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_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()