MifParser

Parses Adobe FrameMaker MIF files.

document = MifParser.parse(io_or_string)

document.each do |element|
  # element.type         :heading | :body | :list | :table | :cell
  # element.text         plain text
  # element.html_text    same text with <b> / <i> / <u>
end

Pipeline: syntax → structure (Para / Tbl) → list classification → typed elements.

MIF has no list construct; lists are inferred from tags and PgfNumString markers.

Elements

Document is enumerable (#paragraphs, #lists, #tables).

Class	Meaning
Paragraph	heading or body (heading?, heading_level)
List	:ol / :ul, marker, level
Table	rows is a grid of Cells
Cell	nested Paragraph / List

Heading and list levels are 0-based (Title1 / first list item → 0). An explicit heading style wins over numbering (Title4 stays 3 even if numbered 4.3).

A typical importer switches on the element itself:

document.each do |element|
  next unless %i[heading body list].include?(element.type)

  text = element.html_text.to_s.strip
  next if text.empty?

  case element.type
  when :heading
    element.heading_level
  when :body
    text
  when :list
    element.list_type
    element.list_level
    element.list_marker
  end
end

Format

document.catalog is PgfTag → Format from <PgfCatalog>. Each paragraph/list has:

• format — catalog entry plus local <Pgf> overlay
• runs — character spans from inline <Font>
• html_text — HTML view of the runs

An empty <Font> resets to the paragraph format.

Collected MIF tags (add names to collect more):

MifParser::Format::PROPERTY_TAGS
# => %w[FWeight FAngle FUnderlining]

format.bold?, format.italic?, format.underline? and format.to_h read those properties. text stays plain so you can render HTML, Markdown, or anything else from format / runs.

License

MIT