Attribute grammar-based interactive system to retrieve information from XML documents

A system to interactively access extensible markup language documents aiming at information retrieval (IR) is described. The system has two main modules: the query editor/processor, where the user specifies his/her needs and the document analyser, which performs operations for query evaluation. The interactive construction of queries is based on the manipulation of intermediate results during query edition and evaluation. Queries are written in IXDIRQL, a query language that extends XPath with selection operations to extract the interesting subset of elements from intermediate results. This helps the user in building queries to retrieve the desired results. Moreover, textual similarity search of traditional IR is also possible in IXDIRQL, leading to a ranked list of elements. To support a syntax-directed edition of queries and its incremental evaluation during the iterative process, IXDIRQL is specified by an attribute grammar (AG). This formalisation enables the use of an automatic generator of the desired working environment. In this system, documents are also represented by AG. This representation uniformly defines structure, content and operations over documents; this allows for a better interoperability between components. The system has been used by real users to check its correct behaviour and the correct specification of queries, using selection operations.