Improved String Alignment

Complete overhaul of the string alignment algorithms used within the codebase. Switched to unboxed, mutable vectors/matrices for pairwise alignments. Updated the Direct Optimization and Implied Alignment algorithms with the improvements presented in Efficient Implied Alignment. Some other quality of life, performance gains, and new features were implemented as well.

Notable changes include:

• Multiple efficiency improvements of dynamic character scoring
• Asymptotic improvement of dynamic character pre-order traversals
• Improved TCM representation.
• Performance tuned file parsers
• Saving & reloading "save states" rework to use binary serialization

See the changelog for a complete listing of new features.

Restructured Architecture

Sweeping restructuring and refactoring of the entire codebase. The build process has been restructured to be modularized into sub-libraries. Refactoring of I/O interaction, textual representations, metadata caching, and computational evaluation model. Added integration tests and benchmarking suite along with expanding existing unit test suites and a setting up continuous integration process to improve maintainability and quality assurance. Many quality of life, performance gains, and new features were implemented along with numerous defect corrections.

Notable changes include:

• Dramatically reduced memory pressure
• Multiple efficiency improvements of dynamic character scoring
• Many useful command line options
• Saving & reloading "save states"
• New node clustering options to Wagner build command
• Asymptotic improvement of Wagner build command

See the changelog for a complete listing of new features.

Stabalized Performance

Corrected runtime defects and applied performance optimizations to make performance acceptable.

Notable changes include:

• Removed space leaks across the FFI bindings
• Sharing rather than repeating work during evaluation
• Increased memoization of dynamic character states
• Minor performance tuning of parsers

Alpha Release - DARPA Reporting

This release is a very minimal viable product. Capable of naive phylogenetic network construction and scoring under parsimony. Parses Fasta, Fastc, Newick, Nexus, and TNT files as input. Reports a custom data rendering, DOT, XML formats.

Code documentation is present and mostly complete. User documentation is missing.

Not ready for general use for extended search on computing clusters.