These are Python bindings for the original Fortran implementation of Box Least Squares (BLS) algorithm from Kovács et al. (2002).

Clone the source code from the GitHub repository and install using the standard python tools:

git clone https://github.com/dfm/python-bls.git
cd python-bls
python setup.py install

For testing in development, you can use

python setup.py build_ext --inplace

to build the bindings directly in the bls directory.

These Python bindings were developed—building directly on the code released by Kovács—at the SAMSI workshop Modern Statistical and Computational Methods for Analysis of Kepler by

• Ruth Angus (Oxford)
• Dan Foreman-Mackey (NYU)

The Python bindings are licensed under the MIT License.

TODO: describe Pythonic binding.

You can get direct access to the Fortran subroutine provided by Kovács through the eebls() function:

import bls
results = bls.eebls(time, flux, u, v, nf, fmin, df, nb, qmi, qma)

where

• time is an N-dimensional array of timestamps for the light curve,
• flux is the N-dimensional light curve array,
• u and v are N-dimensional empty work arrays,
• nf is the number of frequency bins to test,
• fmin is the minimum frequency to test,
• df is the frequency grid spacing,
• nb is the number of bins to use in the folded light curve,
• qmi is the minimum transit duration to test, and
• qma is the maximum transit duration to test.

The returned values are

power, best_period, best_power, depth, q, in1, in2 = results

where

• power is the nf-dimensional power spectrum array at frequencies f = fmin + arange(nf) * df,
• best_period is the best-fit period in the same units as time,
• best_power is the power at best_period,
• depth is the depth of the transit at best_period,
• q is the fractional transit duration,
• in1 is the bin index at the start of transit, and
• in2 is the bin index at the end of transit.