Menu
SkillsMP has collected 22 skills from rust-dd/stochastic-rs. Open a skill to review its source and details.
Showing 22 of 22 collected skills.
Conventions for writing and maintaining stochastic-rs documentation pages under website/content/docs/. Nine section templates (process / distribution / pricer / calibrator / estimator / copula / AI surrogate / concept / tutorial), frontmatter schema, KaTeX…
Enforces scalability, integration, and compatibility requirements when creating any new module in stochastic-rs — covers stochastic, quant, stats, distributions, copulas, and ai
How to implement a model calibrator in stochastic-rs (BSM, Heston, SABR, SVI, …). Invoke when adding a new calibration routine that fits a model to market option prices or implied vols.
How to expose first- and second-order Greeks in stochastic-rs — an inherent greeks(s, k, r, q, tau, option_type) aggregator on a pricer, or the no-argument GreeksExt trait for the two Monte Carlo Malliavin estimators. Invoke when adding a pricer that needs…
Maintenance and extension rules for stochastic-rs Python (PyO3) bindings — invoke when adding/updating distributions, processes, pricers, calibrators, or estimators that need Python exposure
How to add a bivariate copula to stochastic-rs-copulas. Invoke when implementing Clayton, Frank, Gumbel, Joe, Plackett, FGM-style families, or any 2-d Archimedean / extreme-value copula.
How to add a univariate distribution to stochastic-rs-distributions. Covers SimdXxx struct, sampling pattern (transformation / ziggurat / rejection / inversion), DistributionExt closed-form moments/pdf/cdf/cf, KS-test, and the py_distribution! macro.
Conventions for adding / propagating Cargo features across the stochastic-rs workspace. Invoke when adding a new optional dependency, gating a sub-module, or debugging "feature X enabled in crate A but not B" build errors.
Step-by-step recipe for exposing a new Rust type to Python via stochastic-rs-py. Quickstart companion to the comprehensive `python-bindings` SKILL — invoke for the "what files do I touch and in what order" view.
How to add a new diffusion / SDE process to stochastic-rs-stochastic. Invoke when implementing GBM-like, OU-like, Vasicek-like, CIR-like, Heston-like models that satisfy `dX_t = drift dt + diffusion dW_t`.
Development rules for stochastic-rs — enforces project conventions when writing new modules, adding dependencies, or implementing algorithms
Conventions for writing integration tests in stochastic-rs. Pinned-seed mandate, golden-numerics paper-reference pattern, no-debug-test rule, feature-gating discipline. Invoke when adding tests under tests/ or when a test is flaky.
Use when writing or editing Rust `///` or `//!` doc comments — never tag sections with `## v2.3.0 design choice — …` or `## v2.4 deferred — …`. Version history goes in `MIGRATION.md`, git log, or `docs/V*_UPDATE.md`, never in source doc comments.
Step-by-step release workflow for stochastic-rs (rc → stable → patch). Invoke when bumping versions, cutting a tag, publishing to crates.io, or shipping wheels to PyPI.
How to add a Monte Carlo variance-reduction technique to stochastic-rs-stochastic. Covers antithetic, control-variate, stratified, importance, quasi-MC (Halton/Sobol), and MLMC. Returns McEstimate<T> with 95% CI.
How to add a statistical estimator to stochastic-rs-stats. Covers ArrayView1<T> input shape, *Result struct conventions, parametric vs bootstrap p-values, openblas gating, paper-citation requirements, and reference-comparison tests.
How to add a Hurst-parameterised (rough / fractional) process to stochastic-rs-stochastic. Invoke when wrapping fBm / fGn / Volterra-kernel processes — fOU, rBergomi, rough Heston, fBates, fractional CIR, etc.
How to add or extend a GPU / accelerated sampling backend (CUDA, Metal, wgpu, Accelerate) in stochastic-rs-stochastic. Invoke when porting fGN sampling to a new device, or when a backend-generic process needs to reach one.
Conventions for writing criterion benchmarks in stochastic-rs. Group naming, parameter sweep, [[bench]] required-features gating, no-println / no-dead-helper rules. Invoke when adding a new bench or refactoring an existing one.
How to add a jump-diffusion / Lévy / compound-Poisson process to stochastic-rs-stochastic. Invoke for Merton-jump, Kou-jump, Bates-style models, or for layering jumps onto an existing diffusion (GBM → MJD, Heston → Bates).
How to add a neural-network volatility surrogate to stochastic-rs-ai. Covers StochVolModelSpec, StochVolNn, the BoundedScaler / StandardScaler split, gzip-npy training-set loading, the train_save_load roundtrip test, and predict_implied_vol_surface…
Use when writing or editing Rust where you'd otherwise add a `let x: T = …` binding-type annotation — prefer turbofish on the call site (`Foo::<T>::method()`, `.collect::<Vec<_>>()`) so the type travels with the expression instead of pinning the binding.