Xpath Analysis Phase 2

COM 3360 Fall 99 Term Project

Part 1

Evaluating Xpath for Structure-Shyness

Submitted by : Vibhas & Prakash.

Faculty : Prof Karl Lieberherr.

Index

Looking at Structure-Shyness of Xpath *

Summary *

Xpath Basics *

Defining Structure-Shyness *

Analysis Approach *

UML for the test application *

Xpath In Action *

Tool(s) Used. *

Test Files *

Test Cases. *

Comparing Xpath with Dem Java Traversal Primitives *

Traversal Matrix *

Online Resources *

Conclusion *

Future Work *

Looking at Structure-Shyness of Xpath

Summary

Xpath is a standard from W3C (a "recommendation" in W3C parlance as opposed to "working drafts") that can be used to navigate/address parts of an XML document and is meant to be used with XSLT & XPointer standards from W3C which are the languages for document-transformation and Multi-directional links respectively. This project attempts to evaluate the structure-shyness of Xpath.

Xpath Basics

Xpath is used in conjunction with XSLT & XPointer standards. XSLT (XML Style Language Transformations) is used to convert XML document from one form to another (typically HTML). It takes two input files - an XML document along with it’s DTD or equivalent and an XSLT language instructions file and outputs another xml document. XPointer specifies complex Links like bi-directional links and allows the location of a specific part of XML document. Any fragment identifier that points into an XML resource must be an Xpointer.

Both these languages need to navigate an XML document and hence the navigation part was extracted out as a separate language – Xpath.

XPath gets its name from the use of a path notation as in URLs for navigating through the hierarchical structure of an XML document. It uses a compact, non-XML syntax to facilitate it’s use within URIs and XML attribute values and operates on the abstract, logical structure of an XML document, rather than its surface syntax. The primary syntactic structure of XML is an expression that evaluates to yield an object which can be one of the four types below:

node-set (an unordered collection of nodes without duplicates)
boolean (true or false)
number (a floating-point number)
string (a sequence of UCS characters)

Expression evaluation occurs with respect to a context. XSLT and XPointer specify how the context is determined for XPath expressions used in XSLT and XPointer respectively. The context consists of:

a node (the context node)
a pair of non-zero positive integers (the context position and the context size)
a set of variable bindings
a function library
the set of namespace declarations in scope for the expression

The patterns can be matched using notations shown below

Symbol

Meaning

Example

/

Gets all matching children

Chapter / Section / *: will give all children nodes of the node "Section" which also are grandchildren of the node "Chapter".

//

Gets all matching descendants`

Chapter // Paragraph : Will return all paragraphs that are descendants of the Chapter node no matter how many nodes come in between.

@

Gets all matching attributes

@UNITS : will return value of the UNITS attribute for the current node.

Comment()

Gets the comment node

Chapter / * / comment() : Will return the comment node (tag) of all children nodes belonging to the chapter node.

[ ]

Gets nodes satisfying a condition.

Chapter [Annotation] : Will return Chapter nodes that contain Annotation attribute (assuming the attribute is optional).

Defining Structure-Shyness

Rather than a formal definition, we feel it would suffice in the context of this project to define structure-shyness as follows : a piece of code involving any kind of traversal is structure-shy if details of the route followed along traversal are not "hard-wired" into it. This in essence is also the implication of the "Law Of Demeter" (LOD) and forms the basis of flexible software development.

Analysis Approach

We adopted a two step approach. In the first we executed various Xpath traversals on a sample XML application data file. In the second step we attempted to find equivalent Xpath commands for all the Dememter Java traversal primitives as determined by the AP book. Since Xpath is an emerging standard (started in April 1999 and version 1.0 was published only this month), not many tools were found to support this yet. Of those that were found (for example Caucho ‘s Resin 1.0 ) supplied Xpath Java APIs that had to be coded in a program to create our testing tool. We decided to use a tool that required minimum efforts to set up the testing tool. The best appeared to be a tool called XT that was written by James Clark (author of W3C ‘s Xpath).

UML for the test application

Xpath In Action

Tool(s) Used.

XT is a Java utility that needs two input files - a .xml file containing XML document and .xsl file containing XSLT intructions with embedded Xpath instructions. The output would be another XML file. No compiling or elaborate coding is necessary to test the various navigational combinations. Please note that the abbreviated version of Xpath syntax is used throughout this report.

Here is an example of .xml file

<?xml version="1.0"?>
<?xml-stylesheet type="text/xml" href="14-2.xsl"?>
<PERIODIC_TABLE>

  <ATOM STATE="GAS">
    <NAME>Hydrogen</NAME>
    <SYMBOL>H</SYMBOL>    
    <ATOMIC_NUMBER>1</ATOMIC_NUMBER>
    <ATOMIC_WEIGHT>1.00794</ATOMIC_WEIGHT>
    <BOILING_POINT UNITS="Kelvin">20.28</BOILING_POINT>
    <MELTING_POINT UNITS="Kelvin">13.81</MELTING_POINT>
    <DENSITY UNITS="grams/cubic centimeter"><!-- At 300K -->
      0.0899
    </DENSITY>
  </ATOM>

  <ATOM STATE="GAS">
    <NAME>Helium</NAME>
    <SYMBOL>He</SYMBOL>    
    <ATOMIC_NUMBER>2</ATOMIC_NUMBER>
    <ATOMIC_WEIGHT>4.0026</ATOMIC_WEIGHT>
    <BOILING_POINT UNITS="Kelvin">4.216</BOILING_POINT>
    <MELTING_POINT UNITS="Kelvin">0.95</MELTING_POINT>
    <DENSITY UNITS="grams/cubic centimeter"><!-- At 300K -->
      0.1785
    </DENSITY>
  </ATOM>
  
</PERIODIC_TABLE>

And here is an example of .xsl file with embedded Xpath instructions. The SELECT and MATCH attributes in the file below can contain any valid Xpath expression.

