An implementation of a steepest descent algorithm for fitting exoplanet transit data to Mandel and Agol light curves (http://arxiv.org/abs/astro-ph/0210099).
The only thing here that I wrote is 'ad-exofit.py', which implements a version of the algorithm mentioned above. I wrote this mostly out of curiosity; it turns out that this method is a lot slower than the Markov Chain Monte Carlo that's all the rage these days. It also gets stuck in local minima. The solution to this is straight forward -- if there has been no change in the fit for some number of steps, increase the dx until there is change -- I just haven't gotten around to coding this.
There's a neat little feature that lets you watch the fit happen in real time. Definition of exhilirating.
The rest of the code (i.e. most of it) is by Laura Kreidberg (http://astro.uchicago.edu/~kreidberg/) and included here with permission.