[ckan-dev] Temporal extent ingest failing in CSW spatial harvester?

Bruce Crevensten becrevensten at alaska.edu
Fri Jan 30 01:10:30 UTC 2015


Hi, I did further diagnosis and found that I could get a successful ingest
of the temporal extent by changing the XML namespace declaration for gml
from:

"gml": "http://www.opengis.net/gml/3.2",

to:

"gml": "http://www.opengis.net/gml/"

...which is identical to what GeoNetwork (version 2.6.1) is outputting.

Am I encountering a somewhat trivial edge case, or perhaps it could be
possible to provide a default set of namespaces in the harvester, with an
option to override via the configuration object for the CSW harvester?


On Thu, Jan 29, 2015 at 11:58 AM, Bruce Crevensten <becrevensten at alaska.edu>
wrote:

> Hi, I'm working on a custom harvester that extends the CSW harvester and
> while most ISO fields are coming through, the ingest of temporal extent is
> failing.  We're harvesting from GeoNetwork and an example specimen we're
> ingesting is here:
>
> https://athena.snap.uaf.edu/geonetwork/srv/xml.metadata.get?id=22
>
> The XPath expressions in the relevant ISODocument code
> <https://github.com/ckan/ckanext-spatial/blob/master/ckanext/spatial/model/harvested_metadata.py#L762>
> looks like it should be working and I confirmed in a 3rd party tool that
> they do, in fact, work.  I tried a few varying paths in the extended
> get_package_dict function, but no luck.  I also tried running the built-in
> CSW harvester (getting my custom code out of the loop), but it also fails
> to capture temporal extent.
>
> Other XPath expressions I'm using to pull out data from the harvest source
> are working, and the ISODocument is reasonably complete (there are places
> where our metadata may not be entered into GeoNetwork in a way that CKAN
> understands, but the Temporal Extent looks solid).
>
> Our CSW endpoint for this harvest job is here:
>
>
> https://athena.snap.uaf.edu/geonetwork/srv/en/csw?request=GetCapabilities&service=CSW
>
> It doesn't seem like we have a malformed document or XPath, so I almost
> wonder if this relates to a lower-level Python or other system
> configuration in some way?  We're running:
>
> CentOS 6.5
> CKAN 2.2.1 (from source install)
> Python 2.2.6
> LibXML 20900
>
> Ingest of temporal extent also fails using a CKAN 2.0 install we're
> running on a similar platform (same versions of libraries listed above).
>
> If any developer were able to try a CSW harvest job using a known-good
> installation or could give some suggestions on where to look next, it would
> be much appreciated!
>
> --
> *Bruce Crevensten*
>
>
> *Senior Software Engineer*
> (SNAP) Scenarios Network for Alaska & Arctic Planning
> www.snap.uaf.edu
>



-- 
*Bruce Crevensten*


*Senior Software Engineer*
(SNAP) Scenarios Network for Alaska & Arctic Planning
www.snap.uaf.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-dev/attachments/20150129/809b8aaa/attachment-0003.html>


More information about the ckan-dev mailing list