Histosketching Using Little Kmers
This is a complete re-implementation of HULK. The change log states the main changes:
sketch
subcommand:
smash
subcommand:
print
and distance
subcommands is available in the smash
subcommandMinor bug fixes and improvements:
This release bumps HULK to a more stable version. Here are a summary of the main changes:
swap uint64 encoding of k-mers to instead us ntHash (Go implementation)
replace delta+epsilon values in CMS with a soft memory limit for the CMS structure
use Jump hash adjusting/querying CMS counters
allow FASTA input
bug fixes (histosketch metadata, weighted jaccard similarity
First full release of HULK