[Biojava-l] Adapting org.biojava.bio.seq classes to an interface?

[Christian Storm]

Thanks!

Ended up implementing SequenceBuilder from scratch rather than subclassing 
SequenceBuilderBase, but now it works nicely.

Christian


----- Original Message ----- 
From: "Thomas Down" <td2 at sanger.ac.uk>
To: "Christian Storm" <cs at is-analyse.de>
Cc: <biojava-l at biojava.org>
Sent: Thursday, December 16, 2004 3:54 PM
Subject: Re: [Biojava-l] Adapting org.biojava.bio.seq classes to an 
interface?


>
> On 16 Dec 2004, at 12:35, Christian Storm wrote:
>
>> Is there a way of adapting org.biojava.bio.seq classes / interfaces to an 
>> interface?
>>
>> I would like to add functionality to some of the interfaces in 
>> org.biojava.bio.seq, i.e. sequence and feature, by adapting them to an 
>> existing interface.
>>
>> I see two possibilites - a class adapter:
>>
>> OwnSequence extends SimpleSequence implements NewFunctionality
>>
>> or an object Adapter:
>>
>> OwnSequence implements Sequence, NewFunctionality
>>
>> Unfortunately both approaches would mean massive recoding of the parsing 
>> /building process of sequence-objects, so that in essence it would be 
>> simpler for me to write my own classes to handle sequences.
>
> Hi Christian,
>
> The BioJava sequence model was written with the possibility of dropping in 
> new implementations in mind, and it certainly ought to be possible to do 
> this -- whatever approach you choose to take -- without having to rewrite 
> things like parsing code.  A good place to start looking is the 
> SequenceBuilder interface, which receives events from the parsing 
> framework and constructs Sequence objects.  By developing a new 
> implementation of this, you'll be able to use any of the BioJava parsers. 
> Depending on the exact details of your new Sequence implementation, you'll 
> quite possibly be able to get away with a very small subclass of 
> SequenceBuilderBase.
>
>          Thomas.
>