def test_intersection_size(self):
fpr = 0.001
# False positive rate with small numbers is high, therefore let's test with bigger sets
bloom_one = BloomFilter(100000, fpr)
bloom_two = BloomFilter(100000, fpr)
listA = [str(random.getrandbits(14)) for i in range(71000)]
listB = [str(random.getrandbits(12)) for i in range(69000)]
for char in listA:
bloom_one.add(char)
for char in listB:
bloom_two.add(char)
merged_bloom = bloom_one.intersection(bloom_two)
bloom_one_count = bloom_one.count
bloom_two_count = bloom_two.count
listA_uniq_count = len(set(listA))
listB_uniq_count = len(set(listB))
merged_bloom_count = merged_bloom.count
listAB_uniq_count = len(set(listA).intersection(set(listB)))
assert bloom_one_count == listA_uniq_count
assert bloom_two_count == listB_uniq_count
# Intersection guarantees to have all elements of the intersection but the false positive rate might be slightly higher than that of the pure intersection:
assert (listAB_uniq_count * (1 - 2 * fpr) <= merged_bloom_count <= listAB_uniq_count * (1 + 2 * fpr))
def test_intersection(self):
bloom_one = BloomFilter(100, 0.001)
bloom_two = BloomFilter(100, 0.001)
chars = [chr(i) for i in range_fn(97, 123)]
for char in chars:
bloom_one.add(char)
for char in chars[:int(len(chars)/2)]:
bloom_two.add(char)
new_bloom = bloom_one.intersection(bloom_two)
for char in chars[:int(len(chars)/2)]:
self.assertTrue(char in new_bloom)
for char in chars[int(len(chars)/2):]:
self.assertTrue(char not in new_bloom)
def test_intersection(self):
bloom_one = BloomFilter(100, 0.001)
bloom_two = BloomFilter(100, 0.001)
chars = [chr(i) for i in range(97, 123)]
for char in chars:
bloom_one.add(char)
for char in chars[:int(len(chars) / 2)]:
bloom_two.add(char)
new_bloom = bloom_one.intersection(bloom_two)
for char in chars[:int(len(chars) / 2)]:
self.assertTrue(char in new_bloom)
for char in chars[int(len(chars) / 2):]:
self.assertTrue(char not in new_bloom)
def test_nstar_intersection_1(self):
bloom_one = BloomFilter(200, 0.001)
bloom_two = BloomFilter(200, 0.001)
chars = [chr(i) for i in range_fn(0, 200)]
for char in chars:
bloom_one.add(char)
for char in chars[:int(len(chars)/2)]:
bloom_two.add(char)
new_bloom = bloom_one.intersection(bloom_two)
self.assertTrue(bloom_one.nstar() > len(chars)-10 and bloom_one.nstar() < len(chars)+10)
self.assertTrue(bloom_two.nstar() > len(chars)/2-10 and bloom_two.nstar() < len(chars)/2+10)
self.assertTrue(new_bloom.nstar() > len(chars)/2-10 and new_bloom.nstar() < len(chars)/2+10)
def test_nstar_intersection_2(self):
bloom_one = BloomFilter(200, 0.001)
bloom_two = BloomFilter(200, 0.001)
chars = [chr(i) for i in range_fn(0, 200)]
for char in chars[int(len(chars)/2):]:
bloom_one.add(char)
for char in chars[:int(len(chars)/2)]:
bloom_two.add(char)
new_bloom = bloom_one.intersection(bloom_two)
self.assertTrue(bloom_one.nstar() > len(chars)/2-10 and bloom_one.nstar() < len(chars)/2+10)
self.assertTrue(bloom_two.nstar() > len(chars)/2-10 and bloom_two.nstar() < len(chars)/2+10)
#The nstar operator will fail on the intersection of the filters..
self.assertTrue(new_bloom.nstar() > 10)
self.assertTrue(bloom_one.nstar_intersection(bloom_two) < 10)
