ProApproX – An approximation query processor over probabilistic XML
Introduction
ProApproX is a Java query engine for probabilistic XML documents, developed by Asma Souihli during her PhD. It evaluates XQuery queries over probabilistic XML by translating them to queries over a regular XML database, computing the probabilistic lineage of query results as propositional formulas, and choosing between exact and approximate probability computation algorithms based on a cost model. It is described in the following publications:
- ProApproX: A Lightweight Approximation Query Processor over Probabilistic Trees (SIGMOD 2011 demonstration, first version);
- Demonstrating ProApproX 2.0: A Predictive Query Engine for Probabilistic XML (CIKM 2012 demonstration, the version distributed here);
- Optimizing Approximations of DNF Query Lineage in Probabilistic XML (ICDE 2013), for the underlying algorithms.
Download
The code is provided as is, as it was left in 2012, for research
purposes; it was originally distributed from Asma Souihli's Web page at
Télécom ParisTech, which no longer exists. The archive contains the
Java sources (a NetBeans project), the XQuery parser, a built
ProApproX.jar, the third-party libraries it depends on
(BaseX, Saxon, Xalan, Java 3D…), the probabilistic versions of the
Mondial dataset used in the demonstrations, and example queries.