Sergio:Botero

From Octopress to Pelican

If I want to blog about something the last thing I need is the blogging framework getting in the way.

I don't blog much, but every time I wanted to write something I had to basically relearn everything about my previous 'blog framework' to get that working.

I was spending way too much time on 'not writing', making me delaying any idea I wanted to put up. I considered moving to tumblr or something like it, but that was being too radical about it (also I like to feel at least some ownership on the process).

Recently I came across a blog post from Amy Hanlon about Migrating to GitHub Pages using Pelican. The process was very easy and motivated me to migrate from Octopress.

The migration was really easy and this is what I did:

I followed the Pelican setup instructions (I left the fabric part out).
I moved my Markdown files from the previous blog and changed the metadata for the Pelican fields.
Added a new section I never published with the previous blog.
I made a couple of modifications to the pelicanconf.py file.
Publish to Github.

I had to spend some time on step #2 because I was basically editing out the metadata fields, but that is something that you can get done pretty easily once you master some keyboard shortcuts.

The new section I wanted to put up is based and constructed using IPython notebooks. A notebook can be exported as HTML and embedded into a blog. I added a couple of meta tags to the HTML head and Pelican accepted it, no questions asked.

At this point I was really happy with the outcome and was ready to push the new blog to Github, when I noticed something.

In the previous blog, the routes where:

http://sergiobuj.github.io/blog/YEAR/MONTH/DAY/POST-TITLE/

And now I had

http://localhost:8888/POST-TITLE/

In my case this is not very important because of the traffic, but there is a fundamental idea about those resources 'posts' that I would be breaking if I had let this happen.

Imagine someone bookmarked one of your blog posts that they really liked and want to come back and read it again. It won't matter how much you payed a designed for a beautiful 404 page, a 404 page is frustrating for an user.

It might not be my case, but I want consistency. And this turned out to be a very easy fix in Pelican. I added the following to the pelicanconf.py file:

ARTICLE_URL = '{category}/{date:%Y}/{date:%m}/{date:%d}/{slug}/'
ARTICLE_SAVE_AS = '{category}/{date:%Y}/{date:%m}/{date:%d}/{slug}/index.html'
SLUGIFY_SOURCE = 'title'

These options, and many more, are explained in the Pelican settings.

Last step was getting the generated files to the Github page. In the version of Pelican I used, a Github pages support is already available, but before make github can work I had to create the git project and add the remote.

Other articles

Small Victories

Published: Mon 08 April 2013
By Sergio Botero

In blog.

tags: Optimizations Python ClusterPy RiSE

Lately I have been working in optimizing and parallelizing the ClusterPy library and at the moment, I can celebrate some 'Small Victories'.

So far I have been making small changes in places that have huge impact -functions that get called a lot and unnecessary function calling- and the timer now ...
read more
Opportunity in Parallelism

Published: Mon 11 March 2013
By Sergio Botero

In blog.

tags: Parallelism RiSE Python ClusterPY

Beginning this week I will be working for a period of 3 months with the people of the Research in Spatial Economics group (RiSE) at Universidad EAFIT. They developed a library of spatially constrained clustering algorithms named ClusterPy. I will be helping them in making some optimizations and basically, in ...
read more
Studying with MOOCs

Published: Tue 12 February 2013
By Sergio Botero

In blog.

tags: MOOCs UC Davis MIT CECI Labmanager Parallel Programming Computer Science

Looking forward to studying a Master's in Computer Science (I am still waiting to be admitted), I decided to start refreshing some concepts about parallel programming and theoretical computer science with two online courses from Udacity. This concept of Massive Open Online Course (MOOC) works great because you do ...
read more
New year, a note to self

Published: Tue 08 January 2013
By Sergio Botero

In blog.

After being away from the project for the holidays I started again to work on it last week and there is something I have to remember as a note to self: "When taking over a project, finish the new feature first and then improve the project."

And by improving the ...
read more
Labmanager for Learning Management Systems

Published: Sat 01 December 2012
By Sergio Botero

In blog.

tags: CECI Flask IMS Labmanager LMS MIT Moodle Python

Since October 2012 I have been visiting the CECI Lab at the MIT and contributing to a project to allow single sign-on between learning management systems (Moodle, Sakai, dotLRN, etc) and remote laboratories management systems. This will provide access to remote laboratories and other activities to students while they are ...
read more
Rosalind: Learning Bioinformatics

Published: Tue 20 November 2012
Updated: Mon 11 August 2014
By Sergio Botero

In blog.

tags: coding Bioinformatics python learning Rosalind

Rosalind.info is a great project to learn Bioinformatics by solving programming challenges. The system is similar to the one you would find in the Google Code Jam in which you are given a problem and a sample input and output. When you think you have solved the problem, you ...
read more
The Elements of Style

Published: Tue 17 January 2012
By Sergio Botero

In blog.

tags: The Elements of Style Writing William Strunk Jr. E.B. White

In this post I would like to thank Hock (nhocki on twitter) for a great gift. A gift that will teach me a lot about writing and that will, with no doubt, make a good writer out of me (hopefuly). The gift was a book from William Strunk Jr. and ...
read more
Hubzero Submit Installation

Published: Tue 01 November 2011
By Sergio Botero

In blog.

tags: batch system condor instalacion python submit

There are two machines/locations involved in the installation. The hub as HUB and the cluster controller as CLUSTER. Through this installation the $HOME directory on CLUSTER will refer to the home directory of the user that will be used to submit the jobs to the cluster (for most purposes ...
read more
For the next 2.5 months

Published: Mon 08 August 2011
By Sergio Botero

In blog.

tags: EAFIT HPC Hubzero Nanohub Purdue

The Purdue University (IN, US) is working closely with EAFIT University (ANT, CO) on a project to set up a supercomputer on EAFIT and be able to support some of the current scientific research in Colombia by providing the computational capabilities needed to process large sets of data and the ...
read more

Page 1 / 1

Other articles

social