How to use this

Testing the Bloom Filter

• Implementation Details are in bloom_filter.py
• Test functions are in main.py

To view the empirical analysis of ONLY bloom filters, comment lines 108-111 in main.py and run interpreter main on main.py To view the empirical analysis of Bloom vs. Cuckoo filters, run interpreter main on main.py

To adjust the hash functions used for Bloom filters, define hash_fxns in main.py/analyze_bloom_filter() as some set of hash functions or lambdas. Then, call bloomFilter = bloom_filter.BloomFilter(m, k, hash_fxns) to create a new bloomFilter. Be sure to adjust m and k based on the number you are planning to insert.

To insert without our prebuilt experiments, bloomFilter.insert_with_time_elapsed To check without our prebuilt experiments, bloomFilter.check_with_false_prob_and_time_elapsed

Adjust inserting_num = 3800 #2**11 num_random_checks = 500 for prebuilt goals. Be sure to adjust m and k as well based on k = (m/n)*ln(2).

goal_1_theoretial_false_pos_rate tests theoretical rate of insertion on a bloomFilter. goal_2_polynomial_hash tests empirical rate of polynomial hashing on bloomFilter. goal_3_polynomial_hash tests insertion speed on bloomFilter.

Testing the Cuckoo Filter

• Implementation is located in cuckoo_filter.py
• Test functions are in test_cuckoo_filter.py

To view the empirical analysis of ONLY cuckoo filters, run interpreter on test_cuckoo_filter.py.

To adjust the fingerprint size when testing the false positive rates, assign f to be your desired value in the main() function (recommended range of f is 4-20 bits).

Within both tests:

• to adjust the size of the cuckoo filter, set M to desired value (must be a power of 2).
• to adjust the number of elements inserted in the table, set n to desired value.