oreprovider - oreprovider: an OAI-ORE add-on service for Fedora

Introduction

oreprovider is distributed as a Java web application (a war-file). Once downloaded this should be placed in your Tomcat's webapp/ directory (or any other servlet container you may be using) and renamed oreprovider.war (stripping away the information of the version). This will then be extracted to a directory called 'oreprovider/' (you might have to restart Tomcat to achieve this). The application can now be reached at

http://<host>:<port>/oreprovider/rem/<identifier>/

where

<host> is your server running Tomcat.
<port> is the port that Tomcat is running on (usually 8080).
<identifier> is the identifier you want to retrieve a Resource Map for.

At the moment oreprovider only returns Resource Maps serialized as Atom feeds.

Settings

There are some settings that can be customized for oreprovider. The settings file can be found in

oreprovider/WEB-INF/classes/ore.properties

The settings that can be applied there are:

fedora.server - the server and port where Fedora is running.
Example: fedora.server = http://magasin.kb.se:8080/fedora/
Default: http://localhost:8080/fedora/
rem.creator - this value will be set as the creator in the Resource Maps created.
Default: ORE Provider
rem.rights - this value will be set as the rights in the Resource Maps created.
Default: http://creativecommons.org/licenses/by-nc-sa/3.0/
rem.autocreation - if set to true, no annotations have to be present in RELS-EXT. Instead all datastreams for the Fedora object matching the given id will be added as Aggregated Resources. See more about the issue of annotations versus autocreation below.
Example: rem.autocreation = true
Default: false
rem.autocreation.filter - a whitespace separated list of datastreams to automatically filter for when using autocreation . Only datastreams matching a name given here will be included as an Aggregated Resource. If not set, all datastreams for a Fedora object will be added when using autocreation.
Example: rem.autocreation.filter = DC MODS WEBIMAGE
Default: Not set, i.e. all datastreams will be included.
type.<RESOURCE> - this will be the type for an Aggregated Resource matching <RESOURCE>, where <RESOURCE> is a datastream or a disseminator. Multiple values for a <RESOURCE> can be provided as a whitespace separated list.
Example: type.DC = info:eu-repo/semantics/descriptiveMetadata
Example: type.WEBIMAGE = http://purl.org/dc/dcmitype/Image
Example: type.bdef\:text/getDocument = http://purl.org/dc/dcmitype/Text http://purl.org/eprint/type/JournalArticle
[Note: if you want to add a type for a disseminator you have to remember to escape any ':' characters in the key correctly. ]

How Resource Maps are created from Fedora objects

To let oreprovider be able to create Resource Maps for digital objects in Fedora two different modes can be used.

The first way is to modify the RELS-EXT for the Fedora object(s) you want to expose as a Resource Map, annotating exactly which datastreams and disseminators that should be included (possibly from multiple Fedora objects).
Another way is to use the autocreation mode and just let all datastreams for a Fedora object be added to the Resource Map.

The pros of the first mode is that it is very flexible and precise and you have exact control over what will be included in the Resource Maps. You can also include disseminators and are not restricted to only datastreams. Here you are able to have many different Content Models in the repository and treat all of them different. Another big advantage is that you can create Resource Maps for Aggregations that span several Fedora objects (which for example would be the case for a book with a Fedora object for each image of a page).

On the other hand the second mode is really simple to use and no alterations has to be done to an existing repository. So if there is an one-to-one mapping between all Fedora objects and the Resource Maps you want to create, this is the ideal mode. All you have to do to expose your objects as Resource Maps are to download oreprovider.war to the correct location, then modify the properties to point at your Fedora server and set rem.autocreation to true. And that's it.

Using autocreation

This is really simple. Just set

rem.autocreation = true

in the ore.properties file. This will then add all datastreams for a Fedora object with <pid> to the Resource Map, when a call is made to http://localhost:8080/oreprovider/rem/<pid>/ .

[Note: when this mode is enabled all annotations set in the RELS-EXT will be ignored. ]

If you want to have some control over which datastreams should be included you can additionally set rem.autocreation.filter to a whitespace separated list of datastreams that should be included. Any datastreams not in the list will be ignored. If this property is not set, all datastreams will be added (as is the default).

Using annotations in RELS-EXT

Annotating digital objects as Resource Maps

To let oreprovider know that an object in Fedora should be part of a Resource Map you have to say so in the RELS-EXT for the object. This is done by adding the following

<ore:resourceMap xmlns:ore="http://www.kb.se/dl/xml/ore#">ore:<IDENTIFIER></ore:resourceMap>

where <IDENTIFIER> should be the identifier that you want to let the Resource Map be known as. (This could very well be the pid of the object in Fedora.)

An example could be:

<ore:resourceMap xmlns:ore="http://www.kb.se/dl/xml/ore#">ore:kb:9051></ore:resourceMap>

Annotating datastreams and disseminators as Aggregated Resources

Besides doing the above you also have to tell oreprovider which datastreams or disseminators that should make up the Aggregated Resources of the Resource Map. This is done by adding the following

<ore:resource xmlns:ore="http://www.kb.se/dl/xml/ore#"><DATASTREAM></ore:resource>

to the RELS-EXT, where <DATASTREAM> should be the datastream or disseminator in the object that should be added as an Aggregated Resource. (It is this name that type.RESOURCE in the configuration file looks for when adding a type to the Aggregated Resource.)

Of course multiple datastreams or disseminators can be added. An example could look like:

<ore:resource xmlns:ore="http://www.kb.se/dl/xml/ore#">DC></ore:resource>
<ore:resource xmlns:ore="http://www.kb.se/dl/xml/ore#">MODS</ore:resource>
<ore:resource xmlns:ore="http://www.kb.se/dl/xml/ore#">bdef:images/getWebImage</ore:resource>

Resource Maps consisting of multiple Fedora objects

Sometimes you might have a digital object that is built from multiple Fedora objects. This could be the case if you have scanned a book and have one image per page. Then the digital object would consist of the Fedora object for the book as a whole (where you might keep bibliographic metadata), plus all the Fedora objects for the individual pages.

To create a Resource Map for such a digital object you would have to modify the RELS-EXT for the book as a whole, as well as the RELS-EXT for each individual page. The only difference from simple case is that all the RELS-EXTs should have an <ore:resourceMap> with the same identifier. (This could preferably be the pid of the Fedora object for the book as a whole.) Then each RELS-EXT would also list the datastsreams and disseminators that should be included from that specific Fedora object.

A full example

Listed below is an example of the RELS-EXT for a Fedora object consisting of three datastreams, DC, MODS and PDF. It also has a disseminator that will give you the text of the PDF as plain text, called bdef:text/getText .

<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
         xmlns:ore="http://www.kb.se/dl/xml/ore#">
  <rdf:Description rdf:about="info:fedora/kb:1628">
    <ore:resourceMap>ore:kb:1628</ore:resourceMap>
    <ore:resource>DC></ore:resource>
    <ore:resource>MODS</ore:resource>
    <ore:resource>PDF</ore:resource>
    <ore:resource>bdef:text/getText</ore:resource>
  </rdf:Description>
</rdf:RDF>

This Resource Map could now be retrieved from http://localhost:8080/oreprovider/rem/ore:kb:1628/ and would have Aggregate Resources for all of the above.

If you want the Aggregated Resources for the DC and MODS to say that there types are descriptive metadata you should add the following to the configuration file:

type.DC = info:eu-repo/semantics/descriptiveMetadata
type.MODS = info:eu-repo/semantics/descriptiveMetadata

Documentation