#!/bin/bash # # This script installs DocFX and brings all content pages from EFL's documentation # content repository, transforming them from DokuWiki's format to DocFX's. # This step only needs to be done once (or every time the content pages change). # After this step, you can run the gendoc.sh script as many times as you want, # to generate the actual HTML pages, including the reference guide extracted from # libefl_mono.dll and libefl_mono.xml # # This script could be improved a lot, but it gets the job done. # Xavi (xartigas@yahoo.es) # # Check mono version since DocFX has been known to fail with mono 4. # mono_version="$(mono --version | awk '/version/ { print $5 }')" if [ $mono_version \< 5 ]; then echo "Requires at least mono version 5, detected $mono_version" exit 1 fi; # # Download and extract DocFX # if [ ! -d "bin" ]; then rm -rf docfx.zip bin wget https://github.com/dotnet/docfx/releases/download/v2.45/docfx.zip unzip -q docfx.zip -d bin rm docfx.zip else echo "Skipping DocFX download because 'bin' folder already exists." fi # # Clone whole Content site # rm -rf www-content git clone --depth 1 https://git@git.enlightenment.org/website/www-content.git www-content # # Copy all pages related to C# (those inside a folder called 'csharp') to the articles folder # rm -rf articles mkdir articles # Copy them find www-content -wholename "*/csharp/*.md.txt" -exec cp --parents {} articles \; # Except tutorials and guides summary pages (start.md) find articles/www-content/pages/develop/tutorials -name "start.md.txt" -exec rm {} \; find articles/www-content/pages/develop/guides -name "start.md.txt" -exec rm {} \; # Remove the trailing .txt from filenames (DocFX wants only the .md) for f in `find articles -name "*.md.txt"`; do mv $f $(echo $f | rev | cut -c5- | rev) ; done # Copy all media files currently in use to the images folder rm -rf images mkdir images cd www-content/media for f in `find ../../articles -name "*.md" | xargs grep -Poh '(?<=/_media/)[^)]*'`; do cp -f --parents $f ../../images; done cd ../.. # Remove git clone now that we have everything we wanted rm -rf www-content # # Parse all Markdown files to adapt them to DocFX # # Remove special Dokuwiki headers like ~~NOCACHE~~ because DocFX complains about them. find articles -name "*.md" | xargs sed -i -e "s/^~~.*~~$//g" # Transform media links in articles from /_media/* to ~/images/* find articles -name "*.md" | xargs sed -i -e "s#/_media#~/images#g" # Transform API root links from /develop/api/ to @api-reference-root (the UID of the API root) find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\](\)/develop/api/)#\1xref:api-reference-root)#g" # Transform API class links from */api/namespace/class to @Namespace.Class (that should be the UID of the reference page) find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\](\)[^)]*/api/\([^/.]*\)/\([^/.]*\))#\1xref:\u\2.\u\3)#g" # Point links to Legacy API to the www.enlightenment.org site find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\](\)\(/develop/legacy/api/[^)]*)\)#\1https://www.enlightenment.org\2#g" # Point the rest of links to /develop/api/* to the root the API, since we don't know how to transform them find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\](\)/develop/api/[^)]*)#\1xref:api-reference-root)#g" # Add start.md to links pointing to a folder inside ~/develop/ find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\](/develop/[^)]*/\))#\1start.md)#g" # Transform absolute links from /develop/* to ~/articles/www-content/pages/develop/* find articles -name "*.md" | xargs sed -i -e "s#\(\[[^]]*\]\)(\(/develop/[^)]*\))#\1(~/articles/www-content/pages\2)#g" # Transform page anchor links from [label](#Anchor_Title) to [label](#anchor-title). for f in `find articles -name "*.md"`; do awk -v RS=' ' '{ if ($0 ~ ".*\(#.*\)") { gsub( /_/, "-" ); printf "%s ", tolower($0) } else { printf "%s ",$0 } }' $f > /tmp/efl_docfx_setup.tmp mv /tmp/efl_docfx_setup.tmp $f done # # Populate articles TOC file from the articles' titles # cd articles rm -f toc.yml echo "- name: Setup Guide" >> toc.yml echo " href: www-content/pages/develop/setup/csharp/start.md" >> toc.yml echo "- name: Tutorials" >> toc.yml echo " items:" >> toc.yml find www-content/pages/develop/tutorials -name "*.md" -exec sh -c "cat {} | grep '^# ' && echo ' href: {}'" \; >> toc.yml echo "- name: Guides" >> toc.yml echo " items:" >> toc.yml find www-content/pages/develop/guides -name "*.md" -exec sh -c "cat {} | grep '^# ' && echo ' href: {}'" \; >> toc.yml sed -i -e "s/^# \(.*\) #/ - name: '\1'/g" toc.yml sed -i -e "s/ in C#//g" toc.yml cd ..