vcfcache

Read full text
Authors: Julius Müller
Cited by: 0

Abstract

0.5.0 (2026-01-07) Added Enforced genome build metadata in configuration: genome_build is now required in both params.yaml and annotation.yaml Workflow startup logs the genome build from both configs Mismatched genome builds now fail fast vcfcache compare now captures run timestamps and thread counts from snapshots/completion flags Added --no-stats to disable stats/log output from vcfcache annotate Changed BREAKING: Removed all contig renaming/variant-generation support. The --generate-contig-variants option is removed. Annotation now logs contig overlap at startup and fails if there is no overlap. BREAKING: vcfcache annotate now writes directly to an output BCF file instead of creating an output directory. Use --stats-dir to store workflow logs/auxiliary files under _vcstats. Use - or stdout to stream output to stdout. Any tooling or scripts that expected an output directory now need to pass a file path (e.g., sample_vc.bcf). --stats-dir now always uses the input basename ( _vcstats) for the stats directory name vcfcache compare output reformatted with structured sections and a summary table vcfcache compare reports two rates (end-to-end output throughput and tool-step throughput) vcfcache compare now reports extra params only when they differ between runs (with values) Deprecated scripts/release.sh and simplified release steps in RELEASE.md CI workflow now runs on both tag pushes and main branch pushes (for codecov badge updates) Dockerfile optimizations: moved tests to test stage only, added pip cache mounts for faster builds Fixed Contig overlap is now reported on every cached annotation run and no-overlap failures are surfaced early. Fixed AttributeError: 'Namespace' object has no attribute 'verbose' when running vcfcache compare command Added missing parent parser to compare command to inherit --verbose, --quiet, and --debug flags Enhanced vcfcache compare to show detailed timing breakdown similar to old benchmark mode Now displays detailed step-by-step timings for each bcftools command in both cached and uncached runs Shows comprehensive summary with total time, speedup factor, time saved (in seconds, minutes, and hours) Improved workflow.log parsing to correctly extract "Workflow completed successfully in X.Xs" timing Human-readable time formatting (e.g., "1h 23m 45.6s" or "12m 34.5s") NEW: Displays genome_build and threads from params.snapshot.yaml for both runs NEW: Counts and compares total variants in output files using bcftools index -n NEW: Shows both variant count comparison and MD5 checksum verification Removed warning about comparing runs in same mode (useful for comparing different caches) Added comprehensive test suite for compare command (tests/test_compare.py) with 16 tests

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
7 January 2026