[Biojava-l] Re: basic sequence-fetching methods

[Marc Colosimo]

Thomas Down wrote:

>
> For those who hadn't already heard about it, O'Reilly have
> recently run a `hackathon' event, with representatives of all
> the open-bio projects, plus a number of other bioinformatics
> programmers.  Mathew Pocock and myself attended as BioJava
> representatives.  The main thing which came out in the meeting
> was agreement on a set of basic sequence-fetching methods,
> which should be supported by Bio{Perl,Python,Java,Ruby}.  A
> draft document can be downloaded from:
>
>     http://www.technophage.com/open-bio-database.pdf
>

[snip]

That is great to hear that a standard set of methods will be made. I've been,
slowly, working on adding a new sequence reader and notice several things and
I have some comments about them.

Briefly, I like the bioperl IO methods model much better. Basically, if my
memory severs me correctly, you make one call to open a sequence (like what is
in the SeqIOTools). What I have in mind is something  like:

theFile = new SeqIO("filename", "file format");

This could also try to guess the format.

this would solve one of my problems and possible bugs. My reader is for a
binary file format that I need to read in the header before any calls are made
by StreamReader. Also, the way it is set up now, one can accidentally send a
different BufferedReader to the underlining file classes since they are not
encapsulated in the class.

Making these changes should be easy and probably not mess up anything using it
the current way. What do people think about this?

I'm new here so maybe this has already be hashed out awhile ago or I just
don't fully understand the design of BioJava, yet

Marc