Sphinx-based blogging engine
HgBlog is a set of modifications to the Sphinx project to make it slightly more suitable as a blogging engine. It’s built for those of us who love using reStructuredText markup to write documents.
HgBlog assumes a level of familiarity with RST and Mercurial. You can certainly use and enjoy using HgBlog if you’ve never used either one of them. I recommend reviewing a tutorial for Mercurial if you’ve never used it or are unfamiliar with how Mercurial affects your life.
The quickstart wizard handles setting up an HgBlog for you. This includes all of the usual things that the Sphinx quickstart utility does, but it creates a Mercurial repository and installs a hook and intelligent ignores for you. The hook will automatically convert the .rst files that Mercurial is tracking into HTML using Sphinx when you commit changes to the repository.
Additionally, when you pull changes in from a remote clone of the repository, the hook will do the conversion just like when you commit locally. You can set the hook up on remote clones as well. The hook only converts .rst files that are tracked by Mercurial. This means you can work on new blog articles without committing them to the repository to have them not appear online.
Yes, there are. And most of them rely on databases that require regular maintenance and backup. Databases can also slow down your blog. HgBlog offers you a way to serve up your blog articles as static HTML without the overhead of requesting an object from a database, making it fit into a layout, etc. Any webserver should be perfectly capable of serving the content generated by HgBlog.
I’m not saying there’s anything wrong with database-backed blogs. I maintain my own blog that is Django powered (and database-backed). It works fine for me. However, some people might not want to be confined to the rules imposed by a full-on blogging engine (whatever they may be). People have all sorts of reasons for doing things differently. Some people don’t need a reason at all. It boils down to what works for you.
What does HgBlog offer you that should be attractive?
Speed. No need to deal with the formatting headaches of whatever WYSIWYG editor your blogging engine has dictated is the best. Just use reStructuredText markup (which is quite easy to learn if you’ve never used it before) and let Sphinx worry about formatting it.
Consistency. Again, reStructuredText is a very simple format that will produce consistent, nicely-formatted documents.
Portability. Since HgBlog generates static HTML, you can put it on any server. In fact, you don’t even need any server software–just a web browser. Also, Sphinx allows you to export your articles in several formats:
- HTML, multiple files
- HTML, single file
- LaTeX PDF
- Plain text
- man pages
With other tools, you can even turn your .rst files into PDF or ODT documents.
Redundancy. Since every article you want to have on your blog must be checked into Mercurial, a distributed version control system. This means that you can easily clone your blog to another system, which is a very fast and effective way to backup your articles. If the primary “server” for your blog ever dies, you are likely to have at least one full, up-to-date backup of your blog if you’re using Mercurial as it’s designed.
I’ve developed and tested HgBlog using Linux, Python 2.6.4, Mercurial 1.5.1, Sphinx 1.0-pre, docutils 0.6, Jinja2 2.4.1, and Pygments 1.3.1. However, Sphinx suggests the following version requirements. I’m just being safe with my requirement on Mercurial’s version.
There are several ways to install HgBlog:
Using pip (recommended):
pip install -U hgblog
From the CheeseShop
- Download the .tar.gz file from PyPI
- Extract the .tar.gz file
- Run python setup.py install using the setup.py in the extracted directory
hg qclone http://bitbucket.org/codekoala/hgblog cd hgblog hg qapply -a python setup.py install
HgBlog leverages the existing quickstart wizard for Sphinx projects. There are some modifications to reduce the number of steps required, so you should be able to be up and running within a minute using:
All you need to do is:
Once you do that, you should have a few new directories, one of which will be called source. This is where you should write your .rst articles. When you’re done working on a particular article, you can use:
hg add hg ci
…to add and commit it to your Mercurial repository. At this point, Sphinx will be asked to generate the HTML for your blog based on your .rst files.
If you feel like using Mercurial to clone your blog articles to another system, you might be interested in adding to the new repository the same hooks that are installed by the quickstart utility. First off, this requires HgBlog to be installed on the other system. Next, edit the .hg/hgrc file for the new repository:
[hooks] update.hgblog = python:hgblog.generate_html.htmlize_articles commit.hgblog = python:hgblog.generate_html.htmlize_articles
These hooks make it so the HTML version of your pages will be generated each time you commit changes to the local repository and each time you update your local repository using changesets pulled in from elsewhere.
I realize that there are probably several times you might want to have a gander at the resulting HTML for your .rst files at some point before committing changes to the repo. I’ve created a simple command to make this possible:
You should be able to call that from anywhere within your blog’s Mercurial repository, and your HTML files should be properly refreshed.