dicts_sets = {}
with open(inputName) as fr:
for line in fr.readlines():
lines = line.strip('\n').split(",")
tuples = (lines[0], lines[1], lines[2], lines[3])
if lines[2] not in usertupledict:
usertupledict[lines[2]] = []
usertupledict[lines[2]].append(tuples)
for i in usertupledict.keys():
dicts_set[i] = clean(usertupledict[i])
for i in dicts_set.keys():
locationSet = set()
for j in dicts_set[i]:
locationSet.add(mygeohash.encode(float(j[1]), float(j[0]), 2))
dicts_sets[i] = locationSet
def noisyCount(sensitivety, epsilon):
beta = sensitivety / epsilon
u1 = np.random.random()
u2 = np.random.random()
if u1 <= 0.5:
n_value = -beta * np.log(1. - u2)
else:
n_value = beta * np.log(u2)
return n_value
def noisyCounts(sensitivety, epsilon):
def getL1(k,e):
N = 20
# inputName = "G://data//map1_14_40.txt"
inputName = "G://data//map1_14_"
inputName=inputName+str(k)+".txt"
usertupledict = {}
sensitivety = int(2*k)
epsilon = e
# max = 0
# j = ""
dicts_set = {}
dicts_sets = {}
with open(inputName) as fr:
for line in fr.readlines():
lines = line.strip('\n').split(",")
tuples = (lines[0], lines[1], lines[2], lines[3])
if lines[2] not in usertupledict:
usertupledict[lines[2]] = []
usertupledict[lines[2]].append(tuples)
for i in usertupledict.keys():
dicts_set[i] = clean(usertupledict[i])
for i in dicts_set.keys():
locationSet = set()
for j in dicts_set[i]:
locationSet.add(mygeohash.encode(float(j[1]), float(j[0]), 2))
dicts_sets[i] = locationSet
counts = {}
lists = []
for i in dicts_sets.values():
for j in i:
if j[0] in mygeohash.neighbors('d').values():
lists.append(j)
for i in lists:
if not i in counts:
counts[i] = 0
counts[i] = counts[i] + 1
sum = 0
for i in counts.values():
sum += i
zoneCounts = {}
for i in counts.values():
if not i in zoneCounts:
zoneCounts[i] = 0
zoneCounts[i] = zoneCounts[i] + 1
a = sorted(zoneCounts.items(), key=lambda x: x[0], reverse=False)
# max = a[len(a) - 1][0]
# b = {}
# for i in a:
# b[i[0]] = i[1]
# for i in range(1, max + 1):
# if i not in b:
# b[i] = 0
#
# c = []
# for i in b.keys():
# c.append((i, b[i]))
#
# p = []
# p.append(1)
# for i in range(1, N):
# tamp = ((max - 1) * i / N) + 1
# p.append(int(tamp))
# p.append(max)
#
# g = []
# for i in range(N):
# g.append(0)
#
# for i in c:
# for j in range(1, N + 1):
# if i[0] >= p[j - 1] and i[0] < p[j]:
# g[j - 1] += i[1]
# g[N - 1] += c[len(c) - 1][1]
#
# G = []
# for i in range(N):
# G.append((p[i + 1], g[i]))
x = []
y = []
y_noise = []
z = []
z_noise = []
for i in a:
x.append(i[0])
y.append(i[1])
tamp = 0
for j in range(len(y)):
tamp += y[j]
z.append(tamp)
y_noise = laplace_mech(y, sensitivety, epsilon)
z_noise = laplace_mech(z, sensitivety, epsilon)
for j in range(len(z_noise)):
if z_noise[j] < 0:
z_noise[j] = 0
for j in range(len(y_noise)):
if y_noise[j] < 0:
y_noise[j] = 0
return countL1(y,y_noise)
inputName = "G://data//map1_14_20.txt"
sum = 0
userset = set()
usercount = []
locationset = set()
locationset2 = set()
locationset3 = set()
locationset4 = set()
with open(inputName) as fr:
for line in fr.readlines():
sum += 1
lines = line.strip('\n').split(",")
locationset2.add(mygeohash.encode(float(lines[1]), float(lines[0]), 2))
locationset3.add(mygeohash.encode(float(lines[1]), float(lines[0]), 3))
locationset4.add(mygeohash.encode(float(lines[1]), float(lines[0]), 4))
userset.add(lines[2])
usercount.append(lines[2])
print(sum)
print(len(userset))
print(len(usercount))
print("------------------")
print(len(locationset2))
print(len(locationset3))
print(len(locationset4))