Try using Quotient Filters

https://github.com/vedantk/quotient-filter

https://blog.acolyer.org/2017/08/08/a-general-purpose-counting-filter-making-every-bit-count/

Much more efficient than Bloom

The issue names no FiloDB file, test, or entry point; begin by reading the linked quotient-filter implementation and article, then locate how Bloom filters are used in FiloDB. Define the integration scope and a benchmark or test for the claimed efficiency before coding, since completion criteria are not specified.

Written by the indexing model from the issue text.