Vectorized2D

This is a user-friendly wrapper to numpy arrays, for dealing with numerical problems in a vectorized fashion - in the 2D world.

Provided objects include:

1. Array2D - a user-friendly interface to numpy arrays of shape=Nx2.
2. Vector2D - a user-friendly wrapper for arrays of 2D vectors that represent physical quantities.
3. Point2D - a user-friendly wrapper to arrays of 2D points that represent spatial locations in a cartesian coordinate system.
4. Coordinate - a user-friendly wrapper for arrays of 2D points that represent 2D spatial (geographical) coordinates (longitude and latitude) in radians.

Installation

The easiest way to install vectorized2d and get updates is via the Python Package Index (PyPI):

$ pip install vectorized2d

Performance

Vectorized2d uses Numba to gain enhanced performance compared to vanilla numpy.

For example, (per-row) norm calculation:

  >>> import numpy as np
  >>> from vectorized2d import Array2D

  >>> a_np = np.random.random(size=(1000, 2))
  >>> a_2d = a_np.view(Array2D)

  >>> %timeit np.linalg.norm(a_np, axis=1)
  23.1 µs ± 1.25 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)

  >>> %timeit np.sqrt(np.einsum('ij,ij->i', a_np, a_np))
  10.5 µs ± 40.2 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

  >>> %timeit a_2d.norm
  2.63 µs ± 67.9 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)