Visual
Specification and Automatic Transformation of Web Interchanging Documents
Award number: NSF ITR 0218738
PI: Kang Zhang
Project Summary
This NSF IIS
supported project investigates a visual approach to the specification and
transformation of XML-based Web documents. Based on a graph grammar and graph
transformation formalism, the visual approach will allow the structural
requirements of any XML dialect to be specified graphically and automatically
generate a graphical toolkit for the construction of Web documents of
represented in the XML dialect. It will also support automatic translation of
Web documents from one XML dialect to another. With the proposed spatial
extension, the grammar formalism will be able to specify adaptive layout for
any Web presentation.
Background
Transformations of
XML documents XSL must be created manually on a case-by-case basis.
Furthermore, writing an effective XSL code requires some degree of programming
skills and good understanding of XML’s working principle. The scientific
communities have realized the importance of the XML technology and the need for
automatic transformations [1]. Of the general textual approaches to the
automatic translation of XML documents, Xtra [5] aims at automatic
transformations between XML documents by discovering a sequence of
transformation operations from the source and target DTD trees. The operations
are used to generate an XSL script, which can then be applied to source XML
documents to transform them to XML documents conforming to the target DTD.
Approaches based on database schemas [3, 4] use rules to match similar document
components for most common cases, and allow the user to customize the rules for
more complex cases. These approaches offer neither automatic structure
validation, nor means for visual representation and specification.
Visual approaches such as Xing [2] have been proposed. Xing achieves
XML transformation and restructuring using some rules that combine the patterns
of queries and results returned by queries. It uses nested boxes to represent
XML data such what each element tag is written on top of a box while all the
attributes of the element are enclosed in the box. Such representation is
essentially textual, supplemented by hierarchical boxes surrounding the text.
For a large document, the levels of nesting could be overwhelming so that it is
hard for the user to understand the overall structure.
Goals, Objectives and Targeted Activities
This project has the
following objectives:
- To identify and find
practical solutions the fundamental issues in visual transformation and
validation of Web documents.
- To develop a visual XML
document design, validation and translation framework to support the
automatic generation of Web document construction and transformation
tools. The generation framework will be equipped with graphical tools for
specifying the document syntax and transformation rules.
- To extend the existing graph
transformation formalism with spatial specification capability so that not
only the logical structure but also the layout of XML documents could be
specified graphically.
- To evaluate the usability of
the visual features of the developed toolset and enhance the
user-friendliness and user productivity is using the prototypical toolset.
Project Impact
The proposed project
will enhance the accessibility of IT in general and Web documents in particular
to the general public by developing the user-friendly visual tools for Web
document design and transformation. Web designers will find it easy to define
Web document structures through graph grammar specifications and use the toolset
to automatically generate graphical languages, each within a graph editor for a
domain-specific XML, and if desired, for its translation into the XML of
another domain. Application users in the domain without any knowledge of XML
could then use the graph editor to construct domain-specific XML documents in
an intuitive manner.
Teaching and
training of Web engineering and programming language generation will benefit
from our prototyping toolset. Using the toolset, we could step-by-step and
intuitively illustrate how a visual XML language is generated in the same
fashion as for conventional language generation, according to a given set of
specifications. Then how an XML document can be rapidly constructed, parsed,
validated, and transformed into another document style using the generated
visual XML language can also be vividly demonstrated. The toolset and the
concepts associated with it will make excellent teaching and self-training
materials for the courses on programming languages, compilers, and visual interfaces.
Applications of the Prototype
a) XML-Based Biological Information
Translation
A major challenge in bioinformatics is the interoperation of
heterogeneous biological information from multiple sources of sequence data
from genomic and sequence databases. This section goes through an example to
demonstrate Trans VME for graphical transformation of bioinformatics XML
documents through the translation from a NCBI (National Centre for
Biotechnology Information) XML document into a BSML (Bioinformatic Sequence
Markup Language) document.
NCBI was established in 1988 as a national resource for molecular
biology information. NCBI creates public databases, conducts research in
computational biology, develops software tools for analyzing genome data, and
disseminates biomedical information.
Internally NCBI stores data in a variety of ways most appropriate to the
flow of the data and its semantics. These may include normalized relational
databases (eg. for ESTs), ASN.1 (eg. for other types of sequences), or XML (eg.
for journal articles). BSML is an open XML data standard created for more
efficient communication within the life sciences community. BSML was one of the
first eXtensible Markup Languages (XML) to be developed after the World Wide
Web Consortium (W3C) proposed XML.
Generated from a set of transformation rules, the Trans VME named Biotrans,
facilitates the interchange of data from NCBI XML format to BSML format, i.e.
an implementation of Trans operator on data instances. As shown in Figure 1.a,
the transformation rules for the VME Biotrans consist of 7 rules, of which the
rule 1 is shown in the figure. Figure 1.b shows the generated VME, Biotrans,
with a portion of input NCBI XML document displayed as a tree. Biotrans VME
then translates the input document into BSML file shown in Figure 1.c.
