Re: [GUSDEV] Storing a provider name for an external database

[Elisabetta M. <man...@pc...>]

I'm not familiar with the specs of these datasets. The following could 
possibly be used to store the info in a way semantically consistent with 
other datasets we store in GUS (e.g. in RAD) at CBIL, but the issue is 
then how to best link this to the particular data for the dataset you 
have (e.g. for microarray studies we have a linking table linking studies 
to assays).

dataset name = Study.Study.name

data provider's name = Study.contact_id->SRes.Contact.name (if individual 
use SRes.Contact.first and last too)

short version of the provider's name = concatenate to SRes.Contact.name in 
parentheses

uri = Study.Study.bibliographic_reference_id->SRes.BibliographicReference.uri
with bib_type "online-resource".

Elisabetta

---

On Thu, 25 Jan 2007, Chris Stoeckert wrote:

> Hi John,
> I'm putting this on the gusdev list as non-ApiDB users may have
> similar questions or some insights to offer.
>
> As we just discussed offline, your task here is not to capture where
> the dataset was downloaded but rather who submitted the dataset to
> the resource (e.g., GenBank)  we got it from. ExternalDatabase and
> ExternalDatabaseRelease are meant for the former. Provider
> information (who should be given attribution) should be captured in
> the SRes.Contact table. I don't see an explicit attribute for a home
> page (the URL) but perhaps one of the RAD people can suggest a
> solution since they use Contact extensively for attribution of
> microrarray studies and and samples.
>
> Chris
>
> On Jan 24, 2007, at 5:04 PM, John Iodice wrote:
>
>> Chris,
>>
>> The attribution info at the bottom of PlasmoDB sequence pages is
>> currently generated by the JSP code, as a function of the species
>> and sequence source_id.  The attribution data in the GFF files we
>> upload to to the Interoperability Working Group are calculated in
>> the SQL in the model files.  We'd like to store this attribution
>> info in the database instead.
>>
>> There are four character strings to be stored for each dataset:
>>
>> the dataset name (e.g. "Plasmodium falciparum mitochondrial genome")
>> the data provider's name ("London School of Hygiene & Tropical
>> Medicine")
>> a short version of the provider's name ("LSHTM")
>> the URL ("http://www.lshtm.ac.uk/pmbu/staff/dconway/dconway.html")
>>
>> Steve and I talked about this today.  We're proposing to store the
>> dataset name, short name and URL in ExternalDatabase.name,
>> ExternalDatabaseRelease.id_type, and
>> ExternalDatabaseRelease.id_url, respectively.  We're proposing the
>> addition of a new varchar column named "provider" to the
>> ExternalDatabaseRelease table.
>>
>> What do you think of this?  Is there a better way?  Does GUS
>> already support this?
>>
>> Thanks for any feedback you can give us.
>>
>> John
>
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys - and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Gusdev-gusdev mailing list
> Gus...@li...
> https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev
>