[MOBY] [MOBY-l] How to use Bio-Moby for publishing my services(in java)...

[Mark Wilkinson]

Hi VG!

I'm going to re-write the "beginners tutorial" in the next few days to
help people get a start with setting up services.  Also, I just got word
this morning that Rebecca Ernst (apparently at your institute!) is going
to be giving a MOBY workshop sometime soon so you could also contact her
for hands-on help.  For Java specific issues, get in touch with Martin
Senger (he's on this list).

First:  read the API
(http://www.biomoby.org/twiki/bin/view/TWiki/BioMOBYAPI)

Now, setting up services isn't difficult (I set up four of them in about
an hour last night with my left hand while debugging the client code
with my right hand ;-) )  Once you have done the first one, the second
and third are a piece of cake, but it still isn't as easy as I would
like it to be.  The main problem is that, until we are able to generate
XSD from our RDF object definitions, it is frustratingly difficult to
figure out what the XML structure of an object will be.  In addition, I
need to set up some nicer query interfaces to help people find the
correct namespace representations (I need to do this anyway as we may
take over the work of curating the GO Xref abbreviations list, so I need
to give the GO people a simple interface into this data as well).

In the meantime, I suggest that you do queries on the database directly
in order to get an overview of the object structure.  At the moment, the
database is still small, so this is a practical short-term solution.

mysql -u moby_external -h mobycentral.cbr.nrc.ca -p mobyobject

(no password)

The following query will generate the RDF assertions that describe the
object structure:

mysql>  select ot1.object_type, rt.relationship_type, ot2.object_type
from object as ot1, object_term2term as rt, object as ot2 where
ot1.object_id = rt.object1_id and ot2.object_id = rt.object2_id;

or, if you prefer LSID's for the objects, use this query:

mysql> select ot1.object_lsid, rt.relationship_type, ot2.object_lsid
from object as ot1, object_term2term as rt, object as ot2 where
ot1.object_id = rt.object1_id and ot2.object_id = rt.object2_id;


You can get a listing of valid namespaces, and their definitions, by
doing the retrieveNamespaces call from the MOBY API, or you can query
the database directly as follows:

mysql> use mobynamespace
mysql> select namespace_type, description from namespace;

If none of these namespaces are applicable, you are welcome to register
a new one, but please post a message to the group indicating what
namespace you need so that we can come up with a suitable prefix/suffix
structure for it.  Similarly with objects - if you need objects other
than those that are already registered, please post a message to the
group describing what kind of data you are describing and we can have an
on-list discussion of what that object should look like in MOBY.  I want
to do this for the first couple of months until people become
comfortable with constructing MOBYesque objects, and then it will become
a free-for-all (unless we get funding for a curator).

Other than that... since I don't know exactly what you concerns are,
those are the only suggestions I can give you.  You might want to find
Rebecca and talk to her.

Let me know how you get on, 

Cheers!

Mark


On Thu, 2003-07-10 at 03:24, Vijayalakshmi Janakiraman wrote:
> Hi Mark,
> 
> I have already implemented web services using Apache Axis running in my
> own server. 
> 
> But, I would like to use BioMoby for the same. 
> 
> 1. Can you give me a start on how to proceed if I need to use BioMoby
> for registering and publishing my services(Java) ?
> 
> Thanks
> VG
> 
> 
> 
> 
> 
> 
> _______________________________________________
> moby-l mailing list
> moby-l at biomoby.org
> http://biomoby.org/mailman/listinfo/moby-l
-- 
Mark Wilkinson <markw at illuminae.com>
Illuminae