I'm happy to announce version 0.6 of the crate curve-sampling. This crate determines the best points at which to evaluate a function f: ℝ → ℝ (resp. f: ℝ → ℝ²) to accurately represent its graph (resp. its image).

As an example of what this library can do, here is the graph of the function x ↦ x sin(1/x) drawn with 227 evaluations.

For comparison, here is the graph of the same function with a non-adaptive algorithm (using Gnuplot with the same number of evaluations).

Improvement w.r.t. the previous version include attaching arbitrary data to each evaluation (allowing to track the parameters it comes from), iterators over that data, improved documentation, a much more versatile Sampling::fun, support for complex numbers,...

Hope you find it useful!

2 posts - 2 participants

Read full topic