-
Notifications
You must be signed in to change notification settings - Fork 0
/
construct_graphs.py
348 lines (302 loc) · 13.3 KB
/
construct_graphs.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
import itertools
import logging
import math
import numpy as np
import os
import pandas
from pyspark import SparkContext
from pyspark.mllib.linalg import Matrices, SparseMatrix, DenseMatrix
from pyspark.mllib.linalg.distributed import BlockMatrix
from pyspark.sql import SQLContext
from scipy import sparse
from scipy.io import loadmat
import scipy.io as sio
from pyspark.context import SparkContext
GENERATE_SPARSE = False
election_data_path = 'data/'
elecion_files = ['common-ids.csv', 'part1.csv', 'part2.csv']
def edgeDefinitionElection(vx, vy):
# vx = vx / max(1.0, 1.0 * max(abs(vx)))
# vy = vy / max(1.0, 1.0 * max(abs(vy)))
# return np.sum(np.minimum(vx, vy))
similarity = 0
for i in range(len(vx)):
x, y = vx[i], vy[i]
if (x <= 1e3 and y <= 1e3) or \
((1e3 < x <= 1e4) and (1e3 < y <= 1e4)) or \
(x > 1e4 and y > 1e4):
similarity += math.log(max(1, min(x, y)))
return similarity
def createAdjMat(graphNodes, year, sparseG, blockSize, sc, data = 'election'):
if data == 'toy' or graphNodes == 18:
return createAdjMatToy(graphNodes, year, sparseG, blockSize, sc)
if graphNodes == 70e3:
return createSmallAdjMatClimate(graphNodes, year, sparseG, blockSize, sc)
if graphNodes == 256e3:
return createAdjMatClimate(graphNodes, year, sparseG, blockSize, sc)
if data == 'election' or graphNodes > 10e3:
return createAdjMatElection(graphNodes, year, sparseG, blockSize, sc)
logging.warn('createAdjMat: Invalid input [data]')
return None
def createAdjMatToy(graphNodes, year, sparseG, blockSize, sc):
index = (year > 12) + 1 # 12 maps to 1, 16 maps to 2
path = 'BASE_PATH/toy_example/'
A = loadmat(path + 'toy_A' + str(index) + '.mat')['G']
n = A.shape[0]
p = n / 2
subMatrices = [A[:p,:p], A[:p,p:], A[p:,:p], A[p:,p:]]
ids, blocks = [(0,0), (0,1), (1,0), (1,1)], []
for i, id in enumerate(ids):
adj = subMatrices[i]
G = Matrices.dense(p,p, adj.transpose().flatten())
blocks.append((id, G))
blocksRdd = sc.parallelize(blocks, len(ids))
return blocksRdd
def createAdjMatClimate(graphNodes, year, sparseG, blockSize, sc):
# 1994 Jan vs 1995 Jan
id = (year > 12) + 4 # 12 maps to 4, 16 maps to 5
dataFile = 'BASE_PATH/climate_precip/preDataVectors.mat' # precipitation data
dataVec = loadmat(dataFile, squeeze_me = True)['pre199' + str(id) + '01']
numBlocks = int(graphNodes / blockSize)
dataBlocks = [(i/blockSize, dataVec[i:i+blockSize]) for i in range(0, graphNodes, blockSize)]
dataBlocksRdd = sc.parallelize(dataBlocks, numBlocks)
allPairBlocks = dataBlocksRdd.cartesian(dataBlocksRdd)
def radialBasisBlock(pairData):
I, J = int(pairData[0][0]), int(pairData[1][0])
dataI, dataJ = pairData[0][1], pairData[1][1]
n = len(dataI)
allCombinations = itertools.product(dataI, dataJ)
allCombsEdges = [radialBasisKernel(p[0], p[1]) for p in allCombinations]
print 'allCombsEdges ', len(allCombsEdges), (n*n)
if len(allCombsEdges) == (n*n):
