higgs-bosoff / godot-linalg

Linear Algebra library in GDScript for Godot Engine

godot-linalg

Linear Algebra library in GDScript for Godot Engine

All methods are optimised for maximum speed. They take arrays and assume the right dimension for them. If the inputs aren't right they'll crash. Third input is used for the answer, preallocated. There are no conditional branchings. Just use the method appropriate to the situation. The names are coded to reflect that. s = scalar, v = vector and m = matrix. So for example

dot_vm(v, M)

is a dot product between vector and matrix (in that order). Wherever the in_place argument is provided, it is possible to perform the operation on the object itself instead of instantiating a new one (this too optimises performance). So for example

transpose(M, true)

will turn M into its own transpose by reference, whereas

 MT = transpose(M)

will leave M unaltered.

Most method names are self-explanatory. The less intuitive are:

• householder(v): computes the Householder matrix of a vector;
• qr(M): computes the QR decomposition of a square matrix;
• eigs_powerit: computes the eigenvalues and eigenvectors of a square symmetric matrix using the power iteration algorithm combined with a Hotelling deflation (check out this book for more details).

godot-linalg's Issues

Enhancement: why not as a C++ module using Pool/PackedFloatVector?

Transpose function not working for m x n Matrix when m != n

Hey, great library but your transpose logic only works for square matrices, you could try something like this instead:

Transpose

static func transpose(M: Array)->Array:
var ans = []
var m = len(M)
var n = len(M[0])
ans.resize(n)

for i in range(n):
	ans[i] = []
	for j in range(m):
		ans[i].append(M[j][i])

return ans

Cheers