Object-Oriented Mediator Queries to Internet Search Engines

摘要

A system is described where multiple Internet search engines (ISEs), e.g. Alta Vista or Google, are accessed from an Object-Relational mediator database system. The system makes it possible to express object-oriented (OO) queries to different ISEs in terms of a high level OO schema, the ISE schema. The OO ISE schema combined with the mediator database system provides a natural and extensible mechanism in which to express queries and OO views that combine data from several ISEs with data from other data sources (e.g. relational databases). High-level OO web queries are translated through query rewrite rules to specific search expressions sent to one or several wrapped ISEs. A generic ISE query function sends the translated queries to a wrapped ISE. The result of an ISE query is delivered as a stream of semantically enriched objects in terms of the ISE schema. The system leverages publicly available wrapper toolkits that facilitate extraction of structured data from web sources, and it is independent of the actual wrapper toolkit used. One such wrapper toolkit was used for generating HTML wrappers for a few well-known ISEs.