julialg

Julia-style arrays in Python

Python library for mimicking Julia LinearAlgebra array indexing style and display formatting.

Test Result:

Motivation and Summary

Julia's LinearAlgebra package has some nice features, some of which are easier to emulate in Python than others. When comparing the results of Python code vs Julia code, it is annoying to have to mentally switch between the 0-indexed nature of Python and the 1-indexed nature of Julia. Also, the Julia arrays have a prettier array.

This package wraps numpy n-dimensional arrays in a new class, JulArray, and allows for 1-indexed slicing (more mathematically intuitive) instead of 0-indexed slicing (computer science convention). Also improves the prettiness of the representation of the array in a manner similar to Julia's LinearAlgebra package. To be clear, this is a Python package, meant to bring some of the elegance of Julia's interface for tensors to the Python setting

Indexing

The JulArray class wraps a numpy.ndarray but overrides the getitem syntax to allow for 1-indexed style instead of the default 0-indexed style. For example:

>>> import numpy, julialg

# Create a numpy array
>>> a = numpy.arange(1, 11).reshape((2, 5))

# Create a JulArray from the numpy array
>>> j = julialg.JulArray(a)

# Index the numpy array using 0-indexed syntax
>>> a[0, 0:2]
array([1, 2])

# Index the JulArray using 1-indexed syntax
>>> j[1, 1:3].array
array([1, 2])

Notice in the above sample that the JulArray is able to convert both ints and slices from 1-indexed notation to 0-indexed notation to produce the same underlying numpy array.

Display

The JulArray overrides the default representation of the numpy array to be more cleanly formatted (like Julia arrays).

>>> julialg.JulArray(numpy.arange(1.0, 11.0).reshape((2, 5)))
2x5 Array{float64,2}
  1.0000     6.0000
  2.0000     7.0000
  3.0000     8.0000
  4.0000     9.0000
  5.0000    10.0000