[Biojava-l] SimpleRichAnnotation
James Procter
jimp at compbio.dundee.ac.uk
Mon Dec 8 09:41:28 UTC 2008
Thanks for the reply, Richard - you've cleared up the situation immensely.
Richard Holland wrote:
>
> (d) why are Terms using the == equality test instead of equals().
> (d) for speed, and on the assumption that nobody would want to have two
> copies of the exact same ontology in memory at any one time.
this is what I thought - and I understand both sides. However, there is
always a potential bone of contention here. Is there some kind of
mechanism to normalise (or register) an ontology/term instance against
the set of known ontology instances ? This might prove useful in
situations where ontologies have been imported ad-hoc from alternative
sources (such as a web service, where a clone of the object structure is
received by the server which will have their own object references).
> (b) how can you get a singleton ontology instance loaded from BioSQL
> (b) Use the getObject() method on the RichObjectFactory after already
> having called the connectToBioSQL() method.
Ah - this is what I missed - in the documentation or otherwise :)
However - this isn't going to work perfectly in my situation. I am
expecting to need thread safe access (as far as is possible) to several
biosql backed rich sequence databases - and potentially, each instance
may work in a different namespace. I'm sure you can see how this
requirement conflicts with the use of a default ontology singleton for
term import in equivalence methods.
> (a) getProperty()/getProperties() is deprecated but hasn't been marked
Its good to know this - as it happens, I ended up using the getNoteSet
and implementing an equals() based term equivalence mechanism for my
specific sitation.
> I hope this helps.
it does !
thanks again,
Jim.
--
-------------------------------------------------------------------
J. B. Procter (ENFIN/VAMSAS) Barton Bioinformatics Research Group
Phone/Fax:+44(0)1382 388734/345764 http://www.compbio.dundee.ac.uk
The University of Dundee is a Scottish Registered Charity, No. SC015096.
More information about the Biojava-l
mailing list