Tidysq Versions Save

Fixed:

• early return now works correctly for remove_ambiguous()
• ensured correct order of operations during sequence packing
• changed some values to unsigned wherever there was a mismatch

Updates for the CRAN god. Fixed random_sq() out-of-bounds possible problem, so there may be differences in sequences generated using the same seed in this and previous versions.

Fixed:

• upper bound for random_sq() correctly ignores now "*" and "-" values while computing number of letters to draw from
• removed default move/copy constructors/assignment operators for Sequence and ProtoSequence classes
• lots of technical files cleanup

Expanded on v1.0.0, having implemented functions paste() and collapse() that allow the user to connect multiple sequences into one. Also made some optimization within C++ using templates, speeding up translate() and complement() functions significantly. Lastly, added support for object from bioseq package.

Breaking changes:

• dropped argument interpret_as_stop from translate() function, as it is not feasible to implement well-working translation rules for tables with ambiguous codons (27, 28 & 31)

New features:

• implemented paste() (a method for sq class)
• implemented collapse()
• added support for classes from bioseq package, i.e. bioseq_aa, bioseq_dna and bioseq_rna

Improved:

• remade translate() to have codon tables created in compile time; this reduced execution time of translate() by 95%
• remade complement() to have tables created in compile time; this reduced execution time of complement() by 85%

Fixed:

• made random_sq() actually use seed parameter while generating sequences

Not much different to the previous, v0.5.0 release. Features a stable API with working C++ Operation interface.

From now on all future releases will be named according to semantic versioning practices. API changes and bugfixes will be listed as well.

Most of the intended functionalities are there, we don't expect API to change much if at all. Still working on rewriting crucial (and not-so-crucial) code in C++ and better test coverage (though API is already quite well covered).