Toronto Python Meetup Hack Night Project

Hasan's project for the Hack Night was a log-file visualizer to be used with real-time-ish log parsing such that you could "waggle a mouse in this window and see the performance stats in this one" (he does video driver testing by day).  We decided to step that down to parsing a sample data-file (in this case an apache log file) and rendering a time:downloaded-bytes plot.

We started off with a script that just parsed the whole test data-set into RAM and rendered it with OpenGLContext.  After that we made it capable of doing data-window rendering, so that you just see the last 600 data-points rendered into a sliding window.  I fixed a few bugs in that on the subway, and then added some scale labels after I got home.

I've uploaded the code to github, should anyone wish to play with it. The code was quite capable of rendering 47,000 data-points at >100fps, the current 600-at-a-time thing is just so that we could use a small data-set for testing iterative loading.  Lots of code could be optimized, but honestly we just weren't hitting any limits to make it worthwhile doing so.


Comments are closed.


Pingbacks are closed.