This is the first stable release. Bug fixes and performance enhancements will continue to be added to version 1.1+, but all API/UI changes will now occur in the develop branch, or the future V1.2
- Allow any AlignBuddy supported alignment method to be used
- Switch default alignment method to ClustalOmega, using the --pileup option on sequences sorted by length
- Monitoring program created to track health of workers and masters (monitor_dbs.py)
- Much more stable Workers
- Create program to reset worker database when the occasional race condition breaks something (reset_workers.py)
- Split queued jobs into sub-jobs
- Ensure that TrimAl doesn't reduce sequences too much
- Memory upgrades by passing file paths instead of large objects
- No longer collapse clusters after the first
- Major refactoring of classes/functions and improved unit test coverage
- Orphan placement has been replaced with a more general placement.
- All sequences are converted to HMMs and an all-by-all correlation matrix is used to determine which cluster every sequence belongs to at the end of the run.
- The diminishing returns base is now determined on the fly, relative to cluster composition
- New program to pull together cluster sequences (group_by_cluster.py)
- Fully support MCMCMC resume, creating dumpfiles on the fly so a job can be picked back up in the event of a crash.
- Many, many, many bug fixes and general improvements