def __init__(self, config, genHashes):
"""
Args:
config (dictionary):contains configuration data for this hashing object
genHashes (boolean) : whether or not to generate new hash seeds/coefficients
(Task 2 provides hash seeds/coefficients,
Tasks 1 and 3 require you to make them yourself)
"""
self.k = config['k']
self.n = config['n']
self.N = config['N']
self.Prime = util.findNextPrime(self.n)
self.NPrime = util.findNextPrime(self.N)
# generate new random hashes, or not if task2
if genHashes:
# set random seed to be generated seed at config load
# random.seed(config['genSeed'])
# build lists of coefficients self.a and self.b
self.a = []
self.b = []
for i in range(0, self.k):
self.a.append(random.randint(1, self.n))
self.b.append(random.randint(0, self.n))
# if not gen (task 2), then use hashes that are in config dictionary
else:
self.a = config['a']
self.b = config['b']
def __init__(self, config):
"""
Args:
config (dictionary): Configuration of this bloom filter
config['N'] (int) universe size
config['m'] (int) number of elements to add to filter
config['n'] (int) number of bits in bloom filter storage array
config['k'] (int) number of hash functions to use
config['task'] (int) the task this bloom filter is to perform (1,2,3)
config['type'] (int) type of hash function (1, 2, -1==unknown type)
if type 1 hash :
config['seeds'] (list of k ints) : seed values for k hash functions for type 1 hash function
if type 2 hash :
config['a'] (list of k ints) : a coefficients for k hash functions for type 2 hash function
config['b'] (list of k ints) : b coefficients for k hash functions for type 2 hash function
genHashes (boolean) : whether or not to generate new hash seeds/coefficients
(Task 2 provides hash seeds/coefficients,
Tasks 1 and 3 require you to make them yourself)
"""
# task this boom filter is performing
self.task = config['task']
# task 1 and 3 reguire generated seeds for hashes, task 2 uses provided seeds/coefficients
genHashes = (self.task != 3)
# type of hash for this bloom filter
self.type = config['type']
# check if n is prime
n = int(config['n'])
if not util.checkIfPrime(n):
config['n'] = util.findNextPrime(n + 1)
else:
config['n'] = n
P = int(config['N'])
if not util.checkIfPrime(P):
config['P'] = util.findNextPrime(P + 1)
else:
config['P'] = P
if (self.type == 1):
self.hashFunc = HashType1(config, genHashes)
elif (self.type == 2):
self.hashFunc = HashType2(config, genHashes)
# elif(self.type == 3): #add your own hashes
else:
print('BloomFilter for task ' + str(self.task) +
' ctor : Unknown Hash type : ' + str(self.type))
bf_size = self.hashFunc.n
bf_arr = []
for i in range(0, bf_size):
bf_arr.append(0)
self.bf_arr = bf_arr
def __init__(self, config, genHashes, c):
"""
Args:
config (dictionary):contains configuration data for this hashing object
genHashes (boolean) : whether or not to generate new hash seeds/coefficients
(Task 2 provides hash seeds/coefficients,
Tasks 1 and 3 require you to make them yourself)
"""
self.k = config['k']
self.N = config['N']
self.P = util.findNextPrime(self.N)
if config['task'] == 3:
self.n = c * config['m']
else:
self.n = config['n']
#generate new random hashes, or not if task2
if genHashes:
#set random seed to be generated seed at config load
random.seed(config['genSeed'])
#build lists of coefficients self.a and self.b
self.a = random.randint(low=0, high=(self.n - 1), size=self.k)
self.b = random.randint(low=0, high=(self.n - 1), size=self.k)
#if not gen (task 2), then use hashes that are in config dictionary
else:
self.a = config['a']
self.b = config['b']
def __init__(self, config, genHashes):
self.k = config['k']
self.n = config['n']
self.N = config['N']
self.prime = util.findNextPrime(self.n)
self.NPrime = util.findNextPrime(self.N)
#generate new random hashes, or not if task2
if genHashes:
#set random seed to be generated seed at config load
# random.seed(config['genSeed'])
#build lists of coefficients self.a and self.b
a = []
b = []
for i in range(0, self.k):
a.append(random.randint(1, self.n - 1))
b.append(random.randint(0, self.n - 1))
self.a = a
self.b = b
#if not gen (task 2), then use hashes that are in config dictionary
else:
self.a = config['a']
self.b = config['b']
def computeFalsePositive(configData, numTrials, k, c, hashType):
configData['k'] = k
configData['type'] = hashType
configData['n'] = util.findNextPrime(c * configData['m'])
sumFalsePositive = 0
for i in range(0, numTrials):
# initialize bloom filter
bf = BloomFilter(configData)
bfInputData = util.readIntFileDat(configData['inFileName'])
# add data to bloom filter
for j in range(0, configData['m']):
bf.add(bfInputData[j])
falsePositive = 0
# test false positive
for l in range(configData['m'], len(bfInputData)):
if bf.contains(bfInputData[l]):
falsePositive += 1
sumFalsePositive += falsePositive / float(configData['m'])
avgFalsePositive = sumFalsePositive / numTrials
return avgFalsePositive