2017-07-13—FileSets WG Meeting

Owned by scossu
Last updated: Jul 14, 2017 by Juliet Hardesty
3 min read

Date and Time

July 13, 2017, 2pm EST

Connection Information

Google Hangouts:  https://hangouts.google.com/hangouts/_/artic.edu/pcdm-filesets

Moderator: scossu

Notetaker: Juliet Hardesty

Attendees

Agenda

Since a while has passed since the last meeting it will be good to use this meeting to get back up to speed with the project in relationship with other developments in Samvera 

  1. Notetaker?
  2. Recap on current status of the project 
    1. Have we had any further thoughts on the project scope or implementation during this time?
    2. Code status 
  3. Routes forward
    1. General topics on future development
    2. Move development to Hyrax?

Minutes

  1. Notetaker? Julie Hardesty
  2. Recap on current status of the project 
    1. Have we had any further thoughts on the project scope or implementation during this time?
      1. Last time we met we were discussing Hydra and Sufia but neither of those are around anymore
      2. Now it’s Samvera and Hyrax
      3. Hyrax files are still being modified - data modeling is still similar to Sufia and Hydra; move to Hyrax and Samvera didn’t change data model
      4. what are we doing with filesets? what do we want to change?
        1. filesets don't allow for enough flexibility right now - need multiple views of same object; need to separate digital objects from files associated with object
        2. current set up is simple and works but needs more complexity and flexibility
        3. Phydo (HydraDAM2) work was supposed to dovetail with this work but Phydo architecture isn’t doing that yet
      5. Right now, Hyrax puts each file into a separate fileset inside of a single work; derivative file is also placed in its own fileset
      6. WGBH - wants filesets to be equivalent to work (both first class citizens) so multiple files can be put into a single fileset without needing a work
      7. Right now fileset only containing one file is cluttering things up
        1. can contain original file, thumbnail, and extracted text so there are some options but not enough
        2. use cases on wiki for Phydo and Art Institute already
    2. https://wiki.dlib.indiana.edu/display/HD2/PHYDO+Data+Model (top diagram)
      1. What is needed - one fileset coming off of object and then create relationships between ffprobe and Preservation or ffprobe and Production, etc
      2. Also need to know specific roles for each file and might need to add roles to PCDM-use ontology but what is there now is fine to start or create our own namespace for new uses
      3. PCDM-use uses rdf:type as the predicate for any PCDM-use class applied to an object 
        1. <fedora:object> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://pcdm.org/use#ExtractedText>, which can also be written: <fedora:object> a <pcdmuse:ExtractedText>
    3. Code status 
      1. Want to provide validator that would give thumbs up or down based on RDF type - not quite done
        1. PCDM use is subclass of RDF type already so this is problematic
        2. ActiveFedora doesn’t allow for changing RDF type once resource is created - still an action item for Adam Wead to discuss with community
        3. Want roles to be changeable and use a different property - RDF type/PCDM-use is a more permanent and less changeable property on the object
      2. PCDM use can’t be changed with UI, any files added to fileset have PCDM use applied
  3. Routes forward
    1. General topics on future development
      1. Create new property for Preferred File Use to get around using RDF Type and make it changeable
        1. Does this need to be a property or should this be class?
        2. putting actual value of "preferred file”  on a fileset is more of a fileset to object relationship than a file use definition
      2. Action items below are still relevant (for Drew, at least)
    2. Move development to Hyrax?
      1. Does code have to go through any migration steps to work in Hyrax? Not sure where this code is located, need to find it and see if any migration needs to occur for what we have now to work with Hyrax
    3. Other thoughts 
      1. Would be great to have UI where there’s a binary file view page where metadata files can be uploaded with that file
      2. Would also be great to have presets for different binary file types to expect different metadata file attachments
    4. Stefano’s action item might be a little advanced for where we are now but he will try to move forward
      1. WGBH use case - provide search for individual files - for example: when was last time fixity check was run? what are all files for which fixity check failed?
    5. Available now: a preservation modeling gem that can be added to Hyrax and preservation events can be searched and viewed in Hyrax

Previous Action Items

Copying relevant items from the previous meeting.

  • Andrew Myers Complete validation PR
  • Andrew Myers post file to file associations question to the tech list
  • Adam Wead summarise the rdf:type issue to the tech list; what problems would the solution cause
  • scossu Create some search user stories to inform UI development 

Action Items