Window ops

Naive and fast implementations of common window operations.

This library is intended to be used as an alternative to pd.Series.rolling and pd.Series.expanding to gain a speedup by using numba optimized functions operating on numpy arrays. There are also online classes for more efficient updates of window statistics.

Install

pip install window-ops

How to use

Transformations

For a transformations n_samples -> n_samples you can use {[seasonal_](rolling|expanding)}_{(mean|max|min|std)} on an array.

Benchmarks

n_samples = 1_000  # array size
window_size = 8  # for rolling operations
season_length = 7  # for seasonal operations
execute_times = 1_000 # number of times each function will be executed

Average times in milliseconds.

display_dataframe(times, fmt='{:.2f}')

display_dataframe(speedups, fmt='{:.0f}')

Online

If you have an array for which you want to compute a window statistic and then keep updating it as more samples come in you can use the classes in the window_ops.online module. They all have a fit_transform method which take the array and return the transformations defined above but also have an update method that take a single value and return the new statistic.

Benchmarks

Average time in milliseconds it takes to transform the array and perform 100 updates.

display_dataframe(times.to_frame(), '{:.2f}')