[MOBY-dev] jmoby using cached datatypes etc

[Andreas Groscurth]

Hi Paul,

as far as i can oversee that, this sounds quite reasonable for me. I was 
always wondered what a Central is if not a registry... so it sounds ok 
to make a connection between them programmatically. As long as this wont 
destroy current usage of the code - but that does not seems to be the 
case as you described it.

Looking forward for this :)

Andreas

Paul Gordon wrote:
> Hi Andreas & Martin,
>
> The issue is that you should really be calling
>
> new MobyDataObject("AGI_LocusCode", "at4g30110", central)
>
> on the last line of your example, but of course it's not your fault.  
> The org.biomoby.shared.data package takes Registry as an optional 
> argument for all of the constructors, and I had a mechanism that 
> allowed you to associate a Registry object with a cache.  Because we 
> are moving to a unified caching mechanism, the ideal solution from my 
> perspective would be for Registry to become an interface, Central 
> extends Registry, and all of the CentralImpls are children of the old 
> Registry object (which becomes RegistryImpl).  Martin, would you have 
> any objections if I did this?  I think the Registry -> Central 
> hierarchy makes sense as you move from metadata to metadata+data.
>
> Of course, if you wanted to use the default central endpoint and 
> caching directory, the
>
> new MobyDataObject("AGI_LocusCode", "at4g30110")
>
> call would be all that is required under the new system.
>
> Regards,
>
> Paul
>
> groscurt wrote:
>> this is an example code:
>>
>> Central central = new CentralDigestCachedImpl(
>> "/home/pcb/groscurt/moby-live/Java/myCache" );
>> MobyService mobyService = new MobyService();
>> mobyService.setAuthority( "arabidopsis.org" );
>> mobyService.setName( "Locus2Publications" );
>> mobyService.setCategory( "" );
>> mobyService = central.findService( mobyService )[0];
>>
>> MobyRequest mobyRequest = new MobyRequest( central );
>> mobyRequest.setService( mobyService );
>> mobyRequest.setInput( new MobyDataObject("AGI_LocusCode", 
>> "at4g30110") );
>>
>> running this starts to download the ontologies from biomoby.org. I 
>> guess the
>> handling of the MobyRequest stuff is the problem here - but how to solve
>> that ?
>>
>> Best and thanks
>> andreas
>>
>> ----------------original message-----------------
>> From: "Martin Senger" martin.senger at gmail.com
>> To: "Core developer announcements" moby-dev at lists.open-bio.org
>> Date: Wed, 25 Jun 2008 18:54:43 +0800
>> -------------------------------------------------
>>  
>>>> but if i understand it correctly the CentralDigestCachedImpl does 
>>>> access
>>>>       
>> a
>>  
>>>> default moby registry and then caches it locally.
>>>>       
>>> Actually no. The class gets, in its constructor, a local directory 
>>> where
>>>     
>> the
>>  
>>> cached ontology are located, and uses it. Only if the cache is 
>>> empty, it
>>> tries to fill it.
>>>
>>>
>>>    
>>>> What I'm interessted in is not to access any registry at all for 
>>>> getting
>>>> the ontologies.
>>>> My sitatuation is that i query the moby central to find some services
>>>>       
>> based
>>  
>>>> on e.g. their input and output parameters.
>>>>       
>>> The cache does not have the "find" method - we did not want to 
>>> replicate
>>>     
>> the
>>  
>>> whole registry's API. But because all ontologies are locally 
>>> available, it
>>> is easy to get them all (that's why the class is also called "digest",
>>> because it provides the cumulative methods that the registry's API does
>>>     
>> not
>>  
>>> provide in all cases) and iterate over them and make your search as you
>>> wish.
>>>
>>> Martin
>>>
>>> -- 
>>> Martin Senger
>>> email: martin.senger at gmail.com,m.senger at cgiar.org
>>> skype: martinsenger
>>> _______________________________________________
>>> MOBY-dev mailing list
>>> MOBY-dev at lists.open-bio.org
>>> http://lists.open-bio.org/mailman/listinfo/moby-dev
>>>
>>>     
>>
>>
>>
>> _______________________________________________
>> MOBY-dev mailing list
>> MOBY-dev at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/moby-dev
>>
>> !DSPAM:60005,48622972109501804129376!
>>
>>
>>
>>   
> _______________________________________________
> MOBY-dev mailing list
> MOBY-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/moby-dev