Table of Contents
VuFind Developers Call Minutes: February 5, 2013
Attending: Filipe Bento, Chris Hallberg, Demian Katz, David Maus, Sean Purcell, Al Rykhus, Jon Singletary
1. New/Updated JIRA Tickets
- VUFIND-505 - Part of the “complex checkbox facet” problem has been solved through the addition of some new configuration syntax in VuFind 2.
- VUFIND-598 - Al has posted some improved call number normalization routines.
- VUFIND-738 - This ticket tracks a Zend Framework 2 bug related to HTTP chunking. David confirmed that it can now be closed.
- VUFIND-739 - This ticket (now closed) includes a fix for a Wikipedia-related bug in the Author module.
- VUFIND-741 - This ticket addresses a bug in the routine used for validating module names in the 2.0 install script – the problem is now fixed in Git's master branch.
- VUFIND-742 - This ticket proposes upgrading jQuery for the 2.0 release.
2. VuFind 1.4 Update
VuFind 1.4 was successfully released on schedule. No significant problems have been reported yet, but please try the new release and let us know how it works for you.
3. VuFind 2.0 Update
Demian has also added a new “social statistics” page to the Admin module to report on usage of comments, favorites and tags (since it is often interesting to measure the popularity of these features).
VuFind 2 has been upgraded to Zend Framework 2.1.0. This fixes some bugs and adds important new features, including the long-awaited “render.error” event which will help prevent “white screens of death.”
Demian has refactored the Export-related code so that there are fewer static calls; instead there is an Export tools object in the Service Manager. This makes overrides and testing easier.
Record drivers have been refactored to eliminate calls to \VuFind\Config\Reader; all necessary configurations are now injected via constructor using factory configurations.
Demian has changed some search templates to use separate $this→params and $this→results variables (instead of relying on $this→results→getParams()) to simplify code and prepare for search system refactoring. The UrlQueryHelper class has also been refactored to eliminate a dependency on Results objects (it is now purely Params-based).
We should discuss changes to default settings in 2.0. For example, should Author module continue to use ExpandFacets by default?
David and Demian discussed next steps for search integration. David will prepare a branch that Demian will pull into the main VuFind repo for future development.
4. Linked Data
Apart from a recent mailing list post from David about an online class on the subject, no new development this time around.
5. DataImportHandler Update
Filipe has continued experimenting with the Solr DataImportHandler using an XML file (warning: large) containing the MIT OCW record set. He had to work around some limitations of the XPathEntityProcessor and a problem with the DataImportHandler failing to respect <copyField> directives in the Solr schema. Eventually these problems were overcome – here is an example configuration used to push data through the DataImportHandler after first preprocessing it with some XSLT transforms.
Filipe has also experimented with using Google Reader to pull data from RSS feeds for import. In the OCW example, RSS data is available to supplement the initial raw XML document. There are limitations on how many records can be retrieved in this fashion (1,000 per pack) and the resulting XML requires some preprocessing (Filipe used sed) before it can be cleanly imported through the DataImportHandler.
6. Other Topics?
Demian will be at Code4Lib next week and will try to fit a VuFind breakout session into the schedule.
The next call will be Tuesday, February 19, 2013 at 10am Eastern Standard Time (15:00 GMT).