This version of EightBitTransit is an MIT-licensed python code redesigned to make use of GPU parallelization with Numba:
- Can calculate the model of any pixelated image transiting a star
- Can inject the model into a light curve with a given depth (negative allowed, i.e. a spike) and duration
- Can invert a light curve to recover the "shadow image" that produced it. (from original code unmodified).
To install EightBitTransit, run:
pip install eightbittransit
- Numpy
- Scipy
- Numba
- imageio
- itertools
- matplotlib
Optionally (for the example notebook):
- jupyterlab
See ./examples/examples-gpu.ipynb for examples of both the forward and inverse shadow imaging problem, including for dip 5 of Boyajian's Star. This code reproduces figure 12 of Sandford & Kipping 2018 (https://arxiv.org/abs/1812.01618).
See /examples /eightbit-injection.ipynb for examples on how to inject a signal into an existing lighr curve.
Repeated TransitingImage() calls can cause python to run out of memory in certain cases (thanks to textadactyl for pointing this out!)--when you are done with a TransitingImage object, deallocate it explicitly, i.e.:
ti = TransitingImage(...)
some operations on ti, etc.
ti = None