EPData XML Representation

Warning! This is just a place-holder while I investigate whether its practical to register a MIME-type.

Specification

EPData XML is a recursive structure where an object is represented by a root element named as the dataset-class e.g. eprint becomes <eprint>. Each object may contain zero or more elements representing a single metadata field each e.g. eprint.title becomes <title>. A metafield may be of type Subobject in which case it will contain embedded objects.

Multiple metadata values are represented by zero or more <item> elements directly below the metadata field element.

<DATASETID + 's'>
 <DATASETID>
  <METAFIELDID>
   <item>{basic XML type or complex value}</item>
  </METAFIELDID>
  <METAFIELDID>
   <DATASETID>...</DATASETID>
  </METAFIELDID>
 </DATASETID>
</DATASETID + 's'>

Note: we may change the root element to be a generic term to allow a mixture of dataset-classes.

Example

 Content-Type: application/vnd.eprints.data+xml; charset="utf-8"

 <?xml version='1.0' encoding='utf-8'?>
 <eprints xmlns='http://eprints.org/ep2/data/2.0'>
  <eprint id='http://yomiko.ecs.soton.ac.uk:8080/id/eprint/10'>
    <eprintid>10</eprintid>
    <rev_number>18</rev_number>
    <eprint_status>archive</eprint_status>
    <dir>disk0/00/00/00/10</dir>
    <datestamp>2010-08-10 15:34:04</datestamp>
    <lastmod>2011-05-17 15:16:20</lastmod>
    <status_changed>2011-05-17 13:44:44</status_changed>
    <type>teaching_resource</type>
    <metadata_visibility>show</metadata_visibility>
    <item_issues_count>0</item_issues_count>
    <userid>1</userid>
    <creators>
      <item>
        <name>
          <family>Ποια</family>
          <given>Debitis</given>
        </name>
      </item>
      <item>
        <name>
          <family>Dolorem</family>
          <given>Blénken</given>
        </name>
      </item>
      <item>
        <name>
          <family>Колёса</family>
          <given>As</given>
        </name>
      </item>
      <item>
        <name>
          <family>Βάζοντας</family>
          <given>मुश्किल</given>
        </name>
      </item>
    </creators>
    <title>HIC HIC Koum et τοπικές hic</title>
    <ispublished>pub</ispublished>
    <subjects>
      <item>PZ</item>
      <item>AZ</item>
      <item>PM</item>
      <item>LB2361</item>
    </subjects>
    <full_text_status>public</full_text_status>
    <abstract>" '^M
^M
\ /^M
^M
Aut do och Чем modi. Φακέλους et Scholl itaque αρέσει rerum tenetur हुएआदि. Similque команда voluptas или quidem schaddreg voluptatum ίδιο. Nulla animi νέα συνηθίζουν d'Loft inventore quis.^M
^M
Expedita praesentium σημαίνει Нее rei sed et стратегические. Doloribus alias प्रा ea reiciendis Frot. Molestiae αντιλήφθηκαν μου enim. Tenetur πως отнимет во. Nihil voluptatum nobis dolorum laudantium cum Jo τι.</abstract>
    <date>1961</date>
    <publication>आंतरजाल ποια De impedit och आंतरजाल Чем नयेलिए nesciunt.</publicat>onn
    <refereed>TRUE</refereed>
  </eprint>
 </eprints>
 

