This is the I/O library of the BioC++ project. It provides the io module which offers easy-to-use interfaces for the following formats:

• txt formats: plain-text files (line-wise reading) and delimited files (e.g. CSV and TSV).
• seq formats: FastA and FastQ.
• var formats: VCF and BCF.

The primary goal of this library is to offer higher level abstractions than the C libraries typically used in this domain (e.g. htslib) while at the same time offering an excellent performance. It hopes to offer a modern, well-integrated design that covers most typical I/O use-cases Bioinformaticians encounter.

The library provides stand-alone implementations of the formats (and does not require htslib). Support for reading and writing SAM/BAM/CRAM will be available in separate library until full implementations are available here.

Please see the online documentation for more details.

Attention: this library is currently a work-in-progress, and interfaces are not yet stable.

Simple reading of a FastA-file that is transparently decompressed:

bio::io::seq::reader reader{"example.fasta.gz"};

for (auto & rec : reader)
{
  fmt::print("ID:  {}\n", rec.id);
  fmt::print("Seq: {}\n", rec.seq);
}

Reading a variant file and writing a new one that only contains variants that "PASS":

bio::io::var::reader reader{"example.vcf.gz"};
bio::io::var::writer writer{"example.bcf"};

for (auto & rec : reader)
  if (rec.filter.empty() || (rec.filter.size() == 1 && rec.filter[0] == "PASS"))
    writer.push_back(rec);

The format is transparently converted from compressed VCF to BCF if files have the respective extensions / magic headers.

• Header-only → just drop it in your source code or include it as a git submodule!
• The BioC++ core library is the only hard requirement.
• No build-system and no configure steps required.
• Optional CMake support available.
• Integrates well with the C++20 standard library.

• The recommended way to install BioC++ libraries is to follow the instructions in the meta-repository: https://github.com/biocpp/biocpp
• CMake is recommended, because it finds dependencies automatically and sets required flags/paths.
• Quick instructions without CMake:
  1. Clone biocpp-core and biocpp-io into e.g. ~/devel.
  2. Assuming you have ZLib and Bzip2 in system paths, do:

g++ -O3 -DNDEBUG -Wall -Wextra -std=c++20           \
    -I ~/devel/biocpp-core/include                  \
    -I ~/devel/biocpp-io/include                    \
    -DBIOCPP_IO_HAS_ZLIB=1 -DBIOCPP_IO_HAS_BZIP2=1  \
    your_file.cpp

• If you want to use {fmt}, add -I ~/devel/fmt/include -D FMT_HEADER_ONLY=1.