Scientific data integration: wrapping textual documents with a database view mechanism and an XML engine

Abstract

Building a digital library for scientific data requires accessing and manipulating data extracted from flat files or from documents retrieved from the World Wide Web. We present an approach to querying flat files as well as Web data sources through an object database view based on a database system and a wrapper. Generally, a wrapper has two tasks: it first sends a query to the source to retrieve data and, secondly builds the expected output with respect to the virtual structure. Scientific data servers, and in particular the ones publicly available on the Web, usually provide information retrieval techniques to access data. Our wrappers are composed of a retrieval component, based on an intermediate object view mechanism called 'search views' mapping the source capabilities to attributes, and a XML engine to perform these two tasks. If the retrieval component is specific to each data source, this approach shows that the extraction component (the XML engine) can be common. We describe our system and focus on the retrieval component of the Object-Web Wrapper (OWW) for Web sources. The originality of our approach consists of (1) a common wrapper architecture for flat files and Web data sources sharing a XML engine for data extraction, (2) a generic view mechanism to access data sources with limited capabilities, and (3) the representation of hyperlinks as abstract attributes in the object view as well as their use in the search view. Our approach has been developed and demonstrated as part of a multidatabase system supporting queries via uniform Object Protocol Model (OPM) interfaces.