website transformation

Consider the following XML representation of a
hierarchical (DAG) website, where element names
correspond to hyperlink labels and id's correspond
to nodes.

<db id="1">
   <a id="2">
      <d id="5">
         <g id="10"/>
         <h id="11"/>
      </d>
      <e id="6">
         <g idref="11"/>
         <h id="12"/>
      </e>
   </a>
   <b id="3">
      <f idref="5"/>
      <d id="7">
         <g id="13"/>
      </d>
   </b>
   <c id="4">
      <d id="8">
         <g id="14"/> 
      </d>
      <f id="9">
         <g id="15"/> 
         <h id="16"/> 
      </f>
   </c>
</db>

Notice the presence of crosslinks in this
hierarchy (e.g., from node 3 to node 5)
making it a dag and not a tree.

I'd like to prune all paths from the root
to a leaf (nodes 10-16 above) not indexed
under a given hyperlink label.

For example, if I am given the hyperlink
label "d", I'd like to produce the following
XML representation of the transformed website:

<db id="1">
   <a id="2">
      <d id="5">
         <g id="10"/>
         <h id="11"/>
      </d>
      <e id="6">
         <g idref="11"/>
      </e>
   </a>
   <b id="3">
      <f idref="5"/>
      <d id="7">
         <g id="13"/>
      </d>
   </b>
   <c id="4">
      <d id="8">
         <g id="14"/> 
      </d>
   </c>
</db>

And finally, since "d" has been given, I'd like to
collapse each hyperlink labeled "d", resulting in:

<db id="1">
   <a id="2">
         <g id="10"/>
         <h id="11"/>
      <e id="6">
         <g idref="11"/>
      </e>
   </a>
   <b id="3">
      <f idref="10"/>
         <g id="13"/>
   </b>
   <c id="4">
         <g id="14"/> 
   </c>
</db>

The way I see it, the presence of
crosslinks (e.g., from node 3 to 5 above) in the
dag, require three processing steps to achieve
the cumulative website transformation.

(i) forward propagate from the root to find all
    leaf nodes indexed under "d"
(ii) back-propagate from these leaf nodes to
     retain only the sequences indexing this
     set of leaves
(iii) collapse hyperlinks labeled "d"

Can all of these tasks be accommodated with only
one stylesheet?

I have stylesheets for each individual step (see below), but am
not sure if my requirements can be accommodated in one stroke.

(stylesheet for (i) above)

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

<xsl:output method="text"/>
<xsl:strip-space elements="*"/>

<xsl:template match="node()[ancestor-or-self::d
              and count (child::node()) = 0]">
   <xsl:value-of select="@id"/><xsl:text>
</xsl:text>
</xsl:template>

<xsl:template match="text()"/>

</xsl:stylesheet>

(stylesheet for (ii) above)

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

<xsl:output method="xml"/>
<xsl:strip-space elements="*"/>

<xsl:template match="//*[descendant-or-self::node()[@id='10'
              or @id='11' or @id='13' or @id='14']]">
   <xsl:copy>
   <xsl:apply-templates/>
   </xsl:copy>
</xsl:template>

</xsl:stylesheet>

(stylesheet for (iii) above)

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

<xsl:output method="xml"/>
<xsl:strip-space elements="*"/>

<xsl:template match="d"/>

<xsl:template match="*|@*">
   <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
   </xsl:copy>
</xsl:template>

</xsl:stylesheet>

Of course, the hardcoded values in the above
three stylesheets are for illustrative purposes
only and will be included via parameters in
the final implementation.

Thank You and Best Regards,
Saverio

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list