Participants

Eddie Shin
John Scofield
Joe Gilbert
Molly Pickral
Bess Sadler
Matt Zumwalt

Rick Johnson (notetaker)

Agenda

Matt's work done on plugin
Bess: Decision on how to run hydra-jetty dev and test environments related to fixtures
Rick: usage of hasPart/ispartof for file objects versus link objects
Eddie: Closing a bunch of Fedora tickets. Fedora 3.5 imminent.

Action Items:

Bess will try to get separate hydra-jettys implemented (instead of currently 2 solr cores)
Eddie will setup Fedora 3.5 version of jetty to test before doing any upgrades
Rick will keep going the way he is with adding ExternalAssets to hydra_repository and look at changes to ActiveFedora as well to see if any custom predicates should be used

Next week:

Moderator: Matt Zumwalt
Notetaker: Julie Meloni

Notes:

Matt work on plugin update: Focusing on hydra-head plugin. Documentation on how to test plugins was on Rails 3, but figured out how to test on Rails 2, but kind of tricky. Should be tested same as Blacklight. Trying to imitate that. Instead of referencing a gem, it references a sibling. Added something to environment.rb to reference sibling directories that should be used instead of a gem.

Need to always put plugin under test into a separate directory and then have test.

Matt document this further for others?

Temporary measure until can test in Rails 3.

Rspec tests are running, looking at Cucumber next

Will have Hydra-Head plugin hopefully ready sometime next week

Will run installation process by committers next week.

Bess dev and test with hydra-jetty: Either have two fedora/solr pairs (one fedora for dev only, one fedora for test only), or have multiple jetty instances, one for test and one for dev instead of multiple solr cores. Eddie unsure about two fedoras/ two solrs option. Consensus seems to be option 2.

Rick and usage of isPartOf for managed and external datastreams: Rick adding ExternalAsset model handling to hydra_repository/hydra-head plugin to be able to add external datastreams in a similar way to uploading a file. So that means filing same atomic modelling patterns of having an external datastream in a child object of the asset you are adding the external asset to (mirroring behavior of adding/removing file assets). Rick not sure if should be reusing isPartOf relationship since two lists could get mixed when handling them in a Hydra head. Rick is thinking of differentiating two lists using solr queries where to get file_objects it also queries for parts children that have FileAsset and for external objects queries for objects that have ExternalAsset model to tell them apart. Instead should it be using a different predicate to differentiate the two?

Matt thinks no if they should both be treated as files whether local or external. What about redirected however?

John brought up code that he has added to ActiveFedora that will be released soon to be able to have custom predicates using a yaml mapping file. Predicate mappings for ActiveFedora were hard-coded in Semantic Node and moved to a config file and can configure predicate mappings yaml file. If it didn't have a predicate mapping that it understood it would ignore it. So, perhaps there should be RedirectedAsset's that use a different predicate?

John: Solrizer was failing if running without facets and webrat gems. John changed those in Solrizer

Eddie: With Fedora 3.5 coming out how do we deal with this in terms of active development and do we do separate releases of Hydra for different Fedora versions? What versions of Fedora are institutions standardizing on? Most groups say any upgrades will happen slowly and cannot upgrade immediately

Fedora 3.5 has a lot of fixes and internal Fedora code changes like sql and prepared statement changes. Also, REST API responses were invalid. Eddie does not expect any backwards incompatitabilty with 3.5 though.

Big change in Fedora 3.5 is FESL is now really usable and could use to implement policy and rightsMetadata integrations between Hydra and Fedora.

Browser not supported