adj = np.reshape(allCombsEdges, (n,n))
else:
adj = np.zeros((n,n))
if I==J:
adj[range(n), range(n)] = 0
G = Matrices.dense(n,n, adj.transpose().flatten())
return ((I,J), G)
adjMatBlocks = allPairBlocks.map(radialBasisBlock).cache()
return adjMatBlocks
def radialBasisKernel(x, y, sigma = 388.0):
delta = (x - y) ** 2
return math.exp(-delta/(2.0 * sigma ** 2))
def createSmallAdjMatClimate(graphNodes, year, sparseG, block_size, sc):
index = (year > 12) + 13 # 12 maps to 13, 16 maps to 14
prefix = 'BASE_PATH/climate_precip/GJan'
data_file = prefix + str(index) + 'padded.mat'
adj_mat = loadmat(data_file)['G']
n = adj_mat.shape[0]
if not n == graphNodes:
logging.warn('Error in size')
num_blocks, blocks = n / block_size, []
for i in range(num_blocks):
start_row, end_row = i * block_size, (i+1) * block_size
for j in range(num_blocks):
start_col, end_col = j * block_size, (j+1) * block_size
if end_row > n or end_col > n:
print '\n\n\n Error: createSmallAdjMatClimate \n\n\n'
G = adj_mat[start_row:end_row, start_col:end_col]
blocks.append(((i,j), G))
def denseBlock(x):
adj = x[1].toarray()
G_dm = Matrices.dense(block_size, block_size, adj.transpose().flatten())
return (x[0], G_dm)
# blocksRdd = allBlocksIds(graphNodes, blockSize, sc).map(denseBlock)
blocksRdd = sc.parallelize(blocks, num_blocks **2).map(denseBlock)
return blocksRdd
def allBlocksIds(N, p, sc):
num_blocks = N/p
i = range(num_blocks)
allKeys = [x for x in itertools.product(i, i)]
return sc.parallelize(allKeys, num_blocks ** 2)
def createAdjMatElection(graphNodes, year, sparseG, blockSize, sc):
if sparseG:
GENERATE_SPARSE = True
normalizeDonations = False
path = election_data_path
if (year == 12):
donations = loadContributionsCSV(path, elecion_files[1])
elif (year == 16):
donations = loadContributionsCSV(path, elecion_files[2])
# Fix to overflow errors due to large Aij
donations[4,:] = 0.1 * donations[4,:]
donations[6,:] = 0.1 * donations[6,:]
else:
print 'ERROR: Wrong value of year (', year, '), only 12 and 16 allowed'
return
if graphNodes < len(donations):
if False:
d1 = loadContributionsCSV(path, elecion_files[1])
d2 = loadContributionsCSV(path, elecion_files[2])
totalD = np.sum(d1 + d2, 1)
topDonors = totalD.argsort()[-graphNodes:]
donations = donations[topDonors,:]
else:
donations = donations[0:graphNodes,:]
if graphNodes > len(donations):
extraNodes = graphNodes - len(donations)
donations = np.append(donations, np.zeros((extraNodes, donations.shape[1])), axis = 0)
blocks, n = splitInBlocks(donations, blockSize)
donationsRdd = sc.parallelize(blocks, n)
#--------------------------------------
n = donationsRdd.count()
logging.warn('donationsRdd count = ' + str(n) + ', parts = ' + \
str(donationsRdd.getNumPartitions()))
donationsRdd.repartition(n).cache()
a = donationsRdd.take(1)
sqlContext = SQLContext(sc)
#--------------------------------------
logging.warn('donationsRdd parts = ' + str(donationsRdd.getNumPartitions()))
allPairDonations = donationsRdd.cartesian(donationsRdd)
logging.warn('allPairDonations count = ' + str(allPairDonations.count()))
