Hydra Feature Inventory

Feature Inventory Grid  <<<----  ACTION: Fill this in with your project information.

DAMSPAS - Fedora 3

• Never had a GUI for complex objects

• Streaming server, requires login, passes path/ip/timestamp encrypted

• Shib login or campus IP access

• Auth completely from SOLR

• Serve big files without hitting fedora. Chunked I/O

• Excel spreadsheet maps to RDF, or to MODS

• JHOVE and FITS

• Premise (XML schema) access controls

• ActiveMQ w/ Solrizer

• Serving big files from fedora repo w/ Chunked I/O in controller

• Plan to use Sufia for self-deposit
  
  

Agreed it is silly to write object derivatives back to repo.

Hydra 6 / ActiveFedora 7 / BlackLight 4 with the purpose of replacing ContentDM:

• Basic Hydra access controls. Groups from IP range (or by user)

• Ingest features:

  • Form for flat metadata 

  • Form input templates

  • Review process workflow

  • BagIt packaging

  • Convert ContentDM to bags and ingest

  • Auto-complete from QuestionAuthority and other sources

• Working to do complex data RDF:List

• Use Open Library PDF viewer: zoom, page and added full text search

• Images: Open SeaDragon & IIP, pyramidal TIFF

• Looking forward to IIIF

• Generate audio & video, still working on streaming

• Single item edit, no bulk edit: form can create new item or overwrite existing

• Derivatives from HydraDerivatives

• Opaque persistent ID from NOID

• Want to mint DOIs

• Runs entirely on RDF

Passed on Sufia, b/c no collections and not worried about self-deposit.

• Self deposit w/ Curate: mints DOIs, Orchid IDs

• Editor permissions groups, role based

• Application template/installer

• object types: Works, image, dataset, ETD, article

• Views: custom for datasets, ETD

Everybody wants Spotlight to replace Collection landing page.

Three Hydra-based apps so far: an ETD workflow prototype ("ETDflow"), ArchiveSphere, & ScholarSphere

• Both "spheres" use Sufia, hydra-collections, hydra-derivatives, hydra-ldap

• Vagrant instance for Sufia (https://github.com/projecthydra-labs/sufia-vagrant) available

• Ingest options: single file, multiple files, whole folder hierarchies (in Chrome), all via jQuery file uploader

• Browse-everything support for ingest via Box and Dropbox

• Both self-deposit and proxy deposit (where one user deposits on behalf of another) supported

• NOID IDs surfaced in UI and used in Fedora

• All RDF descriptive metadata, mostly DCMI Terms with some simple DC, FOAF, and RDFS

• ETDflow will use worthwhile to publish finalized ETDs in ScholarSphere

• Permissions are set at the file level, and permissions are a combination of visibility (open, private, logged-in users only) and sharing (with users or groups at either read-only or edit/own level)

• ScholarSphere supports "ownership" transfers between users

• Owners can edit and delete files, add and revert versions, modify visibility/access controls

• Sufia uses a subclass of Hydra RightsMetadata called ParanoidRightsDatastream - adds 3 validation checks: https://github.com/projecthydra/sufia/blob/master/sufia-models/app/models/datastreams/paranoid_rights_datastream.rb

• Uses pre-Questioning Authority tools for harvesting RDF-based controlled vocabularies into the app's RDBMS and connecting these vocabularies to fields in the UI

• On upload, ScholarSphere takes the following steps:

  • Files analyzed by FITS for technical metadata on format, stored in 'characterization' datastream

  • Generate one file for almost all types (a thumbnail)

  • Full-text indexing via Tika (included in Solr instance)

  • Automatic Antivirus checking

  • Application of batch metadata

• Basic A/V viewer (audio.js and video.js), no streaming

• Exposes Mendeley, Zotero, and EndNote export functionality

• Versioning in the UI for adding new versions of files and reverting to old versions

• Uses hydra-ldap to connect with institutional group management service, allow sharing w/ groups

• Includes sitemap gem to generate sitemaps and ping search engines

• Enables batch metadata editing in the UI via hydra-batch-edit

• Prodigious use of background jobs (using Resque backed by Redis)

• ScholarSphere's authN is based on devise, adds a custom strategy for HTTP header authentication (for CoSign, Penn State's institutional SSO solution)

• Single-use links for viewing files and downloading files (expire after one use or 24h)

• Using GAnalytics to gather usage statistics, and displaying graphs of usage and downloads in the public UI

• Embedded metadata galore: schema.org microdata for rich google snippets, google scholar tags, opengraph tags for rich Facebook links, Twitter cards for rich Twitter linking

• Using a number of common code patterns such as concerns (mixin modules), actor classes, and service objects

• Multi-file works

• Fedora 4 w/ large file support

• Mediated Deposit (reviewer/approval workflow)

• Update for wiki for "Gem Marketplace"

TIP: Implement visualization of Hydra-Works as Blacklight plugin