[Biojava-l] removeGap problem with SimpleGappedSequence

[david de beule]

Hi,

your first solution (new constructor: new GappedSymbolList(origSyms,
mergeOriginalGaps)
sounds good to me
These new constructors for GappedSymbolList and GappedSequence would be very
helpful.

Thanks in advance.
David


----- Original Message -----
From: "Matthew Pocock" <matthew_pocock at yahoo.co.uk>
To: <mark.schreiber at group.novartis.com>
Cc: <daviddebeule at pandora.be>; <biojava-l at biojava.org>
Sent: Thursday, February 12, 2004 12:02 PM
Subject: Re: [Biojava-l] removeGap problem with SimpleGappedSequence


> Hi,
>
> Seems like we have a bit of an 'expected behavior' and 'implemented
> behavior' gap. If we decide to modify the GappedSymbolList constructor
> to find all gaps in the original sequence, I think we should add it as
> an option:
>
>   new GappedSymbolList(origSyms, mergeOriginalGaps)
>
> and make the current constructor equivalent to this(syms, false).
> Finding all these gaps, making an ungapped underlying symbol list, and
> building the gap insertion data structures is a potentialy expensive
> operation (imagine gapping a genome! you would pull the whole thing into
> memory and do a linear scan), so we should be careful not to force it
> upon the world.
>
> This would also change the contract of getSourceSymbolList() and also
> what happens if that source is modified, wether changes to it are tracked.
>
> This could be worked around by implementing an "UnGappedView" class that
> does the oposite mapping of GappedSymbolList - removes all gaps in the
> source - then we could gap this putting them all back, making it
> editable. I don't wan't to be the one to write it though - writing
> GappedSymbolList made my brain hurt.
>
> Matthew
>
> mark.schreiber at group.novartis.com wrote:
>
> >Sounds like a pretty sensible suggestion. Can anyone think of why this
> >might not be a 'good idea'?
> >
> >If not, i'll add it to the list of things to fix :)
> >
> >- Mark
> >
> >
>
>
>