[pycsw-devel] CSW search by other fields

[Tom Kralidis]

Hi Juan:

On Wed, Sep 30, 2015 at 10:53 AM, Juan Rapoport
<juanrapoport at suremptec.com.ar> wrote:
> Hi Tom,
>
> We are now implementing a catalog using the ISO 191939 profile to store
> raster metadata.
> We store some metadata about the acquisition platform and sensor inside
> fields like supplementalInformation (MD_DataIdentification ->
> gmd:supplementalInformation), collectiveTitle (CI_Citation ->
> gmd:collectiveTitle) or alternateTitle (CI_Citation -> gmd:alternateTitle),
> and we need to query by these fields.
>
> We are starting another project using the ISO 19139-2 standard to catalog
> products from a SAR satellite mission. We didn't define yet the fields we
> are going to use, but this profile allows to specify metadata about the
> satellite, sensor, processing algorithms, etc. and we'll need to search
> using (al least) all these fields.
>

Sounds like the best bet is likely to allow for something
configuration based, i.e.:

[repository]
mapping_foo:bar=name_of_column
mapping_foo2:bar=name_of_column2

Then pycsw would pick up these queryables at runtime for the following
workflows:

- advertising additional queryables in Capabilities
- ingesting new metadata to check/parse/store the additional queryables
- accepting GetRecords queries with the additiona queryables


> In our case, we don't need to add other fields to the response because we
> get records in gmd output schema
> (outputSchema="http://www.isotc211.org/2005/gmd"). This return the same xml
> used to catalog, right?
> Then we parse the response using  jsonix
> (https://github.com/highsource/jsonix) and ogc-schemas
> (https://github.com/highsource/ogc-schemas) to transform the XML response
> into JSON.
> ogc-schemas recently added suport for the ISO 19139-2
> (https://github.com/highsource/ogc-schemas/issues/109).
>
> For now we are using the standard database; I don't know if additional
> columns will be the solution. Maybe using xpath to query inside the xml and
> allow to configure the queryable fields (AdditionalQueryables).
>

This was the originally designed in this manner in the very early days, but
we found that this did not scale very well, so we moved to a pure RDBMS
approach.  This doesn't mean we can't revisit this again if there is interest.

..Tom


> Thanks!
>
>
> 2015-09-29 23:58 GMT-03:00 Tom Kralidis <tomkralidis at gmail.com>:
>>
>> Hi Juan:`
>>
>> On Tue, Sep 29, 2015 at 3:54 PM, Juan Rapoport
>> <juanrapoport at suremptec.com.ar> wrote:
>> > Hi,
>> >
>> > Is it possible to configure which are the queryable fields for the
>> > GetRecords requests?
>> >
>>
>> Can you provide more context or use case(s)?  Is the goal here solely
>> to extend queryables against additional columns from a custom
>> database?  Do these queryables need to be provided in the CSW response
>> (i.e. returnables)?  Can you provide some examples of what sample
>> queries could look like?
>>
>> We've had similar discussions before, and implementation/enhancement
>> options range from allowing an extensible queryables approach via
>> configuration, to custom non-standard CSW application profiles (which
>> I'd tend to avoid unless absolutely required).
>>
>> Cheers
>>
>> ..Tom
>>
>> > Thanks
>> >
>> > --
>> > Juan Rapoport
>> > SUR Emprendimientos Tecnológicos
>> >
>
>
>
>
> --
> Juan Rapoport
> SUR Emprendimientos Tecnológicos
>
> Av. Rivadavia 611 Piso 11º  (C1002AAE)
> Ciudad de Buenos Aires, Argentina
> Tel. +54 (11) 4342-2976/84
> juanrapoport at suremptec.com.ar
> www.suremptec.com