Python implementation of the Hyper LogLog and Sliding Hyper LogLog cardinality counter algorithms.

Installation:

Use pip install hyperloglog to install from PyPI.

Usage:

import hyperloglog
hll = hyperloglog.HyperLogLog(0.01)  # accept 1% counting error
hll.add("hello")
print(len(hll))  # 1
hll.add("hello")
print(len(hll))  # 1 as items aren't added more than once
hll.add("hello again")
print(len(hll))  # 2

If we add a further 1000 random strings (giving a total of 1002 strings) we'll have a count roughly within 1% of the true value, in this case it counts 1007 (within +/- 10.2 of the true value)

# add 1000 random 30 char strings to hll
import random
import string
[hll.add("".join([string.ascii_letters[random.randint(0, len(string.ascii_letters)-1)] for n in range(30)])) for m in range(1000)]
print(len(hll))  # 1007

Changes:

• Added Sliding window HLL version
• Added bias correction from HLL++

References:

1. http://algo.inria.fr/flajolet/Publications/FlFuGaMe07.pdf
2. http://hal.archives-ouvertes.fr/docs/00/46/53/13/PDF/sliding_HyperLogLog.pdf
3. http://research.google.com/pubs/pub40671.html