Drop-in CLI direction
Supports normal bgzip workflows including stdin/stdout pipelines, compression levels, threads, file mode, integrity checks, and aliases.
Tara open source family
BGZF built for reliable HTS pipelines.
TaraBG is a clean-room Rust command-line tool for compressing, validating, indexing, and randomly accessing genomic data in the BGZF format.
Verified against HTSlib.
The audited workload used the supplied HTSlib 1.24 development snapshot and a real 99.2 MB 1000 Genomes chr22 VCF slice.
Test suite41/41 passing
HTSlib operationsValidate · identify · index · query
Region result9,747 identical records
Level 6, four threads0.15 s TaraBG · 0.21 s bgzip median
Compression tradeoff1.40x faster · 3.08% larger output
This is a reproducible workload result, not a universal speed claim. Compression results vary with input, level, thread count, hardware, and filesystem.
Operational compatibility.
Built around observable BGZF behavior and published format requirements, without copying HTSlib implementation code.
- Streaming BGZF compression and decompression with bounded memory
- Parallel compression with preserved block ordering
- BGZF CRC32 and ISIZE validation with canonical EOF markers
- .gzi creation, rebuilding, random reads, and rebgzip workflows
- Text-aware blocks plus fixed binary block mode
- Linux, macOS, and Windows continuous integration
Indexed access
Creates and rebuilds .gzi indexes, reads byte ranges, and re-blocks streams against index boundaries.
Compatibility discipline
Native differential tests, malformed-input coverage, fuzz targets, and cross-decoder checks protect the format contract.
Measure it on your workload.
The repository includes native comparison and HTSlib downstream workload harnesses with byte-for-byte verification.