PDF output pipeline

The PDF process (formerly known as PDF2) runs the preprocess routine, followed by a series of additional targets. These steps work together to create a merged set of content, convert that to XSL-FO, and then format the FO file to PDF.

The PDF process adds many new Ant targets. During a typical conversion from map to PDF, the following targets are most significant.

  • map2pdf2 creates a merged file by calling a common Java merge module. It then calls publish.map.pdf to do the remainder of the work.
  • The publish.map.pdf target does some initialization, and then calls transform.topic2pdf to do the remainder of processing. That target runs all of the following steps.
    • transform.topic2fo is used to convert the merged file to an XSL-FO file. This process is composed of several Ant targets.
      • transform.topic2fo.index runs a Java process to set up index processing, based on the document language. This step generates the file stage1.xml in the temporary processing directory.
      • transform.topic2fo.flagging sets up preprocessing for flagging based on a DITAVAL file. This step generates the file stage1a.xml in the temporary processing directory.
      • transform.topic2fo.main does the bulk of the conversion from DITA to XSL-FO. It runs the XSLT based process that creates stage2.fo in the temporary processing directory.
      • transform.topic2fo.i18n does additional localization processing on the FO file; it runs a Java process that converts stage2.fo into stage3.fo, followed by an XSLT process that converts stage3.fo into topic.fo.
    • transform.fo2pdf converts the topic.fo file into PDF using the available FO processor (Antenna House, XEP, or Apache FOP).
    • delete.fo2pdf.topic.fo deletes the topic.fo file, unless otherwise specified by setting an Ant property or command line option.
Parent topic: DITA Open Toolkit Architecture