def test_params_for_capacity(self):
"""
Tests that the parameters that are generated for a given
capacity and probability are correct given known sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
bytes, k = pyBloom.params_for_capacity(1e6, 1e-4)
assert bytes - pyBloom.extra_buffer() == round(19170117 / 8.0)
assert k == 14 # Parameters uses the ceiling instead of rounding
def get_bloom(cls, bloom_data):
bloom_args = (100000, .01)
if bloom_data:
size, ideal_k = BloomFilter.params_for_capacity(*bloom_args)
bitmap = Bitmap(size)
bitmap.mmap = bloom_data
return BloomFilter(bitmap, ideal_k)
else:
bf = BloomFilter.for_capacity(*bloom_args)
logging.debug("get_bloom(): data size %dkB" %
round(len(bf.bitmap.mmap) / 1024))
return bf
def test_for_capacity(self):
"""
Tests that the for_capacity method makes a sane bloom filter
using parameters that are generated for a given
capacity and probability are correct given known sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
bf = pyBloom.for_capacity(1e6, 1e-4)
# Check the bitmap size
assert (len(bf.bitmap) / 8) - pyBloom.extra_buffer() == round(19170117 / 8.0)
# Check the k num
assert bf.k_num == 14 # Parameters uses the ceiling instead of rounding
def test_expected_prob(self):
"""
Tests that the expected probability of false positives
is correct given known-sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
assert round(pyBloom.expected_probability(19170117, 1e6), 4) == 1e-4
def test_required_bytes(self):
"""
Tests that the number of required bytes that the bloom filter
says it needs is the correct given some known-sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
assert round(pyBloom.required_bits(1e6, 1e-4) / 8.0) == round(19170117 / 8.0)
def test_small_bitmap(self):
"""
Tests initializing with a bitmap that is too small
(e.g. less than or equal to the extra_buffer() size)
"""
with pytest.raises(ValueError):
pyBloom(Bitmap(pyBloom.extra_buffer()), 3)
def test_check_int(self):
"""
Tests checking for an int in a set. This should fail.
"""
bf = pyBloom.for_capacity(1000, 1e-4)
with pytest.raises(TypeError):
1234 in bf
def test_check_none(self):
"""
Tests checking None in a set. This should fail.
"""
bf = pyBloom.for_capacity(1000, 1e-4)
with pytest.raises(TypeError):
None in bf
def test_add_int(self):
"""
Tests adding an int to a set. This should fail.
"""
bf = pyBloom.for_capacity(1000, 1e-4)
with pytest.raises(TypeError):
bf.add(1234)
def test_add_none(self):
"""
Tests adding None to a set. This should fail.
"""
bf = pyBloom.for_capacity(1000, 1e-4)
with pytest.raises(TypeError):
bf.add(None)
def test_length(self):
"""
Tests that length works
"""
bf = pyBloom.for_capacity(1000, 1e-4)
assert len(bf) == 0
[bf.add("test%d" % x) for x in xrange(1000)]
assert len(bf) == 1000
def test_prob(self):
"""
Tests that the bloom filter is only wrong within
a certain threshold.
"""
# Only wrong once per hundred
bf = pyBloom.for_capacity(1000, 0.01)
res = [bf.add("test%d" % x, True) for x in xrange(1000)]
num_wrong = len([x for x in res if x is False])
# Should get about 10 wrong
assert num_wrong >= 5
assert num_wrong <= 15
def test_add_without_check(self):
"""
Tests that adding to a bloom filter while checking
for existing entries works
"""
bf = pyBloom.for_capacity(1000, 1e-4)
# Assert all adds work
assert all([bf.add("test%d" % x, False) for x in xrange(1000)])
assert all([bf.__contains__("test%d" % x) for x in xrange(1000)])
assert len(bf) == 1000
# Assert all adds work
assert all([bf.add("test%d" % x, False) for x in xrange(1000)])
assert len(bf) == 2000
def test_swap_2(self):
"""
Swaps the mmap files from one implementation to another,
check that things work. Start with pyBloom, then cBloom.
"""
bytes, k = pyBloom.params_for_capacity(2e4, 1e-3)
bitmap = Bitmap(bytes, "testswap2.mmap")
bf1 = pyBloom(bitmap, k)
[bf1.add("foo%d" % x) for x in xrange(20000)]
bf1.close()
# Make a new bitmap
bitmap = Bitmap(bytes, "testswap2.mmap")
bf2 = cBloom(bitmap, 50)
assert len(bf2) == 20000 # Should reload size and k
assert bf2.k_num == k
# Check all the entries
assert all([bf2.__contains__("foo%d" % x) for x in xrange(20000)])
bf2.close()
def test_ideal_k(self):
"""
Tests that the ideal K is correct given known-sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
assert round(pyBloom.ideal_k(19170117, 1e6)) == 13
def test_expected_capacity(self):
"""
Tests that the expected capacity is correct given known-sane values.
"""
# From http://hur.st/bloomfilter?n=1e6&p=1e-4
assert round(pyBloom.expected_capacity(19170117, 1e-4)) == 1e6