XPathExpression to select from tesseract hOCR XML output

Question

I have a file of roughly the following shape:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
 <head>
  <title></title>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
  <meta name='ocr-system' content='tesseract 3.02' />
  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word'/>
 </head>
 <body>
  <div class='ocr_page' id='page_1' title='image "D:\DPC2\converted\60\60.tiff"; bbox 0 0 2479 3508; ppageno 0'>
       <!-- LOTS OF CONTENT -->
  </div>
 </body>
</html>

Then I am using JDOM 2.x with the following XPath query:

//htmlFile is an input variable of type java.nio.Path
Document document = xmlBuilder.build(htmlFile.toFile());

XPathFactory factory = XPathFactory.instance();
XPathExpression<Element> xpePages = 
    factory.compile("//html/body/div[@class='ocr_page']", Filters.element());
List<Element> pages = xpePages.evaluate(document);

But it is never able to find any elements, what am I doing wrong in query?

Ian Roberts · Accepted Answer · 2014-04-30T10:36:20.957

Namespaces.

The xmlns="http://www.w3.org/1999/xhtml" means that elements with no prefix in the XML file are actually in the http://www.w3.org/1999/xhtml namespace, and you need to specify this in the XPath expression using a prefix:

XPathExpression<Element> xpePages = 
    factory.compile("/h:html/h:body/h:div[@class='ocr_page']",
                    Filters.element(),
                    null, // no variables
                    Namespace.getNamespace("h", "http://www.w3.org/1999/xhtml"));

You must use a prefix, as in XPath no prefix always means no namespace.

score 2 · Answer 2 · edited May 23 '17 at 10:32

<html xmlns="http://www.w3.org/1999/xhtml"

Means that elements like html are in namespace http://www.w3.org/1999/xhtml

You've got a couple of ways forward

To register a NamespaceContext (~name space manager), which looks rather onerous in this technology stack: https://stackoverflow.com/a/6390494/314291
Alternatively to use namespace-agnostic xpath, e.g.

 //*[local-name()=='html' and namespace-uri()='http://www.w3.org/1999/xhtml']
 /*[local-name()='body' and namespace-uri()='http://www.w3.org/1999/xhtml']
 /* ... etc.

If you are confident that there is no conflict in the namespaces of the elements, you can choose to use just local-name()

//*[local-name()=='html']/*[local-name()='body']* ...

This is the JDOM `XPathFactory`, not the `javax.xml.xpath` one - namespaces are much easier in JDOM. — Ian Roberts, Apr 30 '14 at 10:33

XPathExpression to select from tesseract hOCR XML output

2 Answers2