Example with Files Embedded

<?xml version='1.0' encoding='utf-8'?>
<eprints xmlns='http://eprints.org/ep2/data/2.0'>
  <eprint id='http://yomiko.ecs.soton.ac.uk:8080/id/eprint/102'>
    <eprintid>102</eprintid>
    <rev_number>2</rev_number>
    <documents>
      <document id='http://yomiko.ecs.soton.ac.uk:8080/id/document/808'>
        <docid>808</docid>
        <rev_number>1</rev_number>
        <files>
          <file id='http://yomiko.ecs.soton.ac.uk:8080/id/file/919'>
            <fileid>919</fileid>
            <datasetid>document</datasetid>
            <objectid>808</objectid>
            <filename>metadata_test.docx</filename>
            <mime_type>application/msword</mime_type>
            <hash>935033b54ca1f3b439deaabb8ec6dba8</hash>
            <hash_type>MD5</hash_type>
            <filesize>76927</filesize>
            <mtime>2011-06-17 10:17:39</mtime>
            <url>http://yomiko.ecs.soton.ac.uk:8080/102/1/metadata_test.docx</url>
            <data encoding='base64'>UEsDBBQABgAIAAAAIQC7VeA/CAIAABkMAAATAAgCW0NvbnRlbnRfVHlwZXNdLnhtbCCiBAIooAAC
...
U2V0dGluZ3MueG1sUEsFBgAAAAAlACUAAwoAAGYiAQAAAA==
</data>
          </file>
        </files>
        <eprintid>102</eprintid>
        <pos>1</pos>
        <placement>1</placement>
        <format>application/msword</format>
        <language>en</language>
        <security>public</security>
        <main>metadata_test.docx</main>
      </document>
    </documents>
    <eprint_status>inbox</eprint_status>
    <dir>disk0/00/00/01/02</dir>
    <lastmod>2011-06-17 10:17:39</lastmod>
    <status_changed>2011-06-17 10:17:11</status_changed>
    <type>article</type>
    <metadata_visibility>show</metadata_visibility>
    <userid>1</userid>
    <creators>
      <item>
        <name>
          <family>Tarrant</family>
          <given>David</given>
        </name>
        <id>dct05r@ecs.soton.ac.uk</id>
      </item>
      <item>
        <name>
          <family>Brody</family>
          <given>Tim</given>
        </name>
        <id>tbb2@ecs.soton.ac.uk</id>
      </item>
      <item>
        <name>
          <family>Carr</family>
          <given>Les</given>
        </name>
        <id>lac@ecs.soton.ac.uk</id>
      </item>
    </creators>
    <title>From the Desktop to the Cloud: Leveraging Hybrid Storage Architectures in your Repository</title>
    <full_text_status>restricted</full_text_status>
    <keywords>Computer Science, Repositories, Hybrid Storage, Cloud Storage, Storage Abstraction</keywords>
    <abstract>Repositories collect and manage data holdings using a storage device. Mainly this has been a local file system, but recently attempts have been made at using open storage products and cloud storage solutions, such as Sun's Honeycomb and Amazon S3 respectively. Each of these solutions has their own pros and cons but there are advantages in adopting a hybrid model for repository storage, combining the relative strengths of each one in a policy-determined model. In this paper we present an implementation of a repository storage layer which can dynamically handle and manage a hybrid storage system.</abstract>
    <date>2011-05-28</date>
  </eprint>
</eprints>

IETF-Registration Draft

Name : Tim Brody

Email : tdb2@ecs.soton.ac.uk

MIME media type name : Application

MIME subtype name : Vendor Tree - vnd.eprints.data+xml

Required parameters : None

Optional parameters : 
charset

Same as charset parameter of application/xml as specified in RFC 3023.


files

Indicates that file content has been embedded in the response XML. This is used during content-negotiation to allow the client to request file content be included. If the client does not have permission to access file content this parameter will be ignored.

"base64" is the only supported value and indicates that file content has been embedded using base64 encoded. If the parameter is empty or any other value the parameter meaning is undefined (in a response) or should be ignored (during content negotiation).


Encoding considerations : binary


Security considerations : 

In addition to those of application/xml as specified in RFC 3023, section 10 the following considerations apply:

No executable or active content is defined.

No integrity features are provided by the media type, with the exception that where files are embedded and checksums are provided the ingesting service should verify the decoded file content against its checksum before further processing those files.

No explicit privacy features are provided by the media type but may be provided on an implementation-specific basis.

Embedded files may contain executable or malicious content. If file content is decoded care should be taken before any further processing or publication by e.g. applying a virus-checker.

Identifiers may be included that either intentionally or unintentionally conflict with existing identifiers in a consuming system. Care must be taken that existing records are not unintentionally overwritten by either assigning new identifiers on ingest or ensuring that ingested data is trusted to update existing records.

References to files by URL may be included. During ingest the system may choose to retrieve these files and include them in the resulting record. If the URL is a Web location (http: or https:) the system should exercise caution to avoid being poisoned by large files or by untrusted users as a means of circumventing host-based restrictions. If the URL points to a local file-system location (or other scheme) this must be ignored for any untrusted sources.


Interoperability considerations : 


Published specification : 
http://wiki.eprints.org/EPData_XML_Representation

Applications which use this media : 
EPrints http://www.eprints.org/

Additional information :

1. Magic number(s) : None
2. File extension(s) : .xml
3. Macintosh file type code : "TEXT"
4. Object Identifiers: None



Person to contact for further information :

1. Name : Tim Brody
2. Email : tdb2@ecs.soton.ac.uk

Intended usage : Limited Use
XML serialisation of EPrints Data (or "EPData") for the import/export of the complete record. This is used e.g. for ingesting the results of XSL
transforms from standardised XML formats. The mime-type is necessary to
support correct content-type negotiation when using the EPrints REST
interface although the client will require knowledge of the instance's
database scheme.

Author/Change controller : EPrints.org http://www.eprints.org/

Tim Brody <tdb2@ecs.soton.ac.uk>