Facilitated by the
framework, constructing a visual transformation environment is only to create a
set of rules. The effort for low-level programming is therefore minimized, and
the framework boosts the interoperation among heterogeneous bioinformatics data
instances.
b) Web Documents Adaptation
Various viewing conditions
exist while users surf the Internet, such as varying screen sizes, style preferences,
and different device capabilities. Majority Web documents are for desktop
displaying, such as XML and XHTML, while WML (Wireless Markup Language)
documents are for displaying on mobile devices. As shown in Figure 1, a Trans operator VME, named
XMLtoWML, was developed based on the prototype system.
Figure 1.a defines
7 transformation rules in the rule generator, which in turn generates a Trans
VME for the translation from input XML documents to WML documents. Figure 1.b
shows a source XML document, which has a page with two sections. In the
generated VME, XMLtoWML, the input XML document is translated into a WML
document as shown in Figure 1.c.
With the growing
demanding for mobile Internet surfing, adaptation of Web documents for small
screen display is becoming critical. With the help of the framework,
programming via transformation rules to reuse existing XML or XHTML documents
saves a lot of effort to create new mobile pages or programming adaptation
tools from scratch.
c) Ontology Mapping and Integration
We provide a warehouse design for
heterogeneous ontologies, such as OBOs and other proprietary ontologies. These
proprietary ontologies can be imported into the warehouse through proprietary
designated interface programs, while tools are available for importing OBOs
into the warehouse. To interoperate these ontologies, mapping tables were
designed to store the correspondences between terms of these ontologies.
The IOMG is implemented to
generate mappings among the ontologies retrieved from the warehouse and detects
possible false mappings for domain experts to confirm. Figure 2 shows the IOMG
mapping generation interface.
Figure 2 Oasis Ontology
Mapping
As highlighted, Parkinson disease is successfully mapped to Paralysis agitans, which is synonym to Parkinson disease. The automatic
generation may produce some false mappings too, or miss some true mappings.
Result mappings should be verified by biomedical experts before being saved
into the ontology warehouse. Through a friendly user interface, domain experts
can correct mappings by editing the mappings directly in the graph.
c) Schema Translation
Similar to the Trans operator on data instances,
a Trans
operator VME on data models is also developed by the graph transformation rules.
Early version of
XML, version 1.0, contains a built-in schema language, i.e. DTD, which is going
to be replaced by XML schema because of some limitations of the DTD. Since its
first introduction many DTD documents are available and have to migrate to the
XML schema documents. Based on the framework, we designed a Trans VME to help the migration
from DTD to XML Schema.
As shown in Figure
3.a, the graph transformation rules specify a Trans operator to translate a DTD
document to XML Schema document. For the input DTD document, five rules are
defined and the first one shown in the popup window of Figure 3.a. Since the
schema and data instance specified in a uniform representation, the
transformation of the schema is specified in the similar way.
Figure 3.b shows
the generated DTD to XML Schema transformation environment, called DTD2XS VME,
which displays the input DTD document as a tree. The VME translates the input DTD into an XML
Schema document displayed as a tree in Figure 3.c.
PhD Students Supported
Guanglei Song (currently Senior
Software Developer, eBay Inc., CA)
Jun Kong (currently Assistant
Professor,
Publications Related to This Project
JOURNAL PAPERS:
J. Kong, K. Zhang, and X. Zeng, Spatial Graph Grammars for
Graphical User Interfaces, ACM Transactions on Computer-Human
Interaction, Vol.13, No.2, June 2006, 268-307.
G.L. Song, J. Kong, and K. Zhang, AutoGen: Easing Model Management
through Two Levels of Abstraction, Journal of Visual Languages and
Computing, Vol.17, No.6, 2006, Elsevier Science Inc.,
K. Zhang, J. Kong, M.K. Qiu, and G.L. Song, Multimedia Layout Adaptation
Through Grammatical Specifications, ACM/Springer Multimedia Systems,
Vol.10, No.3, 2005, 245-260.
REFEREED CONFERENCE PAPERS:
G-L. Song, Y. Qian, Y. Liu, and K. Zhang, Oasis: a Mapping and Integration
Framework for Biomedical Ontologies, Proc. 19th IEEE International Symposium on Computer-Based Medical
Systems, Salt Lake City, USA, 22-23 June 2006, 611-616.
K.L. Ates, K. Zhang, and B. Prabhakaran, Visual Querying on Human Motion for the
Disabled, Proc. 2006 IEEE Symposium
on Visual Languages and Human-Centric Computing (VL/HCC’06), Brighton, UK,
4-8 September 2006, IEEE CS Press, 222-223.
K. Ates, J. Kukluk, L. Holder, D. Cook, and
K. Zhang, Graph Grammar
Induction on Structural Data for Visual Programming, Proc. 18th IEEE
International Conference on Tools with Artificial Intelligence (ICTAI'06),
Washington D.C., USA, 13-15 November 2006, 232-239.