adjMatBlocks = allPairDonations.map(constructElectionBlock)
if normalizeDonations:
logging.warn('Before normalizeDonations')
adjMatBlocks = nomalizeBlocks(adjMatBlocks)
logging.warn('Done normalizeDonations')
return adjMatBlocks
logging.warn('Calling BlockMatrix(), size = ' + str(N))
adjMat = BlockMatrix(adjMatBlocks, blockSize, blockSize, N, N)
return adjMat
def constructElectionBlock(pairDonations):
I = int(pairDonations[0][0])
J = int(pairDonations[1][0])
donationsI = pairDonations[0][1]
donationsJ = pairDonations[1][1]
n = donationsI.shape[0]
allCombinations = itertools.product(donationsI, donationsJ)
allCombsEdges = [edgeDefinitionElection(p[0], p[1]) for p in allCombinations]
if len(allCombsEdges) == (n*n):
adj = np.reshape(allCombsEdges, (n,n))
else:
adj = np.zeros((n,n))
if I==J:
adj[range(n), range(n)] = 0
if GENERATE_SPARSE:
G = sparse.csc_matrix(adj)
subMatrixSparse = Matrices.sparse(n, n, G.indptr, G.indices, G.data)
return ((I,J), subMatrixSparse)
else:
G = Matrices.dense(n,n, adj.transpose().flatten())
return ((I,J), G)
def splitInBlocks(donations, blockSize):
N, d = donations.shape
numBlocks = int(math.ceil(1.0 * N / blockSize))
delta = numBlocks * blockSize - N
donations = np.vstack((donations, np.zeros((delta, d))))
N = donations.shape[0]
m = N/numBlocks
listOfDonationBlocks = [(i/m, donations[i:i+m,:]) for i in range(0, N, m)]
return listOfDonationBlocks, numBlocks
def createElectionDiffGraph(blockSize, sc):
A1 = createAdjMat(12, blockSize, sc)
A2 = createAdjMat(16, blockSize, sc)
logging.warn('\n A1: ' + str(A1.numRows()) + ' x ' + str(A1.numCols()) + ' \n A2: ' + str(A2.numRows()) + ' x ' + str(A2.numCols()))
logging.warn('\n A1: ' + str(A1.rowsPerBlock) + ' x ' + str(A1.colsPerBlock) + ' \n A2: ' + str(A2.rowsPerBlock) + ' x ' + str(A2.colsPerBlock))
diffA = subtractSparseBlockMatFast(A1, A2, blockSize)
return diffA
def subtractSparseBlockMatFast(left, right, p):
n = right.numRows()
negBlocks = right.blocks.map(lambda x: elementWiseMultiply(x, -1))
negative_right = BlockMatrix(negBlocks, p, p, n, n)
result = negative_right.add(left)
return result
def elementWiseMultiply(block, M):
mat = block[1]
mat.values = 1.0*M*(mat.values)
return (block[0], mat)
def WriteEdgesPerBlock(b, SQUARE_BLOCK_SIZE):
I = b[0][0]
J = b[0][1]
if I > J:
return [] # ((I,J), 0)
if isinstance(b[1], SparseMatrix):
mat = sparse.csc_matrix((b[1].values, b[1].rowIndices, b[1].colPtrs), shape=(b[1].numRows, b[1].numCols))
links = sparse.find(mat)
edges = zip((links[0]+ (I * SQUARE_BLOCK_SIZE)), (links[1]+ (J * SQUARE_BLOCK_SIZE)),links[2])
else:
mat = np.array(b[1].toArray())
if I == J:
mat = np.triu(mat)
i,j = np.nonzero(mat)
values = mat[i,j]
i = i + I * SQUARE_BLOCK_SIZE
j = j + J * SQUARE_BLOCK_SIZE
edges = []
for ind in range(len(values)):
edges.append((i[ind], j[ind], values[ind]))
return edges
def nomalizeBlocks(adjMatBlocks):
blockMaxs = adjMatBlocks.map(blockMaxWithCheck)
matrixMax = blockMaxs.reduce(max)
normalizedBlocks = adjMatBlocks.map(lambda x: elementWiseDivide(x, matrixMax))
return normalizedBlocks
