2.0.6

• Replace c5d_2xlarge/c5d_9xlarge labels with single high_cpu label
• Automatic CPU-based concurrency: executor.cpus defaults to all available processors; high_cpu processes cap at min(nproc, available)
• Vectorize seqmap construction in dada2_dada.R for faster dereplicate step
• New bin/create_manifest.py --batch argument (defaults to 1)
• Add OCI labels to Dockerfile for GitHub Container Registry
• Documentation: CPU behavior, --config override example

Full Changelog: 2.0.5...2.0.6

2.0.5

What's Changed

• when splitting, use hit with highest score for each seqname, instead of last by @dhoogest in #93
• Readme update by @crosenth in #95
• Updating deps by @crosenth in #94
• Binned quality option by @dhoogest in #98 (GH #99 )

• Optional binned quality scores parameter for DADA2 error learning (GH #99)
• Pin DADA2 to Bioconductor 1.40.0 stable commit
• Add sha256sum regression checks to CI tests

Full Changelog: 2.0.4...2.0.5