search

dnbaker / dashing

Fast and accurate genomic distances using HyperLogLog
GNU General Public License v3.0
161 stars 11 forks source link

Dev #59

Closed dnbaker closed 3 years ago

dnbaker commented 3 years ago

Bug fixes

  1. Cardinality estimates for empty sketches default to 0.
  2. HyperMinHash bug fix for AVX512F without AVX512BW.
  3. Fix uncanonicalized encoding functions -- previously, there was a bug in these which caused it to sketch the indexes instead of the values.

New Features

  1. Add cardcmp and dashing card subcommand. This has the same usage/options as dist, but only outputs cardinality estimates. Use when you only want size estimates, not intersections.
  2. Faster RangeMinHash sketching -- previously, this was maintained in a binary tree set, but this has been replaced by a hash set + priority queue.
lgtm-com[bot] commented 3 years ago

This pull request fixes 1 alert when merging 252fb6081b327bda87048567a355e42f1bc183cb into 841be58ceff2e8907b65277989836a3af62db392 - view on LGTM.com

fixed alerts: