Interrogare collezioni di documenti XML: una interfaccia utente

Oreste Signore - Marco Andreini - Cristian Lucchesi - Silvia Martelli

Ufficio Italiano W3C presso il C.N.R.
Area della Ricerca di Pisa San Cataldo - Via G. Moruzzi, 1 - 56124 Pisa
Email:oreste@w3.org

XML per i Beni Culturali
Esperienze e prospettive per il trattamento di dati strutturati e semistrutturati
Scuola Normale Superiore
Pisa, 25 marzo 2004

Talk layout

• Main goals
• Design issues
• Demo
• Conclusion

This work has been financed by the project QUESTION-HOW (Quality Engineering Solutions via Tools, Information and Outreach for the New Highly-enriched Offerings from W3C: Evolving the Web in Europe), contract IST-2000-28767

User needs in accessing XML data collections

XML has no semantics per se ...

... hence user needs:

• understanding structure
• understanding semantics
• sharing his/her knowledge base with the indexer

Main goals

• Supporting the user in formulating a "semantically correct" query
  • understanding element semantics
  • entering correct data (list of values, dictionaries, constraints)
  • browsing concepts (access to thesauri, ontologies, etc.)
• Supporting user interaction:
  • multilinguality
  • different cultures (format conversions, e.g. for calendars)
  • different expertise levels
  • formulate complex queries without learning XPath
• Adjusting the output (insert hyperlinks, emphasize searched terms, processing the text)
• Do not affect existing data collection and its schema
• Keeping the architecture open and distributed

Architecture: a rough sketch

Architecture: fully web-based

The XML Schema is externally annotated in RDF Metadata are also stored in the system objects. RDF description can be imported in the system. The system can export RDF annotation A component allows to add metadata to the system objects The user can browse the structure and formulate the query For each element/attribute the system knows and can show to the user: semantics constraints The query is prepared in a general format, can be mapped onto different search engines

Architecture: main features

• The XML Schema is externally annotated in RDF
  • Metadata are also stored in the system objects.
  • RDF description can be imported in the system.
  • The system can export RDF annotation
  • A component allows to add metadata to the system objects
• The user can browse the structure and formulate the query
• For each element/attribute the system knows and can show to the user:
  • semantics
  • constraints
• The query is prepared in a general format, can be mapped onto different search engines

The RDF annotation

• Identification of elements
• Properties of elements
  • description
  • searchability
  • role (structure vs content)
• Constraints
  • thesaurus, dictionary, externalList, localList
  • web service to invoke
  • dependencies among values of elements
• Pre-processing options
• Post-processing options

The query

Frustrating traditional approaches:

• select field from a pop up window
• insert parentheses and Boolean operators (selecting from different options)
• difficult to formulate complex queries
• difficult to reuse elementary conditions

Composition:

• A composition of several queryFragments (conceptually a single navigation upon the document tree)
• A queryFragment is made by several queryTokens
• A queryToken is the elementary component (a single interaction upon the queryForm on a single element)

Normalization:

• identify the common ancestor for clauses on elements belonging to independent paths

The sample document collection ...

• As from W3C Working Draft of XQuery and XPath Full-Text Requirements (February 2003)
• XML Schema
• RDF annotation
• The collection

... and a (partial) graphical representation of it

Specifying constraints

• local list
• external list
• thesaurus
• related elements

The XML document collection

• XML Schema is required
• Document collection is:
  • compressed
  • indexed
  • queried
  • retrieved

The XCDE search engine

• XCDE library developed at the Department of Computer Science, University of Pisa (prof. Paolo Ferragina).
• Best suited for collections of a limited number of large size documents
• Query language oriented mainly to the processing and management of textual data
• Query syntax:
  • similar to SQL: SELECT-FROM-RETURN
  • SELECT clause specified by means of an XML piece of well-formed text
  • implements most of the IR functionalities detailed in the W3C Working Draft of XQuery and XPath Full-Text Requirements, as well as other powerful string-based queries (regular expressions and error matches).
  • output of the query (the snippet) can be formatted within the RETURN clause which, again, includes an XML piece of well-formed text.
  • A special attribute named xml_var (pivot) is added to elements within the SELECT clause in order to identify some "interesting points" in each document subtree that matches the query.
    Pivots are used in the RETURN clause to indicate the way in which these points in the matching subtree must be visualized.

Demo

• The collection
• XML Schema
• RDF annotation
• Start demo

Conclusion and future plans

• A consistent framework for intelligent and controlled access to XML data collections has been implemented
• RDF allowed the annotation of the (unchanged) schema
• Administrator can point to any appropriate knowledge source (accessing thesauri via web services)
• The interface is generalized:
  • any document collection
  • any search engine
• Some features are not yet implemented (but they are just "conventional development"
• Future developments can consider:
  • accessing heterogeneous document collections
  • inserting and updating documents

Thank you for your attention

?