popcornell/pygf2

Polynomial arithmetic over GF2

pyGF2

Optimized Polynomial arithmetic over GF2[x].

This module supports polynomial addition gf2_add, multiplication gf2_mul division gf2_div and Extended Euclidean Algorithm gf2_xgcd.

Polynomials must be represented as ndarrays of uint8 including zero-valued coefficients. The rightmost array element is assumed to represent the leading coefficient.

---

For example:

>>>a =np.array([1, 0, 1, 1], dtype="uint8")

---

It is equivalent to the polynomial x^3+x^2+1. In other words the ordering is the same as the one used in MATLAB. Sympy and numpy.polys instead assume the leftmost array element to be the leading coefficient.

The use of such ordering allows to use numpy.fft to perform multiplication between polynomials.

---

In general a great effort has been put to speed-up all the operations making this module suitable for handling very large degree (tens of thousands) GF2[x] polynomials such as the ones required in coding theory.

Speed-wise it is orders of magnitude faster for polynomials of large degree when it is compared to both numpy.polys functions and Sympy.galoistools.

For Example:

• numpy.polymul vs gf2_mul for two random GF2[x] polynomial with 100000 elements: 5.63 s vs 105 ms

• numpy.polydiv encounters overflow problems when dealing with large polynomials. gf2_div solves that by using only modulo 2 operations.

• sympy.galoistools.gf_gcdex vs gf2_xgcd for two random GF2[x] polynomial with 11083 elements: 2 m 5 s vs 1.29 s

---

Requirements : NumPy (tested with NumPy 1.13.3 and Python 3.6.3)

---

pygf2's Issues

Installing pyGF2

Hi Samuele,

I want to run the following project: daniel-weisse/irreducible_polynom#1

Which has a dependency on pyGF2. How can install pyGF2?

Thank you in advance for any kind of help.

SyntaxWarning: "is not" with a literal. Did you mean "!="?

SyntaxWarning: "is not" with a literal. Did you mean "!="?
  if a.dtype is not "uint8":
SyntaxWarning: "is not" with a literal. Did you mean "!="?
  if b.dtype is not "uint8":
For solve the problem, we must change the following:

if a.dtype is not "uint8":
    a = a.astype("uint8")

if b.dtype is not "uint8":
    b = b.astype("uint8")

to:

if a.dtype != "uint8":
    a = a.astype("uint8")

if b.dtype != "uint8":
    b = b.astype("uint8")