Deal with multiple sources

Hind · June 5th, 2006, 03:07 AM

Hi,

I have 2 source files .xmi and .graphml I merge both by writing the second file before the root end tag of the first then I wrote my xslt, I've got a correct output but this was not the right way to deal with multiple sources. How can I convert the following xslt to deal with 2 files rather than a merged file.

Code:

<?xml version='1.0' ?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:UML="org.omg.xmi.namespace.UML" xmlns:a="http://graphml.graphdrawing.org/xmlns/graphml">
    <xsl:output method="xml"/>
    <xsl:template match="/">
        <xsl:copy>
            <xsl:copy-of select="@*"/>
            <xsl:apply-templates/>
        </xsl:copy>
    </xsl:template>    
    <xsl:template match="*">
        <xsl:copy>
            <xsl:copy-of select="@*"/>
            <xsl:apply-templates/>
        </xsl:copy>
    </xsl:template>

    <xsl:template match="UML:Namespace.ownedElement[local-name(parent::*[1]) = 'Model']">
            <xsl:copy>
                <xsl:call-template name="graphml"/>
                <xsl:copy-of select="@*"/>
                <xsl:apply-templates/>
            </xsl:copy>            
    </xsl:template>

    <xsl:template name="graphml">
        <xsl:for-each select="//a:graphml/.//a:cluster">
            <xsl:choose>
                <xsl:when test="count(child::*) &gt; 1">
                <xsl:variable name="cOne" select="child::*[1]/@name"/>
                <xsl:variable name="cTwo" select="child::*[2]/@name"/>
                <xsl:if test="not((count(child::*) = '2') and ((contains($cOne, '.java') and contains($cOne, $cTwo)) or (contains($cTwo, '.java') and contains($cTwo, $cOne))))">
                    <xsl:element name="UML:Component">
                        <xsl:attribute name="xmi.id">
                            <xsl:text>.:000</xsl:text>
                            <xsl:value-of select="@id"/>
                        </xsl:attribute>
                        <xsl:attribute name="isSpecification">false</xsl:attribute>
                        <xsl:attribute name="isRoot">false</xsl:attribute>
                        <xsl:attribute name="isLeaf">false</xsl:attribute>
                        <xsl:attribute name="isAbstract">false</xsl:attribute>
                        <xsl:copy-of select="@name"/>
                        <xsl:element name="UML:ModelElement.clientDependency">
                            <xsl:for-each select="a:node">
                                <xsl:element name="UML:Dependency">
                                    <xsl:attribute name="xmi.idref">
                                        <xsl:value-of select="@xmi.id"/>
                                    </xsl:attribute>
                                    <xsl:apply-templates/>
                                </xsl:element>
                            </xsl:for-each>
                        </xsl:element>
                    </xsl:element>
                </xsl:if>
                </xsl:when>
            </xsl:choose>
        </xsl:for-each>
    </xsl:template>

    <xsl:template match="*[*/UML:Package/@xmi.idref]"/>

    <xsl:template match="UML:Package[not(@xmi.idref)]">
        <xsl:apply-templates select="UML:Namespace.ownedElement/child::node()"/>
    </xsl:template>

    <xsl:template match="a:graphml"/>
</xsl:stylesheet>

Thanks,
Hind

mhkay · June 5th, 2006, 03:34 AM

I can't follow all your code without knowing more about your problem, but my guess would be that in teh template name="graphml" you can simply replace

<xsl:for-each select="//a:graphml/.//a:cluster">

by something like

<xsl:for-each select="document('graphml.xml')//a:graphml/.//a:cluster">

There are lots of oddities in your code. For example the above path expression is equivalent to "//a:graphml//a:cluster"

I personally find this kind of code really verbose:

<xsl:element name="UML:Dependency">
                                    <xsl:attribute name="xmi.idref">
                                        <xsl:value-of select="@xmi.id"/>
                                    </xsl:attribute>
                                    <xsl:apply-templates/>
                                </xsl:element>

Much better to write

  <UML:Dependency xml.idref="{@xmi.id}">
    <xsl:apply-templates/>
  </UML:Dependency>

This match pattern:

match="UML:Namespace.ownedElement[local-name(parent::*[1]) = 'Model']

seems to be a particularly obscure way of writing:

match="*:Model/UML:Namespace.ownedElement"

(Do you really not know the namespace of the parent element?)

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Hind · June 5th, 2006, 07:16 AM

Sorry for bad code, but this is because its my first xsl code.

The problem is that I want to generate component elements as a child of the first "UML:Namespace.ownedElement" element wich is chield of "UML:Model".
"UML:Model" has exactly 2 children "UML:ModelElement.TaggedVlaue" and "UML:Namespace.ownedElement" which may contain some descendants with the same name.

