This post is in incomplete walkthrough of converting my old blog that was built with TRAC to a static site generated blog, as of know with jekyll just to get up quickly with github pages.

Dump the site

  1. Grab and archive the existing site, using wget to mirror it so you have a working copy.
wget --mirror --convert-links --adjust-extension --page-requisites

One could find a way to dump records from the existing site database and then convert the content, but that requires customizing the dump. It also doesn’t make a usable archive of the old site that you can dig through.

Convert the html to markdown

Using Pandoc, generally like so pandoc -f html -t markdown

Find the content, and ignore everything else. The content can be located with the following html tag <div id="blog-main"> , and I followed these tips

echo "converting $1"
cat $1 | sed '1,/<div class="blog-main">/d' | sed '/<div class="asset-footer">/,/<\/html>/d' | pandoc --wrap=none --from html --to markdown_strict > $

Won’t quite work, since we just want what’s in the div tag, but not the tag.

Following additional ideas.

echo "converting $1"
xmllint --html --xpath "//div[@id='blog-main']/node()" $1 | pandoc --wrap=none --atx-headers --from html-native_divs-native_spans --to markdown_github > $

github markdown drops the extra attributes from the headers (header_attributes in pandoc, there’s probably a more direct way), but we want those for out mardown headers.

Now we can run the script over all the files to create markdowns of every post.

find *.html -exec ./ '{}' \;
  1. Next in a new repo setup for Jekyll on Github Pages, going to copy all the markdown files in as posts.
  2. Need to also grab all the files from the raw-attachments folder and put them into assets. These are all the images used in posts and files attached (pdfs, etc…) mv raw-attachments/*/* ../blog/assets

We could have added command line args to set some of the meta tags but just added it to a filter that also sets the title based on the top level 1 header.

pandoc _drafts/ -o _drafts/ --filter assets/

# Now do it in batch
find _drafts/*.md -exec sh -c 'pandoc -s {} -o {} --filter assets/' \;

# now update the datetime
find _drafts/*.md -exec sh -c 'pandoc -s {} -o {} --filter assets/' \;

# and then drop the Attachment section onward
find _drafts/*.md -exec sed -i '/Posted/Q' {} \;

Probably going to do this part by hand, there are so few and not standard.

# links to attachments (can be images) need to be fixed
# before

# The regex to find them

# after

#Now apply with sed
find *.md -type f -exec sed 's|../raw-attachment/blog/.*?/|attachments|g' {} + | less
find *.md -type f -exec sed -i 's|../raw-attachment/blog/.*?/|attachments|g' {} +

# The more complicated ones, need to also drop the .html at the end