---
title: AsciiDoc Parsing
...

# Document Structure

A *document* is a series of *block elements*.

The first element is optionally a *document header*.

The next element is optionally a *preamble*.

* Document
    * Header (optional)
      Must be separated from the remainder of the document by one or more
      blank lines and cannot contain blank lines.
      Can include comments.
      Can include attribute entries, typically *doctype*, *lang*, *encoding*,
      *icons*, *data-uri*, *toc*, *numbered*
      Command-line attributes override header attributes.
        * Title
        * Info (optional)
            * AuthorInfo
              Format: firstname[ [middlename ]lastname][ <email>]
              Multi-word names should be separated by underscores,
              e.g. van_Gogh.
              If the format isn't matched, the entire line becomes a
              FirstName.
                * FirstName
                * OtherNames (optional)
                    * MiddleName (optional)
                    * LastName
                * EmailAddress (optional)
            * RevisionInfo (optional)
              * Either
                * RevisionNumber (optional)
                * RevisionDate
                * RevisionRemark (optional)
              * Or
                * RCS/CVS/SVN $Id$ marker.
                  The *revnumber*, *revdata*, and *author* attributes are
                  extracted and set and displayed in the header.
                  The header author line may be omitted.
                  
    * Preamble (optional)
        * SectionBody
    * Section (0 or more)
      Up to 4 section levels (1-4) in addition to the document title (level 0)
      Has attributes *level* and *sectnum* implcitly defined
        * Title
        * SectionBody (optional)
            * One of (repeated 1 or more times)
                * Either
                    * BlockTitle (optional)
                      Format: .<title>
                    * Block
                * Or
                    * BlockMacro
        * Section (0 or more)

* Block
  Floating titles probably belong in here somewhere.
    * One of
      * Paragraph
      * DelimitedBlock
      * List
          * One of
              * BulletedList
                  * ListItem (1 or more)
              * NumberedList
                  * ListItem (1 or more)
              * CalloutList
                  * ListItem (1 or more)
              * LabeledList
                  * ListEntry (1 or more)
                      * ListLabel
                          * ListTerm (1 or more)
                      * ListItem
                          * ItemText
                          * One of (repeated 0 or more times)
                              * List
                              * ListParagraph
                                A paragraph with *listelement* option set
                              * ListContinuation
      * Table
    * Or (these can appear almost anywhere)
      * BlockId
        Applies to the following *Title*, *Paragraph*, *List*,
        *DelimitedBlock*, *Table*, or *BlockMacro*
        Format: [[<id>]]
      * AttributeEntry
      * AttributeList
        Applies to the following *Block*
        Most blocks support *id*, *role*, and *reftext* attributes
        Format: [<attr>(, <attr>)*]

Notes:
  * Block elements are separated by line boundaries
  * *Header*, *Title*, *Paragraph*, and *ItemText* cannot contain blank lines.


```.haskell
data Document = Doc
  { _docAttrs :: Attributes
  , _header   :: Maybe DocHeader
  , _preamble :: Maybe SectionBody
  , _sections :: Seq Section
  }


data Section = Sec
  { _secAttrs    :: Attributes
  , _secTitle    :: Text
  , _secBody     :: Maybe SectionBody
  , _secChildren :: Seq Section
  }

type SectionBody = Seq Block
  
```