Developer Area/Import//Export/Interoperability/EPET

From Mahara Wiki

< Developer Area‎ | Import//Export‎ | Interoperability
Revision as of 17:55, 9 May 2011 by Brettwilkins (talk | contribs) (Created page with "ePET was one of the [ Portfolio Interoperability Project members]. <div id="section_1"> === Quick Summary === ePET/Mahara interoperabilit…")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

ePET was one of the Portfolio Interoperability Project members.

Quick Summary

ePET/Mahara interoperability is very poor. Interoperability was last tested on 2009/09/10.

Importing ePET data into Mahara is currently broken due to well-formedness issues with ePETs export, however with those issues fixed Mahara should be able to import a great deal of ePETs information in a sensible manner. This will require much improvement in ePETs exporter, however.

ePET is able to import Mahara exports in a rudimentary fashion (most entries turn into blog posts) - but cannot import any files at all, as the importer can't handle .zip files.


ePET has provided a services to export both Leap2A zip files and just the raw XML. Attached to this page is a sample export. You can sign up for their demo and generate your own export, if you need.

The XML file generated is currently broken in the following ways:

  • Namespaces - the file uses old LEAP namespaces for the predicates and types, these should be updated to the most recent (and final for the 2009-03 version) ones. Instead of LEAP_predicates for xmlns:leap, it should be LEAP2A_predicates. And instead of LEAP_classes for xmlns:leaptype, it should be LEAP2A_types.
  • Blog content, which is entered through a textarea in ePET, is put into the XML file as . This means line breaks are collapsed when imported into Mahara - it's probably better if it's just exported with a content type of text.
  • Blog posts are generated with ids like portfoilo:blog_XXX, but blogs link to them using portfolio_blogXXX which results in no blog posts being imported. In fact, there are many places where links are either broken, or not specified in both directions.

Other issues that are currently reducing the quality of the interoperability:

  • ePET has a files area, but is exporting all of its folders as a selection without selection_type#Folder. As such, Mahara's file plugin is not identifying them as folders to be imported. Some files can be imported, such as those attached to blog posts.
  • In addition, child folders are not claiming they have an is_part_of relationship to their parents. While Mahara's importer is "accidentally" smart enough to deal with this - and Mahara itself shouldn't depend on the relationship being specified both ways - it would aid clarity if the relationship was specified both ways.
  • For some reason, many of the entries in the export have an of "Economics" - if Mahara was strictly checking authors, those entries wouldn't be imported.

... And some good news:

  • Profile information from ePET is being imported into Mahara's profile area, including name, address and phone numbers (in ePET you put this information in via the CV).


As of 2009/04/01, ePET seems to be able to import Mahara exports in a rudimentary fashion.

It crashed when a .zip export was tried (IOError, no such file), but accepted a leap2a.xml file. ePET has a wizard you can use to choose where things are imported to, although the only option given for all entries was to import them as blog posts. Hopefully in future it can be arranged so folders and files are imported into the ePET files area.

Folders seemed to be lost completely by the import, and files - while imported as entries into the blog - didn't show the files themselves as naturally a leap2a.xml file by itself doesn't contain the file contents. Hopefully ePET can fix the problem with zip uploads soon so this can be tested.

ePET doesn't seem to store much personal information. Quite a few of Mahara's profile fields were imported as blog entries.

All of Mahara's resume information was imported as blog entries, although some of it looks like it could map quite nicely to certain ePET fields.