K. Zhang, G.L. Song, and J. Kong, Interoperating XML-Style of Digital
Artifacts for Information Reuse, Proc.
2005 IEEE
International Conference on Information Reuse and Integration (IRI’05),
G.L. Song, K. Zhang, B. Thuraisingham, and J.
Cao, Towards Access Control
for Visual Web Model Management, Proc. 2005 IEEE International
Conference on e-Technology, e-Commerce and e-Service (EEE'05), Hong Kong,
China, 29 March - 1 April 2005, IEEE CS Press, 722-727.
G.L. Song, K. Zhang, B. Thuraisingham, and J.
Kong, Secure Model
Management Operations for the Web, In: S. Jajodia and D. Wijesekera
(Eds.) Data and Application Security XIX - Proc. 19th Annual IFIP WG 11.3
Working Conference on Data and Applications Security, Storrs, USA, 7-10
August 2005, LNCS 3654, Springer-Verlag, 237-251.
G.L. Song, K. Zhang, R.K. Wong, and J. Kong, Management
of Web Data Models Based on Graph Transformation, Proc. 2004 IEEE/WIC/ACM International
Conference on Web Intelligence, Beijing, China, 20-24 September 2004, IEEE CS Press, 398-404.
G.L. Song, K. Zhang, and J. Kong, Model Management Through Graph
Transformations, Proc. 2004 IEEE Symposium on Visual Languages and
Human-Centric Computing (VL/HCC'04), Rome, Italy, 26-29 September 2004,
IEEE CS Press, 75-82.
J. Kong and K. Zhang, Parsing Spatial Graph Grammars, Proc. 2004
IEEE Symposium on Visual Languages and Human-Centric Computing, Rome,
Italy, 26-29 September 2004, IEEE CS
Press, 99-101.
J. Kong and K.
Zhang, On a Spatial Graph Grammar Formalism, Proc. 2004 IEEE Symposium on Visual Languages and Human-Centric Computing, Rome, Italy, 26-29 September 2004, IEEE CS Press, 102-104.
G.L. Song, K. Zhang, R.K. Wong, and J. Kong, Management of Web Data Models
Based on Graph Transformation, Proc. 2004 IEEE/WIC/ACM International
Conference on Web Intelligence (WI'04), Beijing, China, 20-24 September
2004, IEEE CS Press, 398-404.
G.L. Song and K. Zhang, Visual XML Schemas
Based on Reserved Graph Grammars, Proc. International Conference on
Information Technology (ITCC’04), Las Vegas, USA, 5-7 April 2004, IEEE CS
Press, 687- 691.
J. Kong, M.K. Qiu, and K. Zhang, Authoring Multimedia Documents
Through Grammatical Specifications, Proc. 2003 IEEE International
Conference on Multimedia & Expo (ICME'2003), Baltimore, USA, 6-9 July,
2003, IEEE CS Press, 629-632.
M.K. Qiu, G.L Song, J. Kong, and K. Zhang, Spatial Graph Grammars for Web
Information Transformation, Proc. 2003 IEEE Symposium on
Visual/Multimedia Languages (VL'03), Auckland, New Zealand, 28-31 October
2003, IEEE CS Press, 84-91.
J. Kong, and K. Zhang, Graph-based Consistency
Checking in Spatial Information Systems, Proc. 2003 IEEE Symposium
on Visual Languages and Formal Methods (VLFM'03), Auckland, New Zealand,
28-31 October 2003, IEEE CS Press, 153-160.
J. Kong and K. Zhang, Toward A Graphical
Approach to Multimedia Document Design, Proc. 23rd International
Conference on Distributed Computing Systems Workshops - 5th International
Workshop on Multimedia Network Systems and Applications, Providence, USA, 19-22
May 2003, IEEE CS Press, 666-671.
Cited References
- K-H. Cheung, Y. Liu, A.
Kumar, M. Snyder, M. Gerstein, and P. Miller, An XML Application for
Genomic Data Interoperation, Proc. 2rd IEEE Symp. on Bioinformatics and
Bioengineering, 4-6 Nov. 2001, 97-103.
- M. Erwig, A Visual Language
for XML, Proc. 2000 IEEE Symp. on Visual Languages,
- L. Milo and S. Zohar, Using
Schema Matching to Simplify Heterogeneous Data Translation, Proc. 24th
Int. Conf. on Very Large Databases, New York City, USA, 24-27 August,
1998, 122-133.
- W.M. Shui and R.K. Wong,
Application of XML Schema and Active Rules System in Management and
Integration of Heterogeneous Biological Data, Proc. 3rd IEEE Symp. on
Bioinformatics and Bioengineering, 10-12 March 2003, 367-374.
- H. Su, H. Kuno, and E.A.
Rundensteiner, Automating the Transformation of XML Documents, Proc. 3rd
Int. Workshop on Web Information and Data Management,