xslt php select the first sibling and all its child nodes

Question

How does one select the first sibling of a xml node with all its child nodes and apply some transformations on it? So far I only succeeded in selected either the first sibling and only the first sibling (no child nodes that is) or everything following the xml node.

Say we have xhtml like this:

<div class="chapter">Chapter <span class="number">1.1</span> Lorum ipsum</div>
<h2 class="article">Article <span class="number">1.</span> Lorum ipsum</h2>
<p>Lorum ipsum</p>

And the result we are after is xml like this:

<chapter>
  <heading>
    <label>Chapter</chapter>
    <number>1.1</number>
    <title>Lorum ipsum</title>
  </heading>
  <article>
    <heading>
      <label>Article</chapter>
      <number>1.</number>
      <title>Lorum ipsum</title>
    </heading>
    <par>Lorum ipsum</par>
  </article>
</chapter>

My guess is that I need to do some regex magic to get the label and title tags right, but if this could also been done using plain xslt that would be great.

Sensible/correct nesting in the input would help a lot now. Is there a deeper reason why the <div class="chapter"> does not enclose everything that belongs to it? — Tomalak
I know, the main reason is it is easier on the client part and that is a Good Thing (tm) ;) However just how much easier it is this way is open for some internal debate over here, must admit I'm more for better nesting. — Jeroen Beerstra
How is not wrapping a bunch of elements that are logical unit in their own div easier on the client? What could this possibly break? — Tomalak
Client as in the people actually providing the original input. Since they are well paid lawyers, the less they have to do in their wysiwig editor, the better. Although as said I'm rather sceptic as to just how much easier it is this way... — Jeroen Beerstra

Tomalak Tomalak · Accepted Answer · 2010-02-11T15:36:32

This XSLT 1.0 transformation:

<xsl:stylesheet 
  version="1.0"
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
  xmlns:xhtml="http://www.w3.org/1999/xhtml"
  exclude-result-prefixes="xhtml"
>
  <xsl:output encoding="utf-8" />

  <!-- the identity template (copies all nodes verbatim, unless
       more specific templates implement different behavior) -->
  <xsl:template match="node()|@*">
    <xsl:copy>
      <xsl:apply-templates select="node()|@*" />
    </xsl:copy>
  </xsl:template>

  <!-- start by applying templates to all chapter divs -->
  <xsl:template match="xhtml:html">
    <text>
      <xsl:apply-templates select="xhtml:body/xhtml:div[@class='chapter']" />
    </text>
  </xsl:template>

  <!-- chapter div: generate heading, apply templates to adjacent h2 -->
  <xsl:template match="xhtml:div[@class='chapter']">
    <chapter>
      <xsl:apply-templates select="." mode="heading" />
      <!-- ... where the first preceding chapter div has the same ID -->
      <xsl:apply-templates select="
        following-sibling::xhtml:h2[
          generate-id(preceding-sibling::xhtml:div[@class='chapter'][1])
          =
          generate-id(current())
        ]
      "/>
    </chapter>
  </xsl:template>

  <!-- h2: generate heading, apply templates to adjacent paras -->
  <xsl:template match="xhtml:h2[@class='article']">
    <article>
      <xsl:apply-templates select="." mode="heading" />
      <xsl:apply-templates select="
        following-sibling::xhtml:p[
          generate-id(preceding-sibling::xhtml:h2[@class='article'][1])
          =
          generate-id(current())
        ]
      "/>
    </article>
  </xsl:template>

  <!-- headings follow the same scheme, so we can use a unified template -->
  <xsl:template match="xhtml:div | xhtml:h2" mode="heading">
    <heading>
      <label>
        <xsl:value-of select="normalize-space(text()[1])" />
      </label>
      <number>
        <xsl:value-of select="normalize-space(xhtml:span[@class='number'])" />
      </number>
      <title>
        <xsl:value-of select="normalize-space(text()[2])" />
      </title>
    </heading>
  </xsl:template>

  <xsl:template match="xhtml:p">
    <par>
      <xsl:apply-templates select="node()" />
    </par>
  </xsl:template>

</xsl:stylesheet>

when applied to

<html xmlns="http://www.w3.org/1999/xhtml">
  <body>
    <div class="chapter">Chapter <span class="number">1.1</span> Lorum ipsum</div>
    <h2 class="article">Article <span class="number">1.</span> Lorum ipsum</h2>
    <p>Lorum ipsum A</p>
    <p>Lorum ipsum B</p>
    <h2 class="article">Article <span class="number">2.</span> Lorum ipsum</h2>
    <p>Lorum ipsum D</p>
    <h2 class="article">Article <span class="number">3.</span> Lorum ipsum</h2>
    <p>Lorum ipsum E</p>
    <p>Lorum ipsum F</p>
    <div class="chapter">Chapter <span class="number">2.1</span> Lorum ipsum</div>
    <h2 class="article">Article <span class="number">1.</span> Lorum ipsum</h2>
    <p>Lorum ipsum G</p>
  </body>
</html>

yields:

<text>
  <chapter>
    <heading>
      <label>Chapter</label>
      <number>1.1</number>
      <title>Lorum ipsum</title>
    </heading>
    <article>
      <heading>
        <label>Article</label>
        <number>1.</number>
        <title>Lorum ipsum</title>
      </heading>
      <par>Lorum ipsum A</par>
      <par>Lorum ipsum B</par>
    </article>
    <article>
      <heading>
        <label>Article</label>
        <number>2.</number>
        <title>Lorum ipsum</title>
      </heading>
      <par>Lorum ipsum D</par>
    </article>
    <article>
      <heading>
        <label>Article</label>
        <number>3.</number>
        <title>Lorum ipsum</title>
      </heading>
      <par>Lorum ipsum E</par>
      <par>Lorum ipsum F</par>
    </article>
  </chapter>
  <chapter>
    <heading>
      <label>Chapter</label>
      <number>2.1</number>
      <title>Lorum ipsum</title>
    </heading>
    <article>
      <heading>
        <label>Article</label>
        <number>1.</number>
        <title>Lorum ipsum</title>
      </heading>
      <par>Lorum ipsum G</par>
    </article>
  </chapter>
</text>

xslt php select the first sibling and all its child nodes

2 Answers