You probably noticed this blog dramatically changed lately and was sometimes broken. That’s because I was in the middle of migrating it from WordPress to a static site generator.

In the last few years, tweaking and maintaining my server proved to be a challenging but rewarding experience. Most of the knowledge I acquired in the process has landed in my bag of tricks and makes me more productive than I was before.

But now that I’ve learned what I had to learn, the adventure turned into a chore. It’s time to save time. And money. And ditch this machine for good.

Another factor in this decision was the end of Cool Cavemen . My band announced their last album to be released this year. Since the beginning we were armed with a website, an online shop, multiple internal tools and a strong web presence to support our activities. There’s no longer any need of all of these.

So I closed the Drupal shop and moved it to Bandcamp . I merged ZenPhoto galleries into WordPress . All WordPress sites I hosted moved to . bbPress forums were closed: I archived private forums to emails , and public ones to plain WordPress content . As for Mailman mailing-lists, they were archived and the ones remaining active moved to Google Groups.

Listing all of this work today makes me realize that I spent most of 2012 preparing the retirement of my server. And my band. Until only one thing remained: this blog.

The next evolution of this site couldn’t be anything but a bunch of statically generated pages. First, static websites are cool. They’re also fast and cheap. And they don’t requires a full nginx/PHP/MySQL stack to be managed, upgraded and monitored  .

I started my quest for a tool:

  • written in Python,

  • supporting Markdown and

  • supported by an active community.

After a little survey, the remaining contenders were Pelican and Nikola  .

If I like Nikola for its fast release cycles and huge list of configurable features, I ultimately choose Pelican. I found it was easier for the latter to naturally match my previous WordPress URLs (good for SEO).

So, let’s install Pelican and its dependencies:

$ aptitude install python-markdown python-pygments python-beautifulsoup pandoc
$ pip install pelican mdx_video

And create a new site:

$ mkdir blog
$ cd ./blog/
$ pelican-quickstart
Welcome to pelican-quickstart v3.1.1.

This script will help you create a new Pelican-based website.

Please answer the following questions so this script can generate the files needed by Pelican.

> Where do you want to create your new web site? [.]
> What will be the title of this web site? Kevin Deldycke
> Who will be the author of this web site? Kevin Deldycke
> What will be the default language of this web site? [en]
> Do you want to specify a URL prefix? e.g.,   (Y/n) Y
> What is your URL prefix? (see above example; no trailing slash)
> Do you want to enable article pagination? (Y/n) n
> Do you want to generate a Makefile to easily manage your website? (Y/n) Y
> Do you want an auto-reload & simpleHTTP script to assist with theme and site development? (Y/n) Y
> Do you want to upload your website using FTP? (y/N)
> Do you want to upload your website using SSH? (y/N) Y
> What is the hostname of your SSH server? [localhost]
> What is the port of your SSH server? [22]
> What is your username on that server? [root]
> Where do you want to put your web site on that server? [/var/www]
> Do you want to upload your website using Dropbox? (y/N) N
Done. Your new project is available at /home/kevin/blog

Basically, that’s it. You’re now free to tweak the configuration and structure of your site.

The next important step is to migrate our content. We can’t import the XML file WordPress produces as is. We need to massage some data first.

During my Disqus migration , I had to update the markup of the code blocks. Same thing apply here. I need to replace occurrences of:

[code lang="shell"]

by this kind of pure HTML:

<pre><code class="shell">

The magic command to perform that job is:

$ sed -e 's/\[\/code\]/<\/code><\/pre>/g' -e 's/\[code lang=\(.*\)\]/<pre><code class=\1>/g' < ./wordpress.xml > ./wordpress-fixed.xml

Pelican consumes by default reStructuredText content, but it supports Markdown too. To convert WordPress content to Markdown, the command is:

$ pelican-import --wpfile -m markdown -o ./content/ ./wordpress-fixed.xml

The thing is, Pelican’s built-in importer produces files requiring, in my opinion, too much cleaning afterwards.

I was right to choose Markdown over reStructuredText: the former is much more popular that the latter. Plenty of tools are available. exitwp is the one I was looking for. It’s a WordPress to Jekyll importer written in Python.

Let’s fetch it:

$ git clone

Before using it, I had to apply a tiny patch to transform Jekyll metadata to Pelican’s:

$ cd exitwp
$ git rev-parse HEAD
$ git diff
diff --git a/ b/
index fad5a4e..61e626e 100755
--- a/
+++ b/
@@ -265,7 +265,7 @@ def write_jekyll(data, target_format):
            yaml_header['published'] = False

        if i['type'] == 'post':
-            i['uid'] = get_item_uid(i, date_prefix=True)
+            i['uid'] = get_item_uid(i, date_prefix=False)
            fn = get_item_path(i, dir='_posts')
            out = open_file(fn)
            yaml_header['layout'] = 'post'
@@ -311,13 +311,15 @@ def write_jekyll(data, target_format):

-            out.write('---\n')
            if len(yaml_header) > 0:
            if len(tax_out) > 0:
-                out.write(toyaml(tax_out))
+                for tax_type, tax_values in tax_out.items():
+                    if tax_type == 'categories':
+                        tax_type = 'category'
+                    out.write("%s: %s\n" % (tax_type, ', '.join(tax_values)))
+            out.write('\n')

-            out.write('---\n\n')
                out.write(html2fmt(i['body'], target_format))

Now call exitwp and move all Markdown files to Pelican:

$ rm -f ./wordpress-xml/*
$ rm -rf ./build
$ cp ../wordpress-fixed.xml ./wordpress-xml/
$ python
$ cd ..
$ mv ./exitwp/build/jekyll/* ./content/
$ rename "s/\.markdown/\.md/g" ./content/*.markdown

We still have to clean-up articles. Like convert HTML code blocks to use Pelican’s tag:

$ find ./content/ -iname "*.md" -exec perl -0777 -i -pe "s/<code class=\"(.*?)\">(.*?)<\/code>/:::\1\2/gs" "{}" \;

We can also fix some metadata like authors, and misquoted titles produced by exitwp:

$ find ./content/ -iname "*.md" -exec sed -i 's/Author: Admin/Author: Kevin Deldycke/' "{}" \;
$ find ./content/ -iname "*.md" -exec sed -i "s/^title: '\(.*\)'/title: \1/" "{}" \;

I’ll take the opportunity to consolidate some tags:

$ find ./content/ -iname "*.md" -exec sed -i "s/tags:\(.*\)leopard\(.*\)/tags:\1Mac OS X Leopard\2/" "{}" \;
$ find ./content/ -iname "*.md" -exec sed -i "s/tags:\(.*\), Mac,\(.*\)/tags:\1, MacBook,\2/" "{}" \;

I also had to fix my YouTube’s tags:

$ find ./content/ -iname "*.md" -exec sed -i "s/^\[youtube \(.*\)/\1/g" "{}" \;
$ find ./content/ -iname "*.md" -exec sed -i "s/\?rel=0\]$//g" "{}" \;
$ find ./content/ -iname "*.md" -exec sed -i "s/\&rel;=0\]$//g" "{}" \;

I decided to get rid of all image thumbnails generated by WordPress. With the command below, I reduced all my thumbnails linking to their full-size version, to their full-size version only:

$ find ./content/ -iname "*.md" -exec sed 's/\[!\[\(.*\)\](.*)\](\(.*\).\(jpe\?g\|png\|gif\))/!\[\1\](\2.\3)/g' "{}" \;

You’re now ready to publish your site to the world:

$ make clean
$ make ssh_upload

If you need an example or inspiration, my current Pelican blog, its theme, configuration and content are all available on GitHub  .