def blockMaxWithCheck(block):
if len(block[1].values) == 0:
return 0.0
else:
return max(block[1].values)
# does block/M
def elementWiseDivide(block, M):
mat = block[1]
mat.values = 1.0*(mat.values)/M
return (block[0], mat)
def loadContributionsCSV(path, filename):
names = pandas.read_csv(path + elecion_files[0], sep='|', skip_blank_lines = False)
nameIndices = {x:i for i,x in enumerate(names['NAME'])}
partyIndices = {'DEM':0, 'REP':1, 'OTHERS':2}
N = names.shape[0]
df = pandas.read_csv(path + filename, sep = '|', skip_blank_lines = False)
df = df.drop(df[df.DONATION <= 0].index)
df['NAME_I'] = df['NAME'].map(nameIndices)
df['PARTY_J'] = df['PARTY'].map(partyIndices)
df = df[df.NAME_I.notnull()]
df = df[df.PARTY_J.notnull()]
donations = np.array(sparse.csc_matrix((df['DONATION'], (df['NAME_I'], df['PARTY_J'])), shape = (N, 3)).toarray())
return donations
def mapIdToNames(path, filename):
names = pandas.read_csv(path + elecion_files[0], sep='|', skip_blank_lines = False)
nameIndices = {x:i for i,x in enumerate(names['NAME'])}
partyIndices = {'DEM':0, 'REP':1, 'OTHERS':2}
N = names.shape[0]
df = pandas.read_csv(path + filename, sep = '|', skip_blank_lines = False)
df = df.drop(df[df.DONATION <= 0].index)
df['NAME_I'] = df['NAME'].map(nameIndices)
df['PARTY_J'] = df['PARTY'].map(partyIndices)
df = df[df.NAME_I.notnull()]
df = df[df.PARTY_J.notnull()]
donations = np.array(sparse.csc_matrix((df['DONATION'], (df['NAME_I'], df['PARTY_J'])), shape = (N, 3)).toarray())
idToNames = {}
ids, names = list(df['NAME_I']), list(df['NAME'])
for i, id in enumerate(ids):
if id > 0 and id in idToNames and not idToNames[id] == names[i]:
logging.warn('Inconsistent mapping found: df[NAME_I] to df[NAME]: ' + str(id) + ': ' + str(idToNames[id]) + ', ' + str(names[i]))
idToNames[id] = names[i]
return idToNames, donations
def generateReport(nodeIdContributorsList):
path = election_data_path
idToNames, donations = mapIdToNames(path, elecion_files[1])
donations16 = loadContributionsCSV(path, elecion_files[2])
print '\n ========== \n ***** report ******* \n ======== \n'
for row in nodeIdContributorsList:
id, contributors = row[0], row[1:]
print 'Ids: ', id, contributors
print 'Names: ', idToNames.get(id), [idToNames.get(x) for x in contributors]
print '2012: ', donations[id], [donations[x] for x in contributors]
print '2016: ', donations16[id], [donations16[x] for x in contributors]
print '----------'
print '\n ========== \n ***** report END ******* \n ======== \n'
return
def electionEdgesReport(topAnomalyEdges):
path = election_data_path
idToNames, donations12 = mapIdToNames(path, elecion_files[1])
donations16 = loadContributionsCSV(path, elecion_files[2])
print '\n\n\n ', topAnomalyEdges
print '\n ========== \n ***** report ******* \n ======== \n'
for score, (i, j) in topAnomalyEdges:
print idToNames.get(i), ' -- ', idToNames.get(j), ' s = ', score
print '2012: ', donations12[i], donations12[j]
print '2016: ', donations16[i], donations16[j], '\n'
print '\n ========== \n ***** report END ******* \n ======== \n'
return