[Biojava-dev] porting HMM classes from bj1.8 to bj3

[Andreas Prlic]

Hi Zhihua,

I think the danger is to include essentially all of biojava 1 in this
module, which is something we want to avoid. You are right, the ideal
solution will contain all 3 aspects (but in the reverse order).
Ideally there will no dependency on biojava 1 things like annotations,
changetype and bioerror. You can re-route file parsing and
representation of sequences to biojava3-core and you can cut away as
much as possible to filter out what is essential for the HMM
functionality...

Andreas

2011/8/16 zhihuali <lzhtom at hotmail.com>:
> Andreas,
>
> I'm using NetBeans because it has a nice support for Maven projects.
> I was able to add a "hmm" module to the legacy bj1.8 project and in the
> process of transferring hmm-related classes from bj1.8 core module to
> the new hmm module. These hmm classes are heavily dependent on other
> core classes. Some of them are non-essential ones for example BioError,
> Annotation, ChangeType.....which may or may not have the corresponding
> classes in bj3.
>
> I think there are three strategies:
> a. keep the dependent classes as legacy classes in the hmm module
> b. re-route the dependency to bj3 classes as much as possible
> c. remove these non-essential dependencies as much as possible
>
> The real solution will be a combination of these three, but I was wondering
> which way should have
>  the highest priority.
>
> Thanks a lot!
>
> Zhihua
>
>
>> Date: Fri, 12 Aug 2011 10:38:08 -0700
>> Subject: Re: [Biojava-dev] portin! g HMM classes from bj1.8 to bj3
>> From: andreas at sdsc.edu
>> To: lzhtom at hotmail.com; HWillis at scripps.edu
>> CC: biojava-dev at lists.open-bio.org
>>
>> Hi Zhihua,
>>
>> good to see that your message made it through in the end. What IDE are
>> you using? I recommend checking out the latest version of
>> biojava-legacy (1.8) and doing the refactoring on the (legacy) core
>> module. A first step could be to break out a HMM module out of the old
>> core. Then this module could be migrated to the 3.X design. I see
>> some difficulty coming up with how to deal with the Symbols from the
>> 1.X world, so let's keep discussing this as you make progress...
>>
>> Andreas
>>
>> On Fri, Aug 12, 2011 at 7:54 AM, Scooter Willis <HWillis at scripps.edu>
>> wrote:
>> > I think first step is to take the HMM classes from 1.8 and migrate with
>> > zero dependencies of any non HMM classes in ! Biojava 1.8. If you can it
>> > to
>> > compile with a minimal numb er of external packages then migration as a
>> > standalone module should be easy. If you work with Strings as inputs
>> > that
>> > should help minimize dependencies.
>> >
>> > On 8/12/11 10:45 AM, "zhihuali" <lzhtom at hotmail.com> wrote:
>> >
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>
>> >>Dear all,
>> >>
>> >>I have been trying to post this message several times now but couldn't
>> >>see it in the list. Hopefully this time it will get through. If you
>> >>received
>> >>multiple copies of this email please forgive me.
>> >>
>> >>I was pointed to here by Andreas and Sylvain from LinkedIn.
>> >>Basically we felt that a great missing piece in bj3 from legacy biojava
>> >>! is HMM classes. And I'd like to take on this task. Maybe the first
>> >>thing is to migrate bj1.8's dp package and related packages to bj3.
>> >>Could anyone give me directions to get it started?
>> >>
>> >>Thanks a lot!
>> >>
>> >>Zhihua Li
>> >
>> >
>> > _______________________________________________
>> > biojava-dev mailing list
>> > biojava-dev at lists.open-bio.org
>> > http://lists.open-bio.org/mailman/listinfo/biojava-dev
>> >
>



-- 
-----------------------------------------------------------------------
Dr. Andreas Prlic
Senior Scientist, RCSB PDB Protein Data Bank
University of California, San Diego
(+1) 858.246.0526
-----------------------------------------------------------------------