v1.1.0 - Bugfixes and dependency management

Bug fixes

• Fixed a crash when the region filtering BED files are empty (#9). This includes the region filtering plot, which
  failed when no positions were filtered.
• --include-ref is now passed to the workflow, the reference genome is included in the SNP matrix and phylogeny when
  this option is set.
• BED files with additional columns or header lines (track, browser, #) are now parsed correctly.
• The region filtering statistics are now calculated per contig (multi-contig references), and include the last
  position of the reference genome.
• PACU_map can now be executed multiple times in the same working directory.
• PACU_map: the TRIMMOMATIC_ADAPTER_DIR environment variable is now used to locate the Trimmomatic adapters (it
  takes precedence over the Conda installation directory).
• PACU_map (Galaxy): the renamed ONT input is now used, and invalid characters are always removed from input names.

Changes

• The stats/filtered_positions.tsv file in the working directory now includes a chr column, positions are 1-based.

Packaging

• PACU now requires Python 3.10 or newer (tested with Python 3.10, Gubbins is not available for newer Python versions
  on Bioconda).
• The dependencies are now consistent between PyPI and Bioconda (Snakemake >=7.32.4,<8, PuLP <2.8, setuptools <82).
  pytest is now an optional dependency (pip install pacu_snp[test]).
• requirements.txt now contains pinned versions (generated with pip-compile).
• Added a Pixi development environment (pixi.toml and pixi.lock). The environment.yml file is now generated from
  it (this also fixes a segmentation fault caused by incompatible bcftools/htslib versions).

Tests & documentation

• Added tests for BED parsing, region overlaps, the Trimmomatic adapter lookup, the report plots, re-running
  PACU_map and the --include-ref option.
• Updated the README (installation instructions, tested tool versions, usage and testing).

v1.0.0 - First major version

Long overdue version bump to v1.0.0!

Improvements

• Mapping: skipped creation of Bowtie2 / Minimap2 indices when they already exist.
• PACU: allowed use of reference FASTA containing multiple sequences when Gubbins is disabled (--skip-gubbins).
• General: improved handling of relative paths for input/output.
• Build: migrated project configuration from setup.py to pyproject.toml.

No major functional changes; this release primarily includes quality-of-life enhancements for CLI workflows.

v0.0.7

New features

• More logging of the command line calls in the Snakemake workflow

Bug fixes

• Fixed a bug where masked positions where not properly masked in the SNP matrix
• Fixed Gubbins dependency check command

v0.0.6

New features

• Added an additional TSV output file with the genomic positions included in the final SNP matrix
• Additional sanitizing for input file names (mainly affects use through Galaxy)
• Clarified some warning messages

Bug fixes

• Fixed a bug where bcftools mpileup would sometimes randomly output empty VCF files because multiple jobs were trying to create the `.fai' index at the same time.

v0.0.5

Various bug-fixes and improvements:

• Added the option --skip-gubbins to skip recombination detection with Gubbins (can be used to work with fragmented reference genomes)
• The log file is now copied to the output directory
• Improved the layout of the output report
• Bugfix: Now uses copy instead of rename to avoid problems when using the read mapping script across file systems

v0.0.4

• PACU is now installable though Bioconda.
• Added an option to disable Gubbins (--skip-gubbins), which can be enabled to use PACU on fragmented reference genomes (i.e., multiple entries in the reference FASTA file).

v0.0.3

First release of PACU that can be installed through pip.
The next step is to make the complete tool available through Conda, including dependencies.