this is intended to compute G in the space complexity in your method . G is the avergae number of trajectories in each feature space

for i, key in enumerate(bucketingResult):

totallG += len(bucketingResult[key])

print 'totallG : ', totallG

# print newFeatureSpace[1540]

# print 'KMeans Clustering'

print 'Start clustering : ', time.asctime( time.localtime(time.time()) )

KMeansResults = KMeans(init='k-means++', n_clusters=19, n_init=100).fit(newFeatureSpace)

print 'End clustering : ', time.asctime( time.localtime(time.time()) )

# for i in range(20):

# print KMeansResults.predict(X[i])

print [KMeansResults.predict(newFeatureSpace[i]) for i in range(10)]

print [(i, KMeansResults.labels_[i]) for i in range(20)]

clusteringResults = [[], [],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[]]

for i, clustering in enumerate(KMeansResults.labels_):

clusteringResults[clustering].append(i)

fileContainer.write('\n')

fileContainer.write(str(clusteringResults))

fileContainer.write('\n')

fileContainer.write(str([len(i) for i in clusteringResults]))

print clusteringResults[0]

print clusteringResults[1]

print [len(i) for i in clusteringResults]

crossDatasetTrueClassification = [[36, 45, 48, 95, 122, 127, 213, 227, 252, 258, 281, 289, 296, 301, 327, 332, 360, 373, 466, 503, 511, 550, 561, 564, 582, 605, 615, 626, 690, 733, 736, 754, 766, 789, 794, 801, 810, 874, 888, 893, 906, 918, 937, 938, 973, 1028, 1039, 1050, 1071, 1095, 1097, 1111, 1174, 1199, 1241, 1256, 1270, 1303, 1327, 1335, 1347, 1351, 1374, 1384, 1391, 1396, 1406, 1428, 1429, 1432, 1465, 1470, 1480, 1498, 1502, 1506, 1516, 1565, 1585, 1591, 1607, 1639, 1715, 1716,1718, 1724, 1761, 1781, 1797, 1814, 1816, 1831, 1843, 1846, 1848, 1857, 1859, 1860, 1869, 1892],

[2, 3, 16, 18, 24, 25, 96, 109, 133, 177, 182, 195, 210, 218, 226, 230, 244, 260, 276, 326, 361, 418, 453, 461, 467, 489, 497, 507, 512, 536, 544, 579, 588, 598, 617, 671, 680, 721, 739, 793, 808, 871, 931, 934, 964, 965, 990, 1005, 1023, 1059, 1060, 1063, 1067, 1076, 1154, 1164, 1166, 1205, 1214, 1219, 1232, 1253, 1260, 1261, 1274, 1284, 1292, 1299, 1336, 1344, 1352, 1377, 1404, 1418, 1431, 1445, 1451, 1460, 1486, 1489, 1501, 1556, 1557, 1590, 1644, 1675, 1685, 1744, 1774, 1786, 1788, 1802, 1807, 1811, 1827, 1834, 1835, 1876, 1886, 1894]]

maxIntersection = 0

for cluster in clusteringResults:

finalCount = 0

for oracleCluster in crossDatasetTrueClassification:

tempcount = len(set(oracleCluster).intersection(set(cluster)))

if finalCount < tempcount:

finalCount = tempcount

maxIntersection += finalCount

print maxIntersection/19.0

fileContainer.write('\n')

fileContainer.write('The clustering Jaccard Similarity is : '+ str(maxIntersection/19.0))

# Jaccard Similarity

def jaccardSim(groundTruth, predictionResult):

return (jaccard/19.0)*100

# Correct Clustering Rate CCR

def ccr(groundTruth, predictionResult):

return (maxIntersection/1900.0)*100

print jaccardSim(crossDatasetTrueClassification, clusteringResults)

print ccr(crossDatasetTrueClassification, clusteringResults)

# print labelAssignment

fileContainer.write('\n')

fileContainer.write('The Correct Clustering Rate is : '+ str(ccr(crossDatasetTrueClassification, clusteringResults)))

fileContainer.write('\n')

fileContainer.write('The clustering Jaccard Similarity is : '+ str(jaccardSim(crossDatasetTrueClassification, clusteringResults)))

wholeTrueClasses = np.zeros(1900)

for i, trueClass in enumerate(crossDatasetTrueClassification):

for trajectory in trueClass:

wholeTrueClasses[trueClass] = i

print list(wholeTrueClasses[:20])

wholePredClasses = np.zeros(1900)

for i, predClass in enumerate(clusteringResults):

for trajectory in predClass:

wholePredClasses[predClass] = i

print list(wholePredClasses[:20])

print normalized_mutual_info_score(wholeTrueClasses, wholePredClasses)

fileContainer.write('\nThe NMI is : '+ str(normalized_mutual_info_score(wholeTrueClasses, wholePredClasses)))

fileContainer.write('\n--------------------------------------------------------------------')

print measurements.ccr(crossDatasetTrueClassification, clusteringResults)

print measurements.jaccardSim(crossDatasetTrueClassification, clusteringResults)

print measurements.NMI(crossDatasetTrueClassification, clusteringResults)