Skip to content

Instantly share code, notes, and snippets.

@wmayner

wmayner/sliding_window.py

Created September 5, 2019 16:15
Show Gist options
  • Save wmayner/272a1a340403fead4703e4c78b5ec7f9 to your computer and use it in GitHub Desktop.
Save wmayner/272a1a340403fead4703e4c78b5ec7f9 to your computer and use it in GitHub Desktop.
Download ZIP
Sliding window function for NumPy arrays
Raw
sliding_window.py
# from
# https://gist.githubusercontent.com/nils-werner/9d321441006b112a4b116a8387c2280c/raw/489137f314c022ca08f0c5089319ded2eb8c5ae1/sliding_window.py
def sliding_window(size, data, step=1, padded=False, axis=-1, copy=True):
"""Calculate a sliding window over an array.
Note:
If the array cannot be windowed evenly, it is truncated at the end.
Args:
data (np.ndarray): The array to be windowed.
size (int): The window size.
step (int): The number of elements between windows. Defaults to 1.
axis (int): The axis to slide over. Defaults to the last axis.
copy (bool): Return a copy instead of a view.
Returns:
np.ndarray: A matrix where the last dimension corresponds to the
window.
Examples:
>>> a = np.array([1, 2, 3, 4, 5])
>>> sliding_window(3, a)
array([[1, 2, 3],
[2, 3, 4],
[3, 4, 5]])
>>> sliding_window(3, a, step=2)
array([[1, 2, 3],
[3, 4, 5]])
"""
if axis >= data.ndim:
raise ValueError("axis value out of range")
if step < 1:
raise ValueError("step size must be positive")
if size > data.shape[axis]:
raise ValueError("window size may not exceed size of selected axis")
shape = list(data.shape)
shape[axis] = np.floor(data.shape[axis] / step - size / step + 1).astype(int)
shape.append(size)
strides = list(data.strides)
strides[axis] *= step
strides.append(data.strides[axis])
strided = np.lib.stride_tricks.as_strided(data, shape=shape, strides=strides)
if copy:
return strided.copy()
else:
return strided
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment