All Versions
Latest Version
Avg Release Cycle
42 days
Latest Release
27 days ago

Changelog History
Page 1

  • v0.30.0

    November 14, 2019
    • Bayesian models now allow to access internals.
    • 🛠 Various small bug fixes.
  • v0.29.0

    September 27, 2019
    • Migrate error handling to the snafu crate (this is an API breaking change).
    • 🛠 Fix edge cases in pairwise alignment.
    • 🛠 Fix error in backward search if symbol isn't found.
  • v0.28.2

    July 05, 2019
  • v0.28.1

    June 28, 2019
    • 🛠 Fix select in RankSelect in cases where many superblocks have the same rank.
  • v0.28.0

    June 19, 2019
    • 👍 Myers bit-parallel pattern matching now supports arbitrarily long patterns via bit vectors (thanks to @markschl).
    • 📚 Minor documentation updates (thanks to @anders-was-here).
  • v0.27.0

    May 31, 2019
    • Implement sequence-read-trait for FASTQ records.
    • Cleanup dependencies.
  • v0.26.1

    May 10, 2019
    • Fix a bug in select_1 and select_0 that would lead to too large answers.
  • v0.26.0

    May 09, 2019
    • ➕ Added a trait system for computing Bayesian statistical models.
    • ➕ Added an implementation of MSA via partial order alignment.
    • 🐎 Performance improvements to FASTQ reader.
  • v0.25.0

    December 12, 2018
    • ➕ Added FQRead and FARead traits to FastaReader and FastqReader to be more flexible with input types. This allows to use readers on gzipped and on plain text input interchangeably.
    • ➕ Added an implementation of Bayes Factors and evidence scoring using the method of Kass and Raftery.
  • v0.24.0

    November 26, 2018
    • API overhaul to become more flexible when accepting text iterators. Now, anything that iterates over something can be borrowed as u8 is allowed.
    • FMIndex and FMDIndex now also allow plain owned versions of BWT, Less and Occ. This should greatly simplify their usage.
    • 🐎 PairHMM and LogProb implementation has seen extensive performance improvements. Among that, (a) the usage of a fast approximation of exp() as presented by Kopczynsi 2017, and (b) banding of the pairHMM matrix with a given maximum edit distance.
    • 👍 All IO records now support serde.