Re: [Biopython] Deprecating BlastTextParser

Peter Cock <[email protected]>
Newsgroups gmane.comp.python.bio.general
Message-ID <CAKVJ-_6jRKk0bd47pyF3KcLkBuKNA9+8naUFdfE7dr=SGzBfLw@mail.gmail.com>
So leave the code and its warnings as they are for the time being,
but perhaps stronger wording in the documentation that using
the "blast-text" parser in SearchIO is not encouraged?

i.e. Add a note here?:
https://github.com/biopython/biopython/blob/biopython-172/Bio/SearchIO/__init__.py#L189

And a stronger caveat than "Newer versions may not always work" here?:
https://biopython.org/wiki/SearchIO

Peter
On Thu, Nov 22, 2018 at 9:59 AM Wibowo Arindrarto <[email protected]> wrote:
>
> Dear Michiel, Peter, and everyone,
>
> I actually prefer to have the text parsing functionality in
> maintenance mode: no addition of new features, but also not removing
> it entirely. We probably can make a few things clearer / more visible
> in our documentation: that this was developed using the text output of
> old BLAST releases and that we will not be adding support or updating
> it to work with text output from new releases.
>
> So far, I have seen only very minor (if any) maintenance work related
> to the text output parser. And I believe it still has its uses in
> those corner cases where one really needs to parse the text output
> from an irreproducible BLAST search. So it seems to me that we would
> be losing more (functionality-wise) by deprecating it, rather than
> keeping it in maintenance mode.
>
> With kind regards,
> Bow
> On Wed, Nov 21, 2018 at 5:25 PM Peter Cock <[email protected]> wrote:
> >
> > Hi Michiel,
> >
> > This is up to Bow really - does he want to continue to support
> > parsing plain text human readable BLAST output in SearchIO?
> >
> > I've not tried the new BLAST+ 2.8.0alpha release, but am not
> > aware of any current problems with the parser - although as
> > you note, the NCBI and parts of our own documentation do
> > not recommend parsing it - and it could easily break with the
> > next BLAST+ release.
> >
> > I agree from a code cleanliness perspective, it would be nice
> > to remove this and Bio/ParserSupport.py as well. Sadly while
> > we've had deprecation warnings in place for direct use of
> > Bio.ParserSupport or Bio.Blast.NCBIStandalone, indirect
> > use via SearchIO has been deliberately handled without
> > a warning.
> >
> > Minor correction: Bio.Blast.NCBIStandalone has got tests in
> > test_NCBITextParser.py, and indirectly via the SearchIO tests.
> > In fact it has a respectable 90% test coverage:
> >
> > https://codecov.io/gh/biopython/biopython/src/master/Bio/Blast/NCBIStandalone.py
> >
> > Likewise for Bio.ParserSupport:
> >
> > https://codecov.io/gh/biopython/biopython/src/master/Bio/ParserSupport.py
> >
> > Peter
> > On Wed, Nov 21, 2018 at 2:54 AM Michiel de Hoon <[email protected]> wrote:
> > >
> > > Dear all,
> > >
> > > The blast_text module in Bio.SearchIO.BlastIO contains a parser for plain-text output from Blast.
> > > NCBI and Biopython itself recommend against parsing plain-text Blast output.
> > > As far as I can tell, this module currently has no documentation and no tests.
> > > The module relies on the deprecated module Bio.Blast.NCBIStandalone, which relies on the deprecated module Bio.ParserSupport.
> > >
> > > Can we deprecate Bio.SearchIO.BlastIO.blast_text ?
> > >
> > > Best,
> > > -Michiel
> > >
> > > _______________________________________________
> > > Biopython mailing list  -  [email protected]
> > > http://mailman.open-bio.org/mailman/listinfo/biopython
> > _______________________________________________
> > Biopython mailing list  -  [email protected]
> > http://mailman.open-bio.org/mailman/listinfo/biopython
_______________________________________________
Biopython mailing list  -  [email protected]
http://mailman.open-bio.org/mailman/listinfo/biopython
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.