def testFalseNegativeNeverHappens(self):
n = 10
inserted = []
bf = BloomFilter(n, n)
for i in range(n):
bf.insert(str(i))
inserted.append(i)
for j in inserted:
self.assertTrue(bf.query(j))
def testFalseNegativeNeverHappens(self):
n = 10
inserted = []
bf = BloomFilter(n, n)
for i in range(n):
bf.insert(str(i))
inserted.append(i)
for j in inserted:
self.assertTrue(bf.query(j))
def testInputIsFalsePositiveUntilInserted(self):
n = 100
bf = BloomFilter(n / 2, n)
fp = [False] * n
inserted = [False] * n
fp_cnt = 0
for i in range(n):
bf.insert(str(i))
inserted[i] = True
for j in range(n):
if inserted[j]:
continue
# It was false positive before, it must continue
# to be (since it was not inserted).
if fp[j]:
self.assertTrue(bf.query(j))
# Update false positives list
elif bf.query(j):
fp[j] = True
fp_cnt += 1
# We're inserting more elements than the size of the array, so
# there must be false positives.
self.assertTrue(fp_cnt > 0)
def testInputIsFalsePositiveUntilInserted(self):
n = 100
bf = BloomFilter(n/2, n)
fp = [False]*n
inserted = [False]*n
fp_cnt = 0
for i in range(n):
bf.insert(str(i))
inserted[i] = True
for j in range(n):
if inserted[j]:
continue
# It was false positive before, it must continue
# to be (since it was not inserted).
if fp[j]:
self.assertTrue(bf.query(j))
# Update false positives list
elif bf.query(j):
fp[j] = True
fp_cnt += 1
# We're inserting more elements than the size of the array, so
# there must be false positives.
self.assertTrue(fp_cnt > 0)
def bloom_filter_run(n, m, k=None):
keys = range(n)
# Random sampling without replacement
random.shuffle(keys)
probs = []
filter = BloomFilter(m, n, k)
inserted = [False] * (n)
for cnt, entry in enumerate(keys):
filter.insert(str(entry))
inserted[entry] = True
false_positives, total = 0, 0
# Compute false positives
for probe in range(n):
if not inserted[probe]:
exists = filter.query(probe)
if exists:
false_positives += 1
total += 1
if total != 0:
prob = false_positives * 1.0 / total
probs.append(prob)
return probs
def bloom_filter_run(n, m, k=None):
keys = range(n)
# Random sampling without replacement
random.shuffle(keys)
probs = []
filter = BloomFilter(m, n, k)
inserted = [False] * (n)
for cnt, entry in enumerate(keys):
filter.insert(str(entry))
inserted[entry] = True
false_positives, total = 0, 0
# Compute false positives
for probe in range(n):
if not inserted[probe]:
exists = filter.query(probe)
if exists:
false_positives += 1
total += 1
if total != 0:
prob = false_positives * 1.0 / total
probs.append(prob)
return probs
