[Biopython-dev] Plain text BLAST parser updated

[Peter]

On Tue, Oct 20, 2009 at 11:47 AM, Peter <biopython at maubp.freeserve.co.uk> wrote:
> Hi all,
>
> Just to let you know I've been doing a little work on the NCBI plain
> text parser, and got it to work on multiquery output from recent
> versions of BLAST (Bug 2090).
>
> http://bugzilla.open-bio.org/show_bug.cgi?id=2090
>
> I would not describe the changes as elegant, but the plain text parser
> has evolved over time to cope with more and more NCBI variations, so
> some ugliness is perhaps to be expected.
>
> If there are any regressions, please report them and we can extend the
> test suite. Likewise, if you have an recent plain text BLAST files
> which didn't work and still don't work - get in touch, it may be easy
> to fix.
>
> [I'd still encourage everyone to use the XML output by default, but
> there are times when the plain text is the only or best option.]
>
> Peter

The irony of this isn't lost on me. An hour after fixing the parser
and testing it with BLAST 2.2.21 output, the NCBI made a release
which broke it again.

The output from the latest "classic" blastall 2.2.22 is fine.

The output from the "new C++" blastx 2.2.22+ (and likely the
other tools like blastp etc which are all separate executables
now) breaks our plain text parser.

I'll also be trying out the XML output which *should* be fine.

Peter