Processes for Generating Asset Actions for the Aquifer American Social History Online Portal

 

The following is a brief outline of how Asset Actions (AAs) are generated for the Aquifer American Social History Online (ASHO) portal.  This is the process for dynamically generating Asset Actions for records that do not already have Asset Actions from the original provider.  If the original data provider already has their own Asset Action packages for their records, those AA packages will be used instead of the dynamically generated ones which are described below.  This process is primarily for image type resources, but it will generate AAs for non-image resources if it is able to generate a thumbnail representation of the resource.

  1. A Ruby script is run on the ASHO portal that submits OAI records to the OAI ThumbGrabber Web Service at the Grainger Engineering Library Information Center.  http://gita.grainger.uiuc.edu/registry/thumbnails/ .  For details on the ThumbGrabber application, refer to the D-Lib magazine article Automated Capture of Thumbnails and Thumbshots for Use by Metadata Aggregation Services, http://dx.doi.org/10.1045/january2006-foulonneau .  In the process of creating a thumbnail the ThumbGrabber service stores data into a relational database that can be used for the generated Asset Actions.  These data include the OAI Identifier, the DC Identifier URL that was dereferenced, the URL of the final web page containing the image that was captured, the URL of the source image, plus the images original size and date stamp, and the alt and title attributes of the img tag, if any.
     
  2. The ASHO portal uses the UnAPI protocol ( http://unapi.info/ ) to expose Asset Actions.  The current UnAPI service exposes MODS (mods), old-style  Asset Actions (aa), and the new ORE Atom serialization for Asset Actions (ore).
     
  3. The UnAPI service takes an identifier from the portal database, looks up the corresponding OAI identifier, and creates an HTTP redirect to the ThumbGrabber service’s Asset Action generator, for example:

    http://nocoma.grainger.uiuc.edu/unapi?id=251639&format=aa  
    is redirected to http://gita.grainger.uiuc.edu/registry/thumbnails/aa.asp?identifier=oai%3Aoai.dlib.indiana.edu%3Aarchives%2Fcushman%2FP04331
     
  4. The returned Asset Action package refers to various other services that are also handled by the ThumbGrabber service’s Asset Action generator, for example:
    1. The getPreview and getThumbnail actions refer to this service:
        http://gita.grainger.uiuc.edu/registry/thumbnails/thumbnail.asp?identifier=[oai_id ]
      which returns the thumbnail image that was generated by the ThumbGrabber.
    2. The getLabel action refers to this service:
         http://gita.grainger.uiuc.edu/registry/thumbnails/label.asp?identifier=[oai_id]
      First, the ThumbGrabber database is queried for a title or description that might have been set when the thumbnail was captured.  These are typically from the alt or title attributes of the img tag of the captured image.  If there is nothing in the database then the service uses the Grainger OAI Registry to resolve the OAI Identifier to an OAI Base URL which is then used to fetch the oai_dc metadata record directly from the source OAI data provider from which the first one of the DC title, description, or subject is extracted and returned as a plain text string.
    3. The getDCRecord action refers to this service:
         http://gita.grainger.uiuc.edu/registry/thumbnails/dc.asp?identifier=[oai_id]
      which uses the Grainger OAI Registry to resolve the OAI Identifier to an OAI Base URL which is then used to fetch the oai_dc metadata record directly from the source OAI data provider and return it to the requester.
    4. The getWebView and getDynamicView actions refer to this service:
        http://gita.grainger.uiuc.edu/registry/thumbnails/dyn.asp?identifier=[oai_id]
      which displays either the original web page from which the thumbnail was derived, or if that page is not available, it creates a default web page from the OAI DC record and the thumbnail image in the database.
    5. The getDefaultContent action refers to this service:
        http://gita.grainger.uiuc.edu/registry/thumbnails/default.asp?identifier=[oai_id]
      which just redirects to the original web page from which the thumbnail was generated.
    6. The getScreenSize action refers to this service:
        http://gita.grainger.uiuc.edu/registry/thumbnails/screen.asp?identifier=[oai_id]
      and the getMaxSize action refers to this service:
        http://gita.grainger.uiuc.edu/registry/thumbnails/max.asp?identifier=[oai_id]
      Both of these services operate the same.  If the URL in the DC metadata record pointed directly to an image, the user will be redirected to that image.  However, if the URL in the DC metadata record pointed to an HTML page, the user will be redirected to the thumbnail image that was derived from that HTML.  Unfortunately, what this means in practical terms is that the getThumbnail , getScreenSize , and getMaxSize actions often resolve to the exact same image which is just the thumbnail image.  The getScreenSize and getMaxSize will always resolve to the same image.

The process of generating the ORE/Atom based Asset Action packages is very similar, using the exact same services as described above.  The primary difference is that more of the metadata is embedded directly in the Atom feed.