What's new

Release notes

v2.1 - 2026-02-09

Major features

Improvements in rextract

• FASTA I/O support: Added high-performance support for FASTA files in rextract using a custom quick parser to avoid I/O bottlenecks.
• Unified sequence handling: Refactored rextract to handle both FASTQ and FASTA formats seamlessly.
• New command-line option: Added -a / --fasta flag to rextract for processing FASTA datasets.
• Enhanced score filtering: Added -z / --maxscore option to allow extracting sequences with classifier scores below a certain threshold, or within a range when used together with minscore.

v2.0 - 2026-01-04

Major features

HTML file size optimization

• Sparse/Delimited data format: Implemented a new optimized XML format (Version 2) that uses comma-delimited values and sparse representation to significantly reduce HTML file sizes
• Automatic format selection: Added intelligent algorithm that automatically chooses between sparse or delimited format based on data sparsity for optimal compression
• Backward compatibility: Maintained compatibility as previous XML format (Version 1) is still available

Python version requirements

• Minimum version increased: Raised minimum Python requirement from 3.6 to 3.12 for improved performance and security
• Modern python features: Leveraged new language features available in Python 3.12+ for better code quality and efficiency, full pylance-compliant
• Dependency updates: Updated all dependencies to versions compatible with Python 3.12+

Enhancements

Core functionality

• Default strain level analysis: Changed default behavior to include strain-level analysis (previously required experimental flag)
• Parallel processing control: Added new --threads argument for fine-grained control of parallel processing
• Improved data representation: Optimized internal data structures for better memory usage and processing speed
• Enhanced error handling: Added more robust error handling and user-friendly error messages
• Performance improvements: Various optimizations throughout the codebase for faster execution

New tools [experimental, under development]

• Regentrifuge (rgf): Added new tool for genomic functional analysis, expanding Recentrifuge capabilities beyond taxonomic analysis
• Gene Ontology support: Integrated comprehensive Gene Ontology (GO) analysis capabilities. Requires external files (equivalent to names.dmp and nodes.dmp of NCBI Taxonomy) where the GO graph has been reduced to a tree

Changes

Breaking changes

• Python version: Dropped support for Python versions below 3.12
• Dependency updates: Updated core dependencies to their latest versions
• Default taxonomic resolution: Strain level is now enabled by default instead of species level

Command line interface

• New --no-strain flag: Disable strain-level analysis (strain is now default, it was an experimental flag previously)
• New --threads argument: Control number of threads for parallel processing (0 = legacy behavior)

Bug fixes

• rextract: Resolved issues with parsing certain edge cases in classifier output files

Documentation

• Updated Wiki: Comprehensive updates to documentation and user guides
• Migration guide: Created detailed migration guide for users upgrading from previous versions

Development

• Modern build system: Migrated from setup.py and requirements.txt to pyproject.toml for modern Python packaging
• Code quality: Improved code quality with updated linting and type checking
• Testing: Enhanced test suite with additional test cases and improved coverage

---

v1.16.0

• Update after important NCBI Taxonomy changes: new rank realm and reintroduction of domain

v1.15.1

• Update after NCBI Taxonomy change and upgrade to 2025

v1.15.0

• Improve generic parser:
  • New automatic full directory option to process all the present files in the directory
  • Compatible with .gz and .bz2 compressed files
• Improve resilience of statistics module upon extreme cases such as empty samples

v1.14.1

• Solve problems with GH Actions so that CI/CD is working fine again.
• Improve rextract to deal with some simulated reads

v.14.0

• Addition of new tool refafilt for processing databases contained in —huge— fasta files (such as NCBI BLAST nt), so:
  • provide filters by minimum and/or maximum sequence length to separate those sequences,
  • correct issue when header of fasta sequence is multiplexed meaning such sequence is redundant in the database
• Correct bug and improve statistics and messages

v1.13.2

• Correct bug in scoring introduced in v1.13.0

v1.13.1

• Update after NCBI Taxonomy change and upgrade to 2024

v1.13.0

• Important reductions in the size of the generated HTML, now < 2/3 of old size
• LOGLENGTH scoring: switch from geometric to arithmetic averaging
• retest: use pandas.testing.assert_frame_equal for df comparisons
• Relocation of a few krona-related constants