I will generate a component based on .graphML file as follow:

for each cluster that contains more than one child
if cluster contains 2 children and the name of ane of them = the name of the other=".jave"
then do nothing
else create a component.

Following is the graphml part of the source file:

Code:

<graphml xmlns="http://graphml.graphdrawing.org/xmlns/graphml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
        xsi:schemaLocation="http://graphml.graphdrawing.org/xmlns/graphml">
        <graph edgedefault="directed" Edge_Weight_Key="edu.uci.ics.jung.graph.decorators.EdgeWeightLabeller@19e15c"
            Clustering_Weight_Key="edu.uci.ics.jung.graph.decorators.EdgeWeightLabeller@11a75a2"
            StringLabeller.LabelDefaultKey="edu.uci.ics.jung.graph.decorators.StringLabeller@210b5b">
            <clusters>
                <cluster id="1">
                    <node id="17" xmi.id=".:00000000000007D6" name="AstroRecord.java" />
                    <node id="6" xmi.id=".:00000000000007D7" name="AstroRecord" />
                    <node id="27" xmi.id=".:0000000000000875" name="DataInput" />
                </cluster>
                <cluster id="2">
                    <node id="25" xmi.id=".:0000000000000805" name="WritePDB" />
                    <node id="21" xmi.id=".:0000000000000804" name="WritePDB.java" />
                </cluster>
                <cluster id="3">
                    <node id="12" xmi.id=".:00000000000007F8" name="CatalogRecord" />
                </cluster>
                <cluster id="4">
                    <node id="4" xmi.id=".:0000000000000B90" name="Vector" />
                </cluster>
                <cluster id="5">
                    <node id="24" xmi.id=".:00000000000007E0" name="ConvertDST" />
                    <node id="15" xmi.id=".:00000000000007DF" name="ConvertDST.java" />
                </cluster>
                <cluster id="6">
                    <node id="24" xmi.id=".:00000000000007E0" name="ConvertDST" />
                    <node id="15" xmi.id=".:00000000000007DF" name="ConvertDST.java" />
                </cluster>
                <cluster id="7">
                    <node id="1" xmi.id=".:000000000000086D" name="DataOutput" />
                    <node id="23" xmi.id=".:0000000000000829" name="RecordIndex.java" />
                    <node id="9" xmi.id=".:0000000000000820" name="PDBElement.java" />
                    <node id="3" xmi.id=".:0000000000000A68" name="Date" />
                    <node id="10" xmi.id=".:000000000000083B" name="RecordOffset.java" />
                    <node id="12" xmi.id=".:00000000000007F8" name="CatalogRecord" />
                    <node id="18" xmi.id=".:0000000000000821" name="PDBElement" />
                    <node id="13" xmi.id=".:0000000000000814" name="DatabaseHeader" />
                    <node id="5" xmi.id=".:000000000000082A" name="RecordIndex" />
                    <node id="19" xmi.id=".:000000000000083C" name="RecordOffset" />
                    <node id="20" xmi.id=".:000000000000080D" name="DatabaseHeader.java" />
                    <node id="26" xmi.id=".:000000000000085D" name="String" />
                </cluster>
                <cluster id="8">
                    <node id="25" xmi.id=".:0000000000000805" name="WritePDB" />
                    <node id="21" xmi.id=".:0000000000000804" name="WritePDB.java" />
                </cluster>
                <cluster id="9">
                    <node id="22" xmi.id=".:00000000000007E3" name="DSTRecord" />
                </cluster>
                <cluster id="10">
                    <node id="8" xmi.id=".:00000000000007CE" name="AstroAppInfo" />
                    <node id="14" xmi.id=".:00000000000007CD" name="AstroAppInfo.java" />
                </cluster>
                <cluster id="11">
                    <node id="2" xmi.id=".:00000000000007EE" name="DumpPDB.java" />
                    <node id="11" xmi.id=".:00000000000007EF" name="DumpPDB" />
                </cluster>
                <cluster id="12">
                    <node id="22" xmi.id=".:00000000000007E3" name="DSTRecord" />
                </cluster>
                <cluster id="13">
                    <node id="10" xmi.id=".:000000000000083B" name="RecordOffset.java" />
                    <node id="19" xmi.id=".:000000000000083C" name="RecordOffset" />
                </cluster>
                <cluster id="14">
                    <node id="16" xmi.id=".:00000000000007FC" name="ParseCatalog" />
                    <node id="7" xmi.id=".:00000000000007F7" name="ParseCatalog.java" />
                </cluster>
            </clusters>
            <edges>
                <edge source="5" target="27" directed="true" RelType="Coupling" edge_weight="28.0" isRemoved="True" />
                ....
            </edges>
        </graph>
    </graphml>
</XMI>