Pagination

If your documents are very long, it can help to enable pagination. This will only show a section of a document on the page at a time, and the user can scroll through the document using the pagination controls.

Enable Pagination

Pagination is configured through the search.xml configuration file.

<Article>
    <!--
    Set to true to enable pagination when viewing documents.
    This setting is not enabled by default, as it may result in issues displaying the document's content,
    because xsl files will not run on the complete document any longer, thus xpaths may fail unexpectedly.
    -->
    <Pagination>false</Pagination>
    <!-- When pagination is enabled, this controls the length of a page. -->
    <PageSize>1000</PageSize>
</Article>

By default, the page size is set to 1000 words.

Pagination Pitfalls

 

If pagination is enabled in your corpus, content outside the current page will not be present in the processed XML.
BlackLab will automatically fixes "unbalanced tags" i.e. automatically opens/closes truncated elements, so the xml should still be valid, but elements that completely lie outside or around the current page will be missing!
Try not to rely too much on the absolute structure of the XML.

Illustration

Imagine only the focused lines are part of the current page, all blurred content is outside the current page.

<doc>
  <meta>
	<title>my document</title>
	<date>01-01-2000</date>
  </meta>
  <contents xmlns="my-namespace">
	<text>
	  <p>
		Here be some text outside the current page
	  </p>
	  <p> 
		This text is on the page 
	  </p> 
	</text>
  </contents>
</doc>

• Absolute xpath like /doc/contents/text/p will fail in this case.
• xpath relying on the namespace will fail in this case, as the xmlns declaration namespace is not on the page.
• Even a simpler xpath like //text will fail (the <text> element is not on the page).

Instead, try to use simpler xpaths, and compare element names using local-name() if your documents are namespaced: e.g. //*[local-name() = 'text'] instead of //my-namespace:text.

Due to how BlackLab is set up, this is a hard problem to solve unfortunately.