drisk 0.1.0

Composable tools for quick Monte Carlo modelling and distribution elicitation.

drisk

Composable tools for quick Monte Carlo modelling, with an emphasis on distribution elicitation and model composition.

drisk provides a compact, notebook-friendly API for probability distributions, elicitation workflows, correlated sampling, decision trees, sensitivity analysis, and composable Monte Carlo models.

drisk is currently in alpha. APIs may change before a stable release.

Links

• Documentation: https://nick-dorsch.github.io/drisk/
• Repository: https://github.com/nick-dorsch/drisk
• Issues: https://github.com/nick-dorsch/drisk/issues

Installation

Install from PyPI:

pip install drisk

Or with uv:

uv add drisk

drisk requires Python 3.12 or newer.

Quick start

import drisk as dr

# Elicit a distribution from intuitive inputs.
dist = dr.LogitNormal.elicit(low=0.1, high=0.25)

# Draw Monte Carlo samples.
samples = dist.sample(1_000)

Features

• Continuous distributions such as Normal, LogNormal, LogitNormal, Beta, PERT, Gamma, and Exponential
• Discrete distributions such as Bernoulli, Binomial, Poisson, Geometric, and NegativeBinomial
• Distribution elicitation, fitting, sampling, plotting, and serialization-friendly parameters
• Mixture distributions with UvMixture
• Correlation matrix helpers and Gaussian / Student-t copulas
• Composable Monte Carlo models via MCModel, MCOperation, and where
• Decision tree modelling with chance, decision, and outcome nodes
• One-at-a-time sensitivity analysis helpers

Documentation

The documentation site includes examples and API reference pages:

https://nick-dorsch.github.io/drisk/

Developer setup

This repository uses:

• mise to install local development tools and run common project tasks
• uv for Python dependency management and command execution

1. Install tools with mise

Install mise if you do not already have it, then install the tools pinned in .mise.toml:

mise install

This installs the project Python version, uv, ruff, and other helper tools for the repository.

2. Install Python dependencies with uv

Sync the project environment from pyproject.toml and uv.lock:

uv sync --all-groups

Use uv run ... for commands that should execute inside the managed environment:

uv run pytest
uv run ruff check .
uv run python -c "import drisk as dr; print(dr.Normal.elicit(0, 1))"

When adding dependencies, use uv add so pyproject.toml and uv.lock stay in sync:

uv add numpy
uv add --dev pytest
uv add --group docs jupyter

3. Run project tasks with mise

List available tasks:

mise tasks

Common tasks:

mise run test      # run the test suite
mise run lint      # run ruff checks
mise run format    # format the codebase
mise run check     # run lint and tests
mise run docs      # preview the Quarto documentation site

Project layout

src/drisk/
  copulas/                  # copula interfaces and implementations
  correlations/             # correlation matrix helpers
  decision/                 # decision tree nodes, branches, and trees
  distributions/            # distribution interfaces and implementations
  models/                   # composable Monte Carlo model helpers
  sensitivity/              # sensitivity analysis utilities
  arithmetic.py             # arithmetic/composition helpers
  random.py                 # seed/RNG helpers
  summary.py                # summary helpers

tests/                      # pytest suite
docs/                       # Quarto documentation site

License

drisk is distributed under the MIT License. See LICENSE for details.