Compiles a dynamic library that adds an approx_count_distinct built on HyperLogLog.
This gives better performance than a traditional count(distinct column) at the cost of precision.
The error rate is set at 0.81% for no other reason than it being the default value used by Redis.
⚠️ NOTE: Use this at your own risk, I make no commitment to maintain or update this library.⚠️