onesandzeros's posterous

I wish I had started with this book when I first got interested in machine learning: it's fun and gives just enough background to try things out without information overloading. There are always references to further explore things out in case you wish - and I did so, with text information retrieval, which is a good sign on how well the book did on introducing the topics at hand. Great chapters are the ones on analysing email (ch 3), blogs (ch 8. Natural language processing and beyond) and Twitter (ch 5). Chapter 7 on Google Buzz is unfortunately outdated - it would be great if a new chapter on Google Plus would be released for new buyers.

Each chapter is light on detailed theoretical explanations but the code samples are just fun to run. And they're based on python, which has a soft spot in my heart :) The author's style is very easy, which makes the ride that much better.

Quick: Mining the social web is a great way to be introduced to the variety of ways data so freely available on the web can be analysed and visualized - it's sort of mind-opening what can already be achieved with so little effort. Good book for beginners.

Here's a way you can reset the database before running your phpunit tests.

In app/phpunit.xml change the bootstrap file we're going to use:

   bootstrap                   = "my_phpunit_bootstrap.php" >   

In my_phpunit_bootstrap.php we import the old bootstrap and add our code to recreate the database and load the fixtures (you could also do this inside your tests):

I also set the test database as pdo_sqlite for faster tests.

Here are some other links I found useful before settling into this solution:

Fully isolated tests in symfony 2

What is the best way to create relationships with functional tests

Using doctrine entity manager in symfony2 unit tests

no better tool to find code snippets in a directory. pretty pretty

Just built my first vim plugin: line-log.vim. It sends the current line in the buffer to a log file, flagging if it's the start or end of the task at hand. The log format is in csv format for easy report generation and borrows heavily from activity-log vim plugin.

The main purpose of this plugin is to make time reporting easy for the forgetful programmer, without additional tools it provides a static log for you, but the format is such that you can easily import or analyze using whatever tool you feel comfortable with.