I’ve now updated the blog software powering my blog, which is very long overdue. In the past, this blog was run off Typo, which apparently now hosts their home site off the github (its moved a dozen times in the past 3+ years).
Typo always worked moderately well for me, however, I found it sluggish (Prolly Rails there), and incredibly horrid on ram usage. It was not at all unusual to see it running past 700 megs of RAM after running for just a few weeks, which is a bit annoying as the machine only has 4GB total and is running quite a few things.
After last weeks SF Python Meetup on MongoDB I figured it was time to get a little actual MongoDB usage under my belt. I also inadvertently implemented enough of the MovableType XMLRPC API as I didn’t want my app to be too extensive, just enough for me to post to my blog.
So in the end, I had a small set of requirements for the replacement:
I wanted to build it myself, because of course, the world definitely needs more blog apps, and I wanted one that used MongoDB. So for those curious, here’s the source code to the blog app.
It’s rather rough, as its fairly custom built just for my needs, nor do I have any plans to expand it into some general purpose blog engine, with themes, etc. The only other thing pending at the moment is to add the ability to comment again, as I haven’t quite gotten that feature in yet. For those trying it out, the README should help get started, but its very rough (thus the name of the package).
When copying some functionality I needed from the Rails app (to retain URL compatibility), I noticed two things it did which I thought was handy. To convert a title into slug for the article, it used a fairly sophisticated scheme relying on two other packages.
First, was the use of a Ruby port of a Perl package called Text::Unidecode which is pretty cool, and converts UTF-8 chars into their closest ASCII version. I figured someone must’ve ported it to Python as well, and sure enough, someone did! It wasn’t on the cheeseshop though, which was unbearable for me, so I’ve posted it to the Cheeseshop so others can easy_install it.
Next up, was a Ruby library called stringex, which add’s a few things to Rails, including a string method called ‘to_url’. That method does a variety of transformations to remove all those funny characters from a title, and do a bunch of other neat changes of common characters to human readable versions (source for those conversions).
I ported the key module of stringex to Python, and it resides in my blog app. If someone would like to extract it and make it into its own package, or even better, if I somehow missed the fact that someone else has ported it already, let me know (tweet me @benbangert!).
I’ll be writing up my thoughts on making a small app with MongoDB, and how it differs from my experience working with CouchDB for PylonsHQ in a later post for those curious.