search

dkoslicki / CMash

Fast and accurate set similarity estimation via containment min hash
BSD 3-Clause "New" or "Revised" License
42 stars 9 forks source link

Multiple k-mer sizes bug #19

Closed dkoslicki closed 4 years ago

dkoslicki commented 4 years ago

Current implementation of multiple k-mer size containment index computation may have some bugs in it as noted here. In particular, training with 30-60-10 and training with 60-60-10 results in different 60-mer containment index values (which it shouldn't).

dkoslicki commented 4 years ago

This is addressed in c8c91f4d3c7437ac0a9bbda7faf8c7fcb528e488, but leaving open until we get a more complete testing environment open as per #14 as only checked with local testing.

dkoslicki commented 4 years ago

Closing as some testing is done, and more testing is being implemented in #14