[Biopython-dev] [Bug 2815] Bio.Application command line interfaces

bugzilla-daemon at portal.open-bio.org bugzilla-daemon at portal.open-bio.org
Tue Apr 28 11:44:06 EDT 2009


http://bugzilla.open-bio.org/show_bug.cgi?id=2815





------- Comment #10 from biopython-bugzilla at maubp.freeserve.co.uk  2009-04-28 11:44 EST -------
(In reply to comment #9)
> (In reply to comment #8)
> > I've checked the MUSCLE wrapper into CVS, and added the -diags option.
> 
> You pulled this from the applic-int branch yes? (Hmm, missed that -diags...)

Yes.  I spotted the -diags because it is an example given if you just run
"muscle".

> Ive also just checked in (to the github branch) some unittests for MUSCLE,
> MAFFT, PRANK, and DIALIGN. These just check that the app runs and the return
> code is 0 - a few other checks are made but not much else. 

I'll look at that.

> > A more general alignment wrapper unit test can simply construct some command
> > line objects and check them against an expected string (without requiring
> > the tools to be installed).
> 
> I will do these - all in one test_ApplicationCommandlines.py unittest suite.

Sounds good.  Maybe just test_AlignApps.py if it is just for
Bio.Align.Applications?

> > Note that I am concerned about the file exists check on the input file
> > argument.  This is helpful, but also prevents certain reasonable usage
> > examples - e.g. the input file is created on the fly and doesn't exist
> > yet, or, the command line constructed will be submitted to a cluster
> > where the path will be valid (even if the path isn't valid on the local
> > machine where Biopython is running).
> 
> Good point. Perhaps the os.path.exists on input files needs to be dropped
> from all wrappers.

Maybe - I dropped most of them from the Muscle and Clustalw ones.  The matrix
arguments are a little trickier, where the argument can be either a special
word of a filename.  See below for a related issue ...

> > Also, perhaps we should think about Bio.Application including automatic
> > quoting for filenames with spaces in them... see the _escape_filename
> > function used in Bio.Clustalw and Bio.Blast.NCBIStandalone.  This would
> > be only for parameters explicitly tagged as filenames.
> 
> Yes, I thought about doing that but havent acted.
> 

Another issue is any file exists check needs to be aware that filenames may be
quoted (due to containing spaces).  i.e. A simple call to os.path.isfile(...)
won't work.  I've integrated your Clustalw wrapper into CVS, and in order to
extend my existing unit tests to use this with spaces in file names, I was
forced to drop the existence check.

Peter


-- 
Configure bugmail: http://bugzilla.open-bio.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.


More information about the Biopython-dev mailing list