This document describes a system for visualizing content stored in metadata stores. It includes a query abstraction layer that separates querying logic from application logic and allows querying metadata stores via web services. Query results are transformed and then visualized using tools like Prefuse, JFreeChart and Swing. The overall system includes components for metadata storage, query transport, query abstraction, result transformation, and visualization, with the prototype called Gview wrapping these components together.
1. Visualizing Metadata Store Content
Automated Learning Group &
Data-Intensive Technologies and Applications
National Center for Supercomputing Applications
University of Illinois at Urbana-Champaign
2. Motivation
Use RDF store for massive storage of processed data and
•
analytical results
Query information out of the store via a conceptual
•
abstraction
Separate querying logic from application logic
•
Visualize content stored on RDF stores
•
Have a working prototype showing the basic
•
functionalities
Visualizing Metadata Stores Contents ALG & DITA 2
3. The Overall Idea
Visualization Layer
Application Logic
Result Transformation Layer
Query Proxy Layer
Query Transport and Interaction Layer
Management & Deploy
Query Abstraction Layer
Query Logic Layer
Visualizing Metadata Stores Contents ALG & DITA 3
5. The Components and the Overall Idea
Prefuse & co.
Visualization Layer
Gview
Application Logic
SAL
Result Transformation Layer
SAL
Query Proxy Layer
SOAP + XML
Query Transport and Interaction Layer
Management & Deploy
Descriptors
Query Abstraction Layer
iTQL
Query Logic Layer
Mulgara
Visualizing Metadata Stores Contents ALG & DITA 5
6. Metadata Stores and Query Abstraction
Mulgara provide XSL/XLST facilities
•
Other implementations provide similar mechanisms Jena/Joseki
•
Allow to define a query unit and publish it as a web service
•
Single queries (iTQL)
Complex query logic (iTQL+XSLT)
Handled on a web app
Runs on SOAP
•
Allow reflection on (RDF based):
•
Descriptions
Parameters (We added type casting)
Abstract return types (Our modification)
Query becomes and abstraction entity
•
Separated from the application logic
Managed and deployed on demand
Visualizing Metadata Stores Contents ALG & DITA 6
10. Transport
• XML response
• We can apply XSLTs to
the results on the XSL if
needed
• XML gets ship back after
the soap request
model =
•
rmi://adelle7.ncsa.uiuc.edu/server1#v
ast07_06_apr_2007
modellucene =
•
rmi://adelle7.ncsa.uiuc.edu/server1#v
ast07_03_apr_07-ft
expression =
•
+cat*
Visualizing Metadata Stores Contents ALG & DITA 10
11. Query Abstraction Proxy (SAL)
• Basic layer to provide:
Programmatic abstractions for queries
A simple API to reflect content on stores
Self contained query objects
Transformers to apply to the results obtained
Contexts
Reflector
Queries
Application
Result Transformer
Visualizing Metadata Stores Contents ALG & DITA 11
13. Visualization Layer
• SAL encapsulates the XML results in transformable
Answer objects
• Answers can be transformed to feed visualization
packages
• For instance answers can be transformed to:
GraphML
TreeML
…
• Extensible via XSLT
• Prefuse, JFreeChart, Swing
Visualizing Metadata Stores Contents ALG & DITA 13
14. Components Wrapped in a Prototype (Gview)
Visualizing Metadata Stores Contents ALG & DITA 14
15. Components Wrapped in a Prototype (Gview)
Visualizing Metadata Stores Contents ALG & DITA 15
16. Components Wrapped in a Prototype (Gview)
Visualizing Metadata Stores Contents ALG & DITA 16
17. Visualizing Metadata Store Content
Automated Learning Group &
Data-Intensive Technologies and Applications
National Center for Supercomputing Applications
University of Illinois at Urbana-Champaign