MarkLogic 8中xdmp.tidy()误删article标签的解决办法及替代方案

The issue here is that the underlying HTML Tidy library used by xdmp.tidy() in MarkLogic 8 doesn’t recognize HTML5 elements like <article> by default—so it strips them as "invalid" tags. Here’s how to fix it:

Preserve

tags with Tidy configuration #

You can pass a custom Tidy configuration to xdmp.tidy() to explicitly mark <article> as a valid block-level element, and enable HTML5 mode. Here’s an example in XQuery:

let $html := "<article><h1>Hello</h1><p>Some content</p></article>"
let $tidy-config := "<config>
  <new-blocklevel-tags>article</new-blocklevel-tags>
  <doctype>html5</doctype>
  <output-xhtml>yes</output-xhtml>
</config>"
return xdmp:tidy($html, $tidy-config)

Key options explained:

• new-blocklevel-tags: Tells Tidy to treat <article> as a valid block-level tag (similar to <div> or <section>), so it won’t strip it.
• doctype: html5: Enables HTML5 parsing rules, which helps Tidy recognize more modern elements natively.
• output-xhtml: yes: Ensures the output is well-formed XHTML, which is what you’re targeting.

If you prefer a string-based config instead of XML, you can write it like this:

let $tidy-config := "new-blocklevel-tags: article
doctype: html5
output-xhtml: yes"
return xdmp:tidy($html, $tidy-config)

Alternative HTML-to-XHTML conversion methods #

If you run into limitations with Tidy, here are two reliable alternatives in MarkLogic 8:

1. Use xdmp:html-parse + xdmp:serialize #

xdmp:html-parse parses HTML into a node tree, and you can then serialize that tree to XHTML using xdmp:serialize with the xhtml method. This approach often preserves HTML5 elements better in older MarkLogic versions:

let $html := "<article>My content</article>"
let $parsed-html := xdmp:html-parse($html)
return xdmp:serialize($parsed-html, 
  <options xmlns="xdmp:serialize">
    <method>xhtml</method>
    <indent>yes</indent>
    <omit-xml-declaration>no</omit-xml-declaration>
  </options>
)

2. Custom XSLT transformation #

You can write an XSLT stylesheet that converts HTML to XHTML while preserving all elements. Here’s a minimal example:

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output method="xhtml" indent="yes" encoding="UTF-8"/>
  
  <!-- Copy all elements and attributes -->
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>
  
  <!-- Convert HTML attributes to XHTML-compliant ones if needed (e.g., "align" to inline style) -->
  <xsl:template match="@align">
    <xsl:attribute name="style">text-align:<xsl:value-of select="."/>;</xsl:attribute>
  </xsl:template>
</xsl:stylesheet>

Invoke it in XQuery like this:

let $html := "<article align='center'>Centered content</article>"
let $xslt := xdmp:document-get("/path/to/your/stylesheet.xsl")
return xdmp:xslt-invoke($xslt, xdmp:html-parse($html))

This gives you full control over the conversion process, including which elements to preserve or modify.

内容的提问来源于stack exchange，提问作者Pavan Sky