OWSLib can't open the CWIC CSW

[rsignell-usgs]

When I try to open NASA's CWIC CSW endpoint:
http://cwic.csiss.gmu.edu/cwicv1/discovery
listed at on the CEOS CWIC page, I get an error:

endpoint = 'http://cwic.csiss.gmu.edu/cwicv1/discovery'  # CWIC
csw = CatalogueServiceWeb(endpoint,timeout=60)

ExceptionReport: 'REQUEST_EXCEPTION: UNRECOGNIZED_REQUEST - Only GetRecords and GetRecordById supported for HTTP POST method.'

I'm guessing this isn't really an OWSlib issue, but more a provider issue, right?
Can you tell what the CWIC folks would need to fix to have OWSlib work with this CSW?

Thanks,
Rich

[tomkralidis]

@rsignell-usgs the CWIC CSW will have to support HTTP POST for GetCapabilities, not just HTTP GET.

[rsignell-usgs]

Okay, thanks, I'll tell them.

[tomkralidis]

@rsignell-usgs turns out for GetCapabilities, HTTP POST is optional, and HTTP GET is mandatory to support. So the CWIC CSW is doing things right. So owslib.csw.CatalogueServiceWeb() needs to be updated to send the GetCapabilities request as HTTP GET.

[tomkralidis]

ref: http://portal.opengeospatial.org/files/?artifact_id=20555, table 54.

[rsignell-usgs]

Great! That will be a nice enhancement for OWSLib. We certainly can't count on optional functionality!

[rsignell-usgs]

How big a job is this?

[tomkralidis]

@rsignell-usgs See branch at https://github.com/tomkralidis/OWSLib/tree/issue-76. I tried it and it works for me (although there are GetRecords which are not related -- another investigation). Does this work for you for GetCapabilities?

[rsignell-usgs]

Yes, this fixed the GetCapabilities issue! (but as I suspect you are, I'm now getting a 500 server error when I try to get records)

[capooti]

Hi @tomkralidis
not sure if it is a related issue and before I open another issue: using this catalog, where POST in getrecords seems to be supported, I get the following error in getrecords:

from owslib.csw import CatalogueServiceWeb
csw = CatalogueServiceWeb('http://www.rndt.gov.it/RNDT/CSW?')
csw.getrecords(keywords=['test'], maxrecords=20)
UserWarning: Please use the updated 'getrecords2' method instead of 'getrecords'.
The 'getrecords' method will be upgraded to use the 'getrecords2' parameters
in a future version of OWSLib.
in a future version of OWSLib.""")
TypeError: int() argument must be a string or a number, not 'NoneType'

any idea? thanks

[tomkralidis]

@capooti thanks for the info. My trace:

>>> c=s('http://www.rndt.gov.it/RNDT/CSW?')
>>> c.getrecords(keywords=['test'], maxrecords=20)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "owslib/csw.py", line 236, in getrecords
    self.results['matches'] = int(util.testXMLValue(val, True))
TypeError: int() argument must be a string or a number, not 'NoneType'
>>> print c.response
<?xml version="1.0" encoding="utf-8"?>
<csw:GetRecordsResponse xmlns:csw="http://www.opengis.net/cat/csw/2.0.2">
<csw:RequestId xmlns:ogc="http://www.opengis.net/ogc" xmlns:ows="http://www.opengis.net/ows" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" />
<csw:SearchStatus />
<csw:SearchResults elementSet="summary" recordSchema="csw:Record" />
</csw:GetRecordsResponse>

. This is not a valid CSW endpoint. CSW 2.0.2 requires the numberOfRecordsReturned and numberOfRecordsMatched attributes within the csw:SearchResults element as part of GetRecords responses.

[capooti]

You are right, it was a problem of their implementation, now they have fixed. Thanks