docx-asciidoc-conversion

Introduction

The objective of the project is to convert the INSPIRE Metadata Technical Guidance 2.0 document from MS Word docx format to AsciiDoc format to facilitate its access and maintenance in a GitHub repository. AsciiDoc was selected over the more common Markdown format because AsciiDoc supports many advanced document features, such as automatic section numbering, that Markdown does not support.

An important issue encountered is that it is not possible for a generic docx to AsciiDoc converter (such as Pandoc used here) to convert all the document content as required, which is mostly due to sections with special formatting, such as TG Requirements, TG Recommendations, Conformance Classes, XML Examples or quotes. To address this issue, a Python script was developed that applies custom styling to relevant sections of AsciiDoc version of the document (obtained using Pandoc converter) based on additional style information included in a Markdown version of the document (obtained using Pandoc as well). Additionally, the Python script fixes other minor issues with the AsciiDoc document, such as section headings, numbering, TOC, etc. Moreover, some of the styles, used to identify relevant sections, are not applied consistently in the docx, so this needs to be corrected in MS Word before the conversion.

Prerequisites

• Pandoc, version 2.14.0.3 used
• Python 3, version 3.8.3 used
• Beautiful Soup Python library, version 4.9.3 used

Conversion Procedure

1. Correct the original docx document, save as md_2_fixed.docx:

   • apply XML Example style to all XML examples (about 7 cases were found in md_2_0_1.docx where this fix was required)
   • remove paragraph breaks from inside footnotes (two cases were found in md_2_0_1.docx where this fix was required)

2. Convert the corrected docx document to AsciiDoc using Pandoc, save as md_2_pandoc.adoc, extract images to media folder:

   pandoc md_2_fixed.docx -f docx -t asciidoc -o md_2_pandoc.adoc --wrap=none --extract-media=. --markdown-headings=atx

3. Convert the corrected docx document to Markdown using Pandoc, enabling the styles extension, save as md_2_styles.md:

   pandoc md_2_fixed.docx -f docx+styles -t markdown -o md_2_styles.md --wrap=none --markdown-headings=atx

4. Run the post_pandoc.py Python script with 3 arguments: adoc file from 2., md file from 3. and attributes.adoc file, it will create md_2_final.adoc file:

   python post_pandoc.py md_2_pandoc.adoc md_2_styles.md attributes.adoc

   The script applies AsciiDoc admonitions to TG Requirement, TG Recommendation and Conformance Class sections, source blocks to XML Examples and blockquotes to quote sections of md_2_pandoc.adoc based on additional style information included inmd_2_styles.md. It also corrects initial (unnumbered) section headings, removes manual TOC entries, inserts automatic TOC. AsciiDoc does not yet support custom section numbering (see this issue), therefore 'Annex X' text is inserted in relevant headings by the script.

5. Replace image1.wmf image with a png version in the media folder.

Included Files

md_2_0_1.docx original docx MS Word, Metadata Technical Guidance 2.0 document, used in 1.

md_2_fixed.docx corrected docx MS Word, Metadata Technical Guidance 2.0 document, created in 1., used in 2. and 3.

md_2_final.adoc final AsciiDoc , Metadata Technical Guidance 2.0 document, created in 4.

attributes.adoc attributes to include in the AsciiDoc document, used in 4.

post_pandoc.py Python script, used in 4.

media folder containing images extracted from docx document by Pandoc, created in 4., modified in 5.