A co-worker of mine set up a Democratic Jukebox public test server today.
Have fun trying it out:
Last week I’ve published my latest fun projekt, the democratic music player: Democratic Jukebox.
Ever wanted to listen to music with a larger group of people e.g. in your office? Who decides what to play? Make your music player democratic and give everyone the chance to promote their favourite song.
The jukebox provides a web interface to search your music library and vote for songs to be played. The more votes a song gets, the sooner you will listen to it.
Jukebox is written in Python using Django and some other packages.
The project source code and installation details are published on Github.
Fork it, play with it, add features, post issues, give feedback – have fun!
I found a nice little piece of software again today while cleaning up my development stuff…
Sphinx is an open source full text search server, designed from the ground up with performance, relevance (aka search quality), and integration simplicity in mind. It’s written in C++ and works on Linux (RedHat, Ubuntu, etc), Windows, MacOS, Solaris, FreeBSD, and a few other systems.
Working with sphinx is pretty easy but it’s query log is – like any other access log – not meant to be read or analyzed by humans. Some time ago I wrote a little tool to analyze these log files and extract information about slow queries – the bad ones making your users sad.
Because sphinx itself is open source and free to use, I decided today that I will also share the tool with anyone out there. Have a look at GitHub and try it out.
Get it here: Sphinx query log analyzer
face.com is a face recognition service.
Unfortunately they only return face positions and not crop information.
Getting the best thumbnails from picture showing people isn’t even trivial when you already know where the heads are. Big pictures with small faces require different margins than small pictures with huge faces.
Various people on one picture are even another challenge.
At work, we’ve implemented an algorithm to find the best thumbnail crop positions using the face.com face detection service. Feel free to use it on your own projects.
Our current project at work uses face.com for face detection on user uploaded images. It works really smooth and the detection rate is fantastic. However face.com limits API requests to 5000 queries per hour. I’ve written a tiny nagios plugin to monitor the currently used requests and alert the team when we will exceed the limit. Feel free to use it, fork it and do whatever you want with it: