Straightforward NNLS solvers in Python & Julia
It's called noodles because when I look at "NNLS" (Non-negative least squares) I think of the word noodles.
This repo implements the Lawson-Hanson algorithm for solving NNLS problems in both python and julia. The Python version requires numpy but is otherwise pure python, the Julia version requires no external packages.
These implementations are very rudimentary and straightforward. There are some optimizations to be made but not yet implemented (TBI, to be implemented).
The Bro-Jong (1999) speedups are not implemented here, as they only become useful when the number of iterations gets high (which is vaguely correlated with how many elements are in your x matrix, or correspondingly the number of templates to fit). They can also be useful when solving for many fits at once (TBI, maybe), so its not out of the question to include them here.
Brief timing tests are in the /nb folder.