RabbitBAM

An ultra-fast I/O framework for BAM files.

Dependancy

• gcc 8.5.0 or newer
• htslib 1.15 or newer
• libdeflate 1.12 or newer

Installation

git clone https://github.com/RabbitBio/RabbitBAM
cd RabbitBAM
bash configure.sh <path-to-htslib-installation-directory> <path-to-libdeflate-installation-directory>
source env.sh
make clean && make

Benchmark

refer to ./rabbitbam -h

Usage

To use RabbitBAM in other software, you need to first compile RabbitBAM according to the Installation instructions, then link it to your new software and include the appropriate header files.

Here is an example of file reading and writing (see RabbitBAM-QC and RabbitBAM-SORT for more examples):

Source:

#include <iostream>
#include <cstdlib>
#include <unistd.h>
#include <thread>

#include "BamTools.h"
#include "BamReader.h"
#include "BamWriter.h"

int main(int argc, char* argv[]) {
    std::string input_file;
    std::string output_file;
    int n_thread = 1;

    int opt;
    while ((opt = getopt(argc, argv, "i:t:o:")) != -1) {
        switch (opt) {
            case 'i':
                input_file = optarg;
                break;
            case 't':
                n_thread = std::atoi(optarg);
                break;
            case 'o':
                output_file = optarg;
                break;
            default:
                std::cerr << "Usage: " << argv[0] << " -i input_file -t thread_num -o output_file\n";
                return EXIT_FAILURE;
        }
    }

    if (input_file.empty() || output_file.empty() || n_thread <= 0) {
        std::cerr << "Invalid arguments. Usage: " << argv[0] << " -i input_file -t thread_num -o output_file\n";
        return EXIT_FAILURE;
    }

    std::cout << "Input file: " << input_file << "\n";
    std::cout << "Total thread number: " << n_thread << "\n";
    std::cout << "Output file: " << output_file << "\n";

    int compress_level = 6;
    int block_size = 256;
    bool single_parser = true;

    int n_thread_read = n_thread / 5;
    int n_thread_write = n_thread - n_thread_read;
    if(n_thread_read <= 0) n_thread_read = 1;
    if(n_thread_write <= 0) n_thread_write = 1;

    BamReader *reader = new BamReader(input_file, n_thread_read, single_parser);
    BamWriter *writer = new BamWriter(output_file, reader->getHeader(), n_thread_write, compress_level, block_size, single_parser);

    bam1_t *b;
    if ((b = bam_init1()) == NULL) {
        fprintf(stderr, "[E::%s] Out of memory allocating BAM struct.\n", __func__);
    }

    long long num = 0;
    while (reader->getBam1_t(b)) {
        num++;
        writer->write(b);
    }

    writer->over();
    printf("num %lld\n", num);

    return 0;
}

Compile:

export HTSLIB_INSTALL_PATH=<path-to-htslib-installation-directory>
export RABBITBAM_INSTALL_PATH=<path-to-RabbitBAM-installation-directory>
export LIBDEFLATE_INSTALL_PATH=<path-to-libdeflate-installation-directory>
export LD_LIBRARY_PATH=$RABBITBAM_INSTALL_PATH:$LD_LIBRARY_PATH

g++ -o main main.cpp -I$HTSLIB_INSTALL_PATH/include -I$RABBITBAM_INSTALL_PATH/htslib -I$RABBITBAM_INSTALL_PATH -I$LIBDEFLATE_INSTALL_PATH/include -L$HTSLIB_INSTALL_PATH/lib -L$RABBITBAM_INSTALL_PATH -L$LIBDEFLATE_INSTALL_PATH/lib -lhts -lz -fopenmp -lpthread -lrabbitbamtools -lrabbitbamread -lrabbitbamwrite

Run:

time ./main -i in.bam -o out.bam -t 40

Note: This is for the single parser case. If you need multiple parsers, please refer to the example parallel_test in block_mul.cpp.