Improve hash function in MinHash family

broadinstitute / catch

A package for designing compact and comprehensive capture probe sets.

MIT License

76 stars 16 forks source link

Improve hash function in MinHash family #45

Closed haydenm closed 2 years ago

haydenm commented 2 years ago

This PR makes several small changes designed to improve the behavior of the MinHash family's hash functions:

Changes their inner function, used for strings (k-mers), from Adler-32 to md5 [1527bec]
Updates the construction of sequence signatures to tolerate sequences that were previously considered too short (i.e., too few k-mers for the signature) [fc6e8a6]

coveralls commented 2 years ago

Coverage increased (+0.06%) to 94.386% when pulling fc6e8a6fe3047ebe3278d42ce375742fc32d84a2 on hm-better-hash into f22207d33fed9f172c5090573bc7cc1425576c04 on master.