HEASARC registry metadata and simple publishing protocols

Gretchen Greene greene at stsci.edu
Mon Sep 29 07:34:03 PDT 2003


Tom,Jeongin,

The way we were harvesting Roy and Ray's OAI registries into the
registry here was to use the OAI perl cgi scripts that support the
protocols

oia.pl?verb=ListRecords&metadataPrefix=oai_nvo&set=Services

or

oai.pl?verb=ListRecords&metadataPrefix=vor


AND what was really good was that they encapsulated a VOResource
instance in the OAI record.  For our harvester we simply stripped off
the OAI and parsed the VOResource.   

the hard part is the moving target of voresource makes this hard to
implement,  but this is where we were trying to go with the harvester.
It as very straightforward to capture the elements and load them into
the registry.  We also had registry publication that output the
voresource yet again we will need to repeat this development when the
schema issues are revolved (hopefully soon???).

Gretchen



-----Original Message-----
From: owner-registry at eso.org [mailto:owner-registry at eso.org] On Behalf
Of Thomas McGlynn
Sent: Friday, September 26, 2003 3:27 PM
To: registry at ivoa.net
Subject: HEASARC registry metadata and simple publishing protocols


Jeongin has been working building the metadata for the HEASARC tables
that will go in a registry.  For the moment, we're thinking that
the HEASARC will not have an active, does-everything registry, we'll
simply provide the documentation on our datasets for inclusion in
other registries.  Probably we'll do more in the future.

Jeongin has created something like 300 XML descriptions as separate
files.
Does anyone have any experience going from there to building an OAI
interface that allows others to harvest these descriptions?  E.g.,
as I understand the OAI interface, it basically allows a harvester to
ask for anything that has been updated since a given date.
I'm wondering if something as simple as a shell script like

     cat OAIHeader > output
     find xmlDirectory -mtime $time exec cat {} \>\> output \;
     cat OAITrailer >> output
     cat output

where we just concatenate all the entries that have been modified since
the
date the user has requested might satisfy the OAI harvesters.  Of course
we'd need to parse the CGI arguments and such too.  Is there anything
else we need to do to be OAI compliant?  If we can really do this with
very simple software it will be a real encouragement for others to
publish
their metadata.

It will be a couple of weeks before we're ready to publish these tables
-- they
are under internal review -- but pointers to simple OAI implementations
would be welcome.

	Thanks,
	Tom McGlynn




More information about the registry mailing list