[Go-essp-tech] changing number in ESG dataset link

[Robert S. Drach]

Hi Martina, Estani,

The existing gateway remoteMetadataService can be used to get the 
catalog URI for a dataset. It's documented at:

https://wiki.ucar.edu/display/esgcet/Remote+Metadata+Query+API

The attached script illustrates usage. For example:

% esgcatalog_uri 
http://pcmdi3.llnl.gov/esgcet/remote/hessian/guest/remoteMetadataService 
cmip5.output1.INM.inmcm4.1pctCO2.day.atmos.day.r1i1p1
http://pcmdi3.llnl.gov/thredds/esgcet/18/cmip5.output1.INM.inmcm4.1pctCO2.day.atmos.day.r1i1p1.v20110323.xml

Best regards,

--Bob

Martina Stockhause wrote:
> Hi, Bob,
>
> as Estani mentioned, we have a problem with changing root URLs 
> (<number> in the below string) for the ESG dataset links:
> http://<datanode>/<service>/esgcet/<number>/<dataset_id>.html
>
> For the DOI data (aggregation on DRS experiment level) we prepare 
> links to the ESG datasets belonging to it. Currently, we construct 
> them out of the root URL for the asked location and the <dataset_id>. 
> If that root URLs change over the CMIP5 data at a location, this does 
> not work any longer.
>
> What can we do about it?
>
> 1. Since the dataset_id is unique, why do we need to know that 
> <number> at all for the ESG dataset link?
> Is it possible to set-up a service that redirects to the right URL, 
> e.g. http://<datanode>/<service>/esgcet/<dataset_id> redirecting to 
> the above URL?
>
> That would be our preferred solution, since all other solutions means 
> a lot of work for my db admin and for me.
> And for these options 2 and 3 the minimal preconditions are:
> - This root URL does not change after DOI assignment.
> - The root URL is the same for all ESG datasets of a single DRS 
> experiment on a single DN.
>
> 2. You mentioned an API that supplies this number.
> How is it called? The call has to be open (without security checks).
> We would implement calls for all ESG datasets before DOI assignment 
> and adapt our metadata, esp. for the location of the data replica.
>
> 3. Is there a range for these numbers? Are they assigned as increasing 
> numbers or totally free numbers?
> If the range is an increasing number of a small range, we could check 
> for the right number and go on as in 2.
>
> Estani will discuss this with you in Ashville to find the best 
> solution to this problem.
>
> Thanks and best wishes,
> Martina
>

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: esgcatalog_uri
Url: http://mailman.ucar.edu/pipermail/go-essp-tech/attachments/20110505/e72e6003/attachment.pl