It’s very hard to convert an unstructured document like a PDF into the logical structure of an XML file - automatically.
XML (Extensible Markup Language) is a format based around simplicity and general usability across the internet. It is a format that is both human-readable and machine-readable. Much modern publishing utilises XML. It creates a clear hierarchy of each element’s function through a series of logical tags which identify each part of the text.
For example, your title would be tagged as: <title>example title</title> and if you had a header following the title, it would be: <head>example header</head>. These logical tags allow both computers and humans to read XML documents which would otherwise be unstructured with no identifying tags. XML was designed to carry data and focus entirely on what that data is. As a format, XML makes it incredibly easy to transfer, share or change data.
In light of this, a publishing firm will often use specific and detailed templates that will take an XML file and convert it to an HTML or PDF file. The template would add the layout according to the publisher’s rules. It’s a standard procedure to go from XML to PDF, but the other way round – converting from PDF to XML– is a headache.
Zanran’s technology allows you to take PDFs and create XML files from them. To convert PDF to XML, Zanran’s technology uses a mixture of layout and semantic analyses which enables the technology to extract specific types of content. The content can be metadata (authors’ names, titles, etc) or sections/chapters with headers and paragraphs.
