This Ruby library (gem) parses and translates .docx Word documents into kramdown syntax, which allows for easy subsequent translation into html or TeX code via the excellent kramdown library. kramdown is a superset of Markdown. See http://kramdown.gettalong.org/ for more details.

A .docx file as written by modern versions of Microsoft Office is just a .zip file in disguise. It contains a directory tree containing XML files. Parsing of these compressed XML trees is rather staightforward, thanks to the zip and nokogiri Ruby libraries.

docx-converter contains a parser which translates all common Word document entities into corresponding kramdown syntax. It extracts images and converts them into .jpg files with a maximum width or height of 800 pixels.

Output files and directories will be created according to the webgen conventions. This is useful when you want to generate a static website with the webgen gem after you have converted your .docx file into html. The file naming is in the format ss.nnnn.ll.page, where ss is a 2-digit sort number, nnnn is the main file name, ll is the language code. For more information on webgen see http://webgen.gettalong.org/

docx_converter was written for our project publishr_web, see http://documentation.red-e.eu/publishr/index.html

Supported Word elements:

• Paragraph
• Line break
• Page break
• Bold
• Italic
• Paragraph styles "Heading1", "Heading2" and "Title"
• Character styles "Strong" and "Quote"
• Footnotes
• Tables
• Images including captions
• Non-breaking spaces

On Debian Linux:

apt-get install libmagic-dev apt-get install libmagickwand-dev gem install docx-converter

Look into the .gemspec file to see all gem dependencies.

Installation may vary on other operating systems.

From the command line:

docx-converter inputfile format output_directory

format can be either kramdown, html or latex. For example:

docx-converter ~/Downloads/testdoc1.docx latex /tmp/docxoutput

output_directory will be created if it doesn't exist. A subdirectory /src will be created by default, which is merely a convention to be identical with the webgen file system standard.

If you want to use docx_converter from a Ruby script, you can use the API like this:

r = DocxConverter::Render.new(options)
rendered_filepaths = r.render(:html)

options is a hash with the following keys

• :output_dir: The directory to be created for the output files. A subdirectory /src will be created by default, which is merely a convention to be identical with the webgen file system standard.
• :inputfile: The path to the .docx file to be parsed
• :image_subdir_filesystem: The subdirectory name into which images will be put. It will be created below the /src subdirectory.
• :image_subdir_kramdown: Usually this is identical to :image_subdir_filesystem and should only be different when you do further manual postprocessing with the kramdown output. This string will be added as a prefix for images in the final kramdown output. An example: ![image description](/image_subdir_kramdown/imagename.jpg).
• :language: The language to be used for the generated file names. See webgen conventions above.
• :split_chapters: when true, the output files will be split between headings which have the Word paragraph style "Heading1". This is useful for large documents. When false, no splitting is done and all content will be output to the file 01.chapter01.ll.page. Footnotes will be split correctly into the various chapters.