Last active
October 14, 2019 20:30
-
-
Save fabiocerqueira/975c2096b91dae4c9770607c862f1c2e to your computer and use it in GitHub Desktop.
Example BloomFilter implementation in Python
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from hashlib import md5 | |
import redis | |
class BloomFilterBaseBackend: | |
def initialize(self, size): | |
self.size = size | |
def validate(self, position): | |
if position < 0 or position >= self.size: | |
raise IndexError( | |
f"Invalid position {position}, min: 0, max: {self.size - 1}" | |
) | |
def set_bits(self, positions): | |
raise NotImplementedError() | |
def get_bit(self, positions): | |
raise NotImplementedError() | |
class BloomFilterSyncRedisBackend(BloomFilterBaseBackend): | |
def __init__(self, key, host="localhost", port=6379, db=0): | |
self.connection = redis.Redis(host, port, db) | |
self.key = key | |
def set_bits(self, positions): | |
pipe = self.connection.pipeline() | |
for p in positions: | |
self.validate(p) | |
pipe.setbit(self.key, p, 1) | |
pipe.execute() | |
def get_bits(self, positions): | |
pipe = self.connection.pipeline() | |
for p in positions: | |
self.validate(p) | |
pipe.getbit(self.key, p) | |
values = pipe.execute() | |
out = dict(zip(positions, values)) | |
return out | |
class BloomFilterMemoryBackend(BloomFilterBaseBackend): | |
def initialize(self, size): | |
super().initialize(size) | |
self.bitarray = 1 << self.size | |
def set_bits(self, positions): | |
for p in positions: | |
self.validate(p) | |
self.bitarray |= 1 << p | |
def get_bits(self, positions): | |
out = {} | |
for p in positions: | |
self.validate(p) | |
value = self.bitarray & (1 << p) | |
out[p] = 1 if value else 0 | |
return out | |
def python_hash(key): | |
return hash(key) | |
def md5_hash(key): | |
# md5 is not recommended, but it is just for testing... | |
return hash(md5(key.encode()).hexdigest()) | |
class BloomFilter: | |
def __init__(self, size=10000, cache=None, hash_functions=None): | |
self.size = size | |
self.count = 0 | |
if cache is None: | |
cache = BloomFilterMemoryBackend() | |
self.cache = cache | |
self.cache.initialize(size) | |
if hash_functions is None: | |
self.hash_functions = [python_hash, md5_hash] | |
def _get_positions(self, key): | |
return [hash_func(key) % self.size for hash_func in self.hash_functions] | |
def add(self, key): | |
if self.count >= self.size: | |
raise IndexError("BloomFilter is full") | |
positions = self._get_positions(key) | |
self.cache.set_bits(positions) | |
self.count += 1 | |
def check(self, key): | |
positions = self._get_positions(key) | |
return all(self.cache.get_bits(positions).values()) | |
def __contains__(self, key): | |
return self.check(key) | |
def __len__(self): | |
return self.count | |
if __name__ == "__main__": | |
print("BloomFilter in-memory") | |
in_memo_bf = BloomFilter() | |
in_memo_bf.add("babs") | |
print(f"babs in in_memo_bf = {'babs' in in_memo_bf}") | |
print(f"fabio in in_memo_bf = {'fabio' in in_memo_bf}") | |
print("-" * 50) | |
print("BloomFilter using sync redis with mykey") | |
bf = BloomFilter(cache=BloomFilterSyncRedisBackend("mykey")) | |
bf.add("fabio") | |
print(f"fabio in bf = {'fabio' in bf}") | |
bf.add("pedro") | |
print(f"pedro in bf = {'pedro' in bf}") | |
print(f"xuxu in bf = {'xuxu' in bf}") | |
print("-" * 50) | |
print("BloomFilter using sync redis with anotherkey") | |
bf2 = BloomFilter(cache=BloomFilterSyncRedisBackend("anotherkey")) | |
print(f"fabio in bf2 = {'fabio' in bf2}") | |
bf2.add("xuxu") | |
print(f"xuxu in bf2 = {'xuxu' in bf2}") |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment