[BioRuby] [GSoC] weekly report #9

[Artem Tarasov]

Firstly, I wouldn't call that a killer feature. On Un*x you should be able
to use /dev/stdin and /dev/stdout (or a named pipe) as input/output
filenames, that's the way people pipe Picard tools.

Many Un*x tools (including samtools) facilitate that by using dash as a
shortcut for stdin/stdout, but this is not a requirement.

Clearly, STDIN can't be used for random access, and some parts of my code
currently rely on assumption that input stream is seekable. I should make
that optional, and then named pipes can be used as input.

On Tue, Jul 17, 2012 at 11:59 AM, Pjotr Prins <pjotr.public14 at thebird.nl>wrote:

> Are you going to support STDIN/STDOUT? Another killer feature!
>
> On Tue, Jul 17, 2012 at 10:47:49AM +0400, Artem Tarasov wrote:
> > Hello everybody,
> >
> > My progress report for the past week is available at
> > http://lomereiter.wordpress.com/2012/07/17/gsoc-weekly-report-9/
> >
> > I've implemented sorting and merging, both parallelized and quite fast.
> > Also my merging tool improves on ideas taken from Picard source code and
> > merges SAM headers as well as sorted alignment records.
> >
> > For those who use Debian, packages for amd64 and i386 are now available:
> >
> > https://github.com/lomereiter/sambamba/downloads
> >
> > At the moment, alternatives to the following samtools commands are
> > developed: view, index, sort, merge, flagstat. The current limitation is
> > that most tools don't work with stdin/stdout and work with BAM files only
> > (does anybody still use SAM?). Nevertheless, they wisely use multi-core
> > processors and usually give a better speed.
> > _______________________________________________
> > BioRuby Project - http://www.bioruby.org/
> > BioRuby mailing list
> > BioRuby at lists.open-bio.org
> > http://lists.open-bio.org/mailman/listinfo/bioruby
> >
>