Like several others I've decided to switch my blog to a non-blog platform. I did this mainly because the whole blog platform web interface approach is not something that works well for me. Often when I think about writing something I want to focus on the writing and not on the whole process of creating a post. Instead, what usually happens is that I log into my blog platform see 1,400 spam comments, feel compelled to delete them all, and then give up on the whole idea. Dealing with media was always another hassle to deal with. In the end I knew I wanted to streamline the process.

Luckily, while I was rolling this around in my head others were doing the same thing. It turns out there's quite a few options out there in the Python world for using static blog generators to deliver up content. In the end I went with blogofile, but customized it a bit for my needs. With blogofile I'm able to write up a post in ReStructuredText format, commit it to Git, and push the repository to my server. From there a Git post-receive hook handles the process of rebuilding my blog. It's a simplified process that allows me to focus on writing content and not the process of getting a blog post out there.

One thing you'll notice is that my blog no longer has comment support. I'm on the fence on this one. Comments can often be helpful to further clarify, distill, or correct things that are written. But, dealing with spam is a hassle and not something I have the energy for. If I do enable comments it will be done by using Disqus as others have done. Until then, just contact me through twitter.

def from_html_to_rst(content):
    p1 = Popen(["/Users/mtrier/.cabal/bin/pandoc",
                "--reference-links",
                "-r",
                "html",
                "-t",
                "rst"], stdin=PIPE, stdout=PIPE)
    p1.stdin.write(content)
    p1.stdin.close()
    return p1.stdout.read()

posts = session.query(Post).filter(Post.status==2).order_by(desc(Post.pub_date))

for post in posts:
    year_directory = os.path.join(post_directory, str(post.pub_date.year))
    mkdir_p(year_directory)
    with open("%s/%s-%s.rst" % (year_directory,
                                str(post.post_id).zfill(4),
                                post.slug.split('/')[-1]), 'w') as f:
        # write out YAML
        f.write('---\n')
        f.write("title: %s\n" % post.title)
        f.write("date: %s\n" % post.pub_date.strftime("%Y/%m/%d %H:%M:%S"))
        f.write("tags: %s\n" % ', '.join([p.name for p in post.tags]))
        f.write("categories: %s\n" % ', '.join([c.name for c in post.categories]))
        f.write("author: Michael Trier\n")
        f.write('---\n')
        try:
            f.write(from_html_to_rst(post.text))
        except Exception as e:
            print "Post %s can't be decoded." % post.post_id

$ brew install haskell-platform
$ cabal update
$ cabal install pandoc