feat: Add static or dynamic stepsize

[mhovd]

No description provided.

[Copilot]

Pull request overview

Adds configurable step-size strategies for the SDE (Euler–Maruyama) solver and exposes them through the public API and native DSL, enabling fixed, per-event, or adaptive stepping.

Changes:

• Introduces SdeStepSize (fixed / event-based / adaptive) and routes it into the Euler–Maruyama solver.
• Adds builder methods on equation::SDE and NativeSdeModel to configure step-size behavior.
• Re-exports SdeStepSize at the crate API level.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
src/simulator/equation/sde/mod.rs	Threads step-size configuration through the SDE simulation pipeline and adds SDE builder methods.
src/simulator/equation/sde/em.rs	Implements SdeStepSize and updates EM solver to support fixed/event/adaptive stepping.
src/lib.rs	Re-exports SdeStepSize from the crate prelude surface.
src/dsl/native.rs	Adds step-size configuration to the compiled/native SDE model pathway.

Comments suppressed due to low confidence (2)

src/simulator/equation/sde/em.rs:199

• solve_fixed can enter an infinite loop if dt is non-finite or <= 0.0 (because step becomes 0/negative/NaN and t never advances). This should be validated to avoid hangs when users configure SdeStepSize::Fixed (or EventSteps producing an invalid dt).

    fn solve_fixed(&mut self, t0: f64, tf: f64, dt: f64) -> (Vec<f64>, Vec<DVector<f64>>) {
        let mut t = t0;
        let mut times = vec![t0];
        let mut solution = vec![self.state.clone()];

        while t < tf {

src/simulator/equation/sde/em.rs:226

• solve_adaptive initializes dt to max_step without validating bounds and without clamping to the remaining interval before the first attempted step. If tf - t0 < max_step, the solver can step past tf and return a final time point > tf; if bounds are non-finite or <= 0, the loop can fail to make forward progress.

        let mut t = t0;
        let mut dt = max_step;
        let safety = 0.9;

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Bencher Report

Project	pharmsol
Branch	sde-stepsize
Testbed	mhovd-pgx

⚠️ WARNING: Truncated view!

The full continuous benchmarking report exceeds the maximum length allowed on this platform.

🐰 View full continuous benchmarking report in Bencher

[mhovd]

@Siel, not sure if we actually want this in main or not, you decide.
I would appreciate if you can check the implementation for the EM solver and the handling of the stepsize, I am in shallow water in that area.