v3.0.0a4

Alpha release

• Experimental features: generative models, clustering, dimensionality reduction
• Integrated LIgO for simulation (LigoSim instruction)
• Internally rely on bionumpy for data storage
• Major documentation updates with focus on new developers (new docs for integrating new components, minor updates to MLMethod and DatasetEncoder interfaces, scripts for automatically testing new components)
• Added functionalities for analysing antigen-binding associated motifs

v2.2.5

What's Changed

• some new reports
• bug fixes in importers, encoders, and data model
• CompAIRR helper updates

v2.2.4

What's new

• update and reorganize requirements
• documentation updates

v2.2.3

What's Changed

• Make CompAIRR encoder more RAM efficient
• Bugfix element generator make subset
• Speed up CompAIRRSequenceAbundance encoder
• Caching speed improvement MatchedSequences/MatchedReceptors
• Bugfix: consider positive_class when making binary mapping
• Support custom import of Olga columns
• Set up LOOCV for sequence datasets
• Add manual splitting for sequence and receptor datasets
• Label mapping bugfixes

Full Changelog: v2.2.2...v2.2.3

v2.2.2

What's Changed

• minor updates related to versions of dependencies

Full Changelog: v2.2.1...v2.2.2

v2.2.1

What's Changed

• updated documentation
• minor bug fixes in ROC curve reports
• minor bug fixes in class mappings

Full Changelog: v2.2.0...v2.2.1

v2.2.0

What's Changed

• Windows-specific bug fixes
• Kmer abundance encoder
• Bug fixes: CompAIRRSequenceAbundance encoder memory, other minor fixes

Full Changelog: v2.1.2...v2.2.0

v2.1.2

What's Changed

• added support for sequence dataset in word2vec encoder
• updated reference encoders with different summarizing strategies
• minor bug fixes

Full Changelog: v2.1.1...v2.1.2

v2.1.1

Minor updates:

• add keep_fraction option to FeatureComparison report
• fix optimization metric in sklearn
• copy reports to HTML output for ExploratoryAnalysis instruction
• fix encodings based on nucleotide sequences
• make the standard dataset name in Galaxy Tools "dataset"
• try to fix dependency issues (minor updates added to FAQ)

v2.1.0

Major change: integrate immuneML with CompAIRR (https://github.com/uio-bmi/compairr):

• add CompAIRRSequenceAbundance and CompAIRRDistance encoder,
• update docker to include CompAIRR,
• minor bug fixes.