# Commol Documentation

> A high-performance compartment modelling library for mathematical modeling

Commol is a Python library, with a Rust engine, to build, simulate and calibrate compartment models. Models are built with ModelBuilder, simulated with Simulation and fitted with Calibrator. Examples use abstract names (bins A, B, C; parameters k1, k2).


## Getting started

- [Home](https://munqu.github.io/commol/latest/index.md): Overview and where to go
- [Installation](https://munqu.github.io/commol/latest/getting-started/installation/index.md)
- [Quick Start](https://munqu.github.io/commol/latest/getting-started/quickstart/index.md)

## User guide

- [Core Concepts](https://munqu.github.io/commol/latest/guide/core-concepts/index.md): Compartments, parameters, transitions, accumulators, the process and the structure, the four engines and what a rate means in each, stratifications, units of the axis
- [Model Builder](https://munqu.github.io/commol/latest/guide/building-models/index.md): Every ModelBuilder method and argument, series parameters, the Aggregate structure and its neighbors, build, save and load
- [Transitions](https://munqu.github.io/commol/latest/guide/transitions/index.md): Absolute and proportional rates, several sources, $compartment, category transitions, per_compartment, stratified rates, impulses, amounts below zero, what build checks
- [Mathematical Expressions](https://munqu.github.io/commol/latest/guide/mathematical-expressions/index.md): Operators, variables, subgroup totals, functions and security limits of rate formulas
- [Axis Patterns](https://munqu.github.io/commol/latest/guide/axis-patterns/index.md): AxisPattern pulses, windows, ramps, sinusoids, combinations and schedules per subgroup
- [Units](https://munqu.github.io/commol/latest/guide/units-and-equations/index.md): Unit checking of rates, the step unit, positions and numbers in a rate, impulses, and dates on the axis
- [Describing a Model](https://munqu.github.io/commol/latest/guide/describing-models/index.md): Model.describe, its sections and blocks, text, Markdown, LaTeX, JSON and DOT renderings, the ODD layout
- [Running Simulations](https://munqu.github.io/commol/latest/guide/simulations/index.md): Simulation.run and run_at, results by output name, seeds and replicates, accumulators, totals by bin and category, windowed totals, differential equations, figures and plot_compared
- [Stochastic Simulation](https://munqu.github.io/commol/latest/guide/stochastic/index.md): The StochasticEvents and StochasticSteps engines, how rates become events, whole units, seeds and replicates, methods, calibration with replicates
- [Calibration Basics](https://munqu.github.io/commol/latest/guide/calibration/index.md): CalibrationProblem, observations, unknowns, losses, stochastic engines, results and writing results back
- [Initial Conditions](https://munqu.github.io/commol/latest/guide/calibration-initial-conditions/index.md): Calibrating bins, categories and compartments, the remainder and the order of application
- [Observations](https://munqu.github.io/commol/latest/guide/calibration-observations/index.md): The fields of an observation, several outputs, sums of outputs, weights, scale factors and windowed observations
- [Optimizers](https://munqu.github.io/commol/latest/guide/calibration-optimizers/index.md): Nelder-Mead and particle swarm settings, several starts, reproducibility, troubleshooting
- [Constraints](https://munqu.github.io/commol/latest/guide/calibration-constraints/index.md): Constraints on unknowns and outputs, and how violations count
- [Uncertainty](https://munqu.github.io/commol/latest/guide/uncertainty/index.md): The three questions, bands and prediction intervals, which method answers which question, the parameter sample, the predictive, scenarios, scores and the entry point
- [Observation Models](https://munqu.github.io/commol/latest/guide/uncertainty/observation-models/index.md): Observation models of each series, noise parameters, maximum likelihood, the Bayesian posterior and its evidence, prediction intervals with noise, the log score and the parametric bootstrap
- [Identifiability](https://munqu.github.io/commol/latest/guide/uncertainty/identifiability/index.md): Structural against practical identifiability, the structural rank, collinearity, a sample, starts, combinations and the report
- [Resampling](https://munqu.github.io/commol/latest/guide/uncertainty/resampling/index.md): The Bayesian, residual and wild block bootstraps, the block jackknife, block lengths and prediction intervals from residuals
- [Sequential Monte Carlo](https://munqu.github.io/commol/latest/guide/uncertainty/smc/index.md): Priors, the sequential Monte Carlo sampler, the Gibbs posterior and its scale eta, ABC and its distances, and the diagnostics
- [Local Approximations](https://munqu.github.io/commol/latest/guide/uncertainty/local/index.md): The curvature and the unbounded scale, the Laplace approximation and the delta method, the sandwich, the Laplace