minrva
Web Services

Each of the Android modules are lightweight. They receive all data through wi-fi connection. Thus, we have set up web apps to provide data feeds in the form of JSON (or whichever other format of data was requested by the client).

The flexibility of our feeds and the ability to set up a RESTful service is provided by a technology called Jersey. Jersey automatically converts Java objects into JSON, XML, JSON-P etc. Although, I would like to note that YQL is a great service for those who do not have the ability to change the format of the data feed.

A typical request to one of the services that we have set up will look like this:

http://localhost:8080/MinrvaServices/LoanableTech

A typical return feed is below (abbreviated): 

[ { “bibId” : “6399958”,

    “count” : “2”,

    “name” : “MacBook Pro “,

    “rank” : “1”

  },

  { “bibId” : “6426672”,

    “count” : “4”,

    “name” : “Macbook Charger “,

    “rank” : “2”

  },

  { “bibId” : “5241959”,

    “count” : “12”,

    “name” : “Laptop “,

    “rank” : “3”

  }

]

LoanableTech Module

LoanableTech module orders the loanable technology by popularity (last 30 days) and then interprets the item statuses in order to determine how many items are still remaining.

Display Module

Display module uses Minrva’s shared Item object to display information about the current book loaded into the system.

Scanner Module

Scanner module acts as an interface for book input by scanning Cod-a-bar. The app was built using the popular ZXing API.

Suggestions Module

Suggestions module currently mooches off of Amazon’s suggestion algorithms, but will soon be replaced by local analysis of library transactions (after some data mining…).

VuFind Module

VuFind module hooks into the Integrated Library System’s (ILS) VuFind search (Solr powered XML search engine).


Dataflow:

1) Populate suggestions menu based on current book loaded into Minrva. 
2) Get search query either through user free-text input or drop down menu selection. 
3) Make query url friendly
4) Send query to RESTful service
5) RESTful service scrapes VuFind results
6) RESTful service sends JSON to mobile device for parsing

See http://m.carli.illinois.edu/vf-uiu/ for ILS VuFind implementation.

Wayfinder Module

   

Wayfinder module prototype has panning capability and customizable maps. 

Dataflow:

1) The module sends a call number to the Wayfinder Minrva web service 
2) The RESTful service parses the call number and searches a local MySQL database for map coordinates.
3) The service returns JSON to the mobile device thanks to Jersey. 
4) The mobile device draws the coordinates on a map that it has also downloaded from Minrva’s image server. 

Optical Character Recognition, Linked Data, and Artificial Intelligence

The minrva app is coming to app stores soon! As we wind up the first semester of the modular project and move on to OCR research, Nate and I are hiring people to help with Artificial Intelligence problems in library mobile apps that will include connecting an OCR engine with a suggestion database for student research. Also - we have funding on a related formative evaluation project to create subsets of the library catalog in RDF and run comparative studies on linked library data. We plan to use the RDF models in our machine learning research and development, too — so should be an exciting summer!

Here is a model of the modular design that Minrva is built upon.

Here is a model of the modular design that Minrva is built upon.


at project minrva we experiment with modular mobile app design for next generation library services. the plan is to develop, share, and collaborate on modules for customization and enhancement to local library user needs. this is our initial feature set/project tasks for the spring 2012 semester:
-search (catalog, other sources)-barcode scanner (scan books in the library)-suggestions-book info (the central state)-augmented reality services (using open CV and OCR)-favorites/checked out titles 

at project minrva we experiment with modular mobile app design for next generation library services. the plan is to develop, share, and collaborate on modules for customization and enhancement to local library user needs. this is our initial feature set/project tasks for the spring 2012 semester:

-search (catalog, other sources)
-barcode scanner (scan books in the library)
-suggestions
-book info (the central state)
-augmented reality services (using open CV and OCR)
-favorites/checked out titles