PAGE-XML namespace handling: prefer default over prefixed `pc:` usage

[johnlockejrr]

Hi and thanks for the great work on eynollah!

I noticed that some generated PAGE-XML files use a prefixed namespace (pc:) in the root element, like:

<pc:PcGts xmlns:pc="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15" ...>

While technically valid, this format is less common in the PAGE ecosystem and may cause compatibility issues with tools expecting the default namespace. The more typical and interoperable format is:

<PcGts xmlns="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15" ...>

Would it be possible to adjust the output to use the default namespace instead? This would improve compatibility with validators and downstream tools.

Thanks again!

[bertsky]

The namespace prefix is based on OCR-D core's PAGE-XML object model that was generated by generateDS from the original PAGE-XML XSD schema file, which uses pc itself.

So I am surprised by your statement this is uncommon in the PAGE ecosystem. (But I can see your point in that most public GT datasets don't use a prefix.)

Anyway, it's really easy to post-process XML files to change their namespace prefix via XSLT. In this case, just use...

<xsl:stylesheet version="1.0"
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
  xmlns="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15">
<xsl:output method="xml" version="1.0" omit-xml-declaration="yes" encoding="UTF-8" indent="yes"/>

<xsl:template match="*">
  <xsl:if test="not(starts-with(namespace-uri(),'http://schema.primaresearch.org/PAGE/gts/pagecontent/'))">
    <xsl:message terminate="yes">
      <xsl:text>input document is not of type http://schema.primaresearch.org/PAGE/gts/pagecontent, but </xsl:text>
      <xsl:value-of select="namespace-uri()"/>
    </xsl:message>
  </xsl:if>
  <xsl:element name="{local-name()}">
    <xsl:apply-templates select="@*|node()|text()"/>
  </xsl:element>
</xsl:template>
<xsl:template match="@*|text()">
  <xsl:copy>
    <xsl:apply-templates select="node()|@*|text()"/>
  </xsl:copy>
 </xsl:template>
</xsl:stylesheet>

... (for example via xsltproc rm-prefix.xsl your-page.xml)

[johnlockejrr]

PAGE-XML is a PRImA Research project, take a look at their schema files,they're on github, OCR-D is using it, they didn't develop it.

Original full paper: https://www.primaresearch.org/www/assets/papers/ICPR2010_Pletschacher_PAGE.pdf

No matter, is easy with Python and xml or lxml to change the namespaces.

[bertsky]

PAGE-XML is a Prima Research project, take a look at their schema files,they're on github, Ocrd-d is using it, they didn't develop it.

I just linked to PRImA on Github. It seems you missed the point: they themselves put pc in their schema file, and this is why generateDS picked it up as the default.

[johnlockejrr]

Yes, I know about the xsd schema but the output files I'm not sure they should have namespaces prefixed with pc:. Try Aletheia, Prima Research software to work with such a format never uses pc: prefixed namespaces in their generated PAGE-XML files.

pc: is like a variable in a template.

[johnlockejrr]

So, to be more precise:

The use of a prefixed namespace like pc: in PAGE-XML is technically valid per XML standards, but not idiomatic within the PAGE ecosystem. Is for interoperability, and that’s exactly what the spec intends.

🧠 The Core Issue

✅ What’s Valid

<pc:PcGts xmlns:pc="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15">

This is valid XML. The prefix pc: is just a label for the namespace URI.

✅ What’s Standard

<PcGts xmlns="http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15">

This is the canonical form used by:

• PriMA Research tools (Aletheia, PAGE Viewer)
• Public datasets (e.g., cBAD, DIVA-HisDB, READ)
• PAGE validators and schema-aware tools

🔍 Why Prefixes Cause Trouble

• XPath queries become more complex (pc:Page vs Page)
• Schema validation may fail if tools expect default namespace
• Tooling assumptions break — many PAGE consumers hardcode the default namespace