OAI PMH Provider is not responding

uwieske

I have installed CA and configured oai_provider.conf in order to expose my repository data to the outside world. (in external data is expressed in Dublin Core). I added an object and have given it an object identifier (idno), i.e. something like "ITEM.001". I have tried a couple of example profiles just to test OAI. This is my URL request http://www.nm-dev.orh/ca/service.php/OAI/dc?verb=GetRecord&identifier=oai%3Awww.nm-dev.org%3AITEM.001&metadataPrefix=oai_dc

I got an "Identifier is invalid" error message in my response.

<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"> <responseDate>2012-11-02T19:25:13Z</responseDate> <request verb="GetRecord" identifier="oai:www.nm-dev.org:ITEM.001" metadataPrefix="oai_dc">http://www.nm-dev.org/ca/service.php/OAI/dc</request> <GetRecord/> <error code="idDoesNotExist">Identifier is invalid</error> </OAI-PMH> I did also dig in the code of providence / app / lib / ca / Service / OAIPMHService.php. I see the following. The code is using the primary key of the ca_objects table in stead of idno (my ITEM.001 value). I suppose this is not the intent. ...??? So, I changed my URL request to pass the primary key value of this object (which I sought in the ca_objects table) instead of my object identifier value ITEM.001. I got no response back. Blank response. It is as if the code execution breaks.

I also discovered a DataExporter class which is instantiated. I am not sure what is the meaning of this piece of functionality? Does this has to do with Import/Export mapping screen under Menu item Administration of Providence?

Please I need some help here? There is no (almost nothing) documentation on OAI PMH Provider and Import/Export mapping. I dont understand this part of Providence functionality.

Thanks a lot! Urs

seth

Hi,

There needs to be an export mapping in place that establishes correspondence between metadata elements in CA and the metadata format(s) to be served over OAI. It's probably dying because there's no mapping.

We apologize for the lack of documentation. The import/export mapping editor in the current version is sorely lacking and we have been working on a replacement. Once that's done we'll update the documentation. Look for it in the next few months.

For now, you can try creating a mapping using the current crappy editor. What are you looking to serve via OAI? DublinCore?

uwieske

Hi Seth,

Thanks for your reply.

Well, having read your reply I realize that the mapping part is my problem. To tell you the truth, I don't really understand how to use the Mapping screen, i.e. what and how should I use the fields (because of little documentation). Can you point me through the mapping steps? I just need some understanding what to fill in and the format of my input. I read something like "ca_object.idno = /dc:identifier" in the Wiki. I filled this in the "Mapping code" field but unsuccessful results. Then I tried the unit fields, which made more sense to me if I judge by their corresponding field label (help)text. Again unsuccessful results. Basically I don't have a clue what and how to fill the fields without guidance of documentation. So If you could help me through the mapping process I would be grateful. :-)

We will be using CA Providence as our Digital Asset Management (repository) and use Omeka as publishing/presentation component to online public. We want to use Dublin Core for data interchange between CA and Omeka which will mirror a subset of what's in the CA repository. That's why I want the OAI-PMH Provider of CA to work. (Also, the Mapping part of cource).

I appreciate any help with the mapping. If this does not work, we are obliged to look for another digital collection repository. We really like CA because of its model structure which can accommodate almost any structure and besides cataloging it also has support for Loan, Movement etc etc.

Thanks! Urs

seth

Hi,

To create a mapping you'll need to create "units" for each field or group of fields (Eg. forename, surname, middlename for an entity) you want to map. Each unit contains one or more mappings between CA bundles (aka fields) and targets in the output format – OAI_DC in this case.

The CollectiveAccess side of the mappings need to have bundle specifiers in the form described here: http://docs.collectiveaccess.org/wiki/API:Getting_Data#Bundle_specifiers

The OAI_DC side of the mappings needs a simple tag path, in the form /dc:title for title.

The base values for each side were intended as shortcut prefixes that are simply prepended onto the mapped values to create the full CA bundle specifiers and tag paths.

If you set the base for the CA side to ca_objects and the mapping to preferred_labels.name, then the bundle specifier ca_objects.preferred_labels.name is what will be used. Similarly, if the prefix for the OAI_DC tag path is "/" and the mapped tag is dc:title, then /dc:title will be used in the mapping.

If this all sounds confusing and rather ill-conceived that's because, well, it is. We've been working on designs for a replacement UI that is visual and drag-and-drop. It is something we want to do soon, but there's still a lot of work to do.

Please give the current UI a try and if it proves unworkable (quite likely) perhaps we can rig up a way for you to specify the mapping in a configuration file, as a bridge until the new UI is available.

seth