<?xml version="1.0"?>
<xsl:stylesheet version="1.0" 
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

    <xsl:template match="/PERIODIC_TABLE/*">
      <html>
        <body>
          <h1>Atomic Number vs. Melting Point</h1>
          <table>
            <th>Element</th>
            <th>Atomic Number</th>
            <th>Melting Point</th>
            <xsl:apply-templates/>
          </table>
        </body>
      </html>
    </xsl:template>
    
   <xsl:template match="ATOM">
    <tr>
      <td>
        <xsl:value-of select="NAME"/>
      </td>
      <td>
        <xsl:value-of select="child::ATOMIC_NUMBER"/> <- Alt Syntax
      </td>
      <td>

<xsl:value-of select=" ATOMIC_NUMBER|ATOMIC_WEIGHT|SYMBOL"/>

      </td>
    </tr>
  </xsl:template>

    
    <xsl:template match="MELTING_POINT">
      <xsl:value-of select="."/>
      <xsl:value-of select="@UNITS"/>
    </xsl:template>
    
</xsl:stylesheet>

Test Files

We used an Invoice-Data application schema as our test example. This consisted of the following files located on CCS machines at - ~vibhasp/com3360/project:

XT :

invoice.xml : XML data file with schema (DTD).

invoice.xsl* : XSL with embedded Xpath traversals

(* = 1,2 etc for test case number).

Invoice.out* : output files for test cases.

(* = 1,2 etc for test case number).

Demeter Java :

invoice.cd : Equivalent Class Dictionary for the XML document.

invoice.beh : Behaviour file specifying traversal test-cases.

invoice.input : Input Invoice Data.

project.prj : Demeter Java project file.

Test Cases.

A combination of test cases that include elements, attributes, "/", "//" and the wild card character * will need to be tested for eveluating structure-shyness. See the section Test Matrix for the actual tests done.

Test Case No.	Description	Xpath Commands	Comments
1.	Fetching an element	Invoice	Returns all Invoice Nodes including children.
2.	Fetching an attribute	@Maker	Returns all Maker attributes.
3.	Fetching a grandchild element for all parent	Invoice/*/Product	Returns all grandchild Product nodes of Invoice.
4.	Fetching all descendant children	Invoice_collection //Address	Returns all descendant Address nodes of root.
5.	Fetch an element using it's position	//Invoice[2]	Returns the second Invoice nide.
6.	Fetch all Product Names for Invoices having 2 entries.	Invoice [Entries/@n = "2"] // @Prod_name	Returns all Prod_name attributes for invoices having 2 entries.

Comparing Xpath with Dem Java Traversal Primitives

Traversal Matrix

Table 1: The following test cases were equivalent for both Demeter Java & Xpath.

Test Case No.	Demeter Java Traversals	Xpath Equivalent	Results/Comments
1.	{Invoice_collection to Product}.	Invoice_collection//Product	Equivalent
2.	*Join Operation* {Invoice_collection to Address via sales}.	Invoice_collectio//Person[Type = "sales"] // Address	. Equivalent

Table 2: The following test cases were where Xpath had no equivalents for Demeter Java .

Test Case No.	Demeter Java Traversals	Xpath Equivalent	Results/Comments
1.	{Invoice_collection to Address bypassing sales}.	Invoice_collectio//Person[not(Type = "sales")] // Address	I believe the Xpath expression should have worked but coud not get it to run.
2.	*Merge Operation* {Invoice_collection to {Name,Address}.		Still searching for equivalent syntax.

Table 3: The following test cases were where Demeter Java had no equivalents for Xpath.

Test Case No.	Demeter Java Traversals	Xpath Equivalent	Description	Results/Comments
1.		@Maker	Get Attribues	The main problem here is there is no direct way of distinguishing child node from attributes in Dem java.
2.
3.
4.

Online Resources

http://www.w3.org/TR/xpath : The W3C standards for Xpath

http://www.schema.net/ : Source for XML DTDs.

http://www.jclark.com/xml/xt.html : Tool by James Clark (authour of W3C Xpath standard).

http://www.ccs.neu.edu/research/demeter/DemeterJava : Demeter Java home page.

Conclusion

Xpath is a fairly structure-shy language for expressing traversals. However on comparing this with Dem Java constructs, we found few traversal primitives that are not easily translated into Xpath (as noted in Table 2).

Future Work

Since Xpath is an emerging standard, this effort will need to be revisited as more features are added to the language.

Symbol	Meaning	Example
/	Gets all matching children	Chapter / Section / *: will give all children nodes of the node "Section" which also are grandchildren of the node "Chapter".
//	Gets all matching descendants`	Chapter // Paragraph : Will return all paragraphs that are descendants of the Chapter node no matter how many nodes come in between.
@	Gets all matching attributes	@UNITS : will return value of the UNITS attribute for the current node.
Comment()	Gets the comment node	Chapter / * / comment() : Will return the comment node (tag) of all children nodes belonging to the chapter node.
[ ]	Gets nodes satisfying a condition.	Chapter [Annotation] : Will return Chapter nodes that contain Annotation attribute (assuming the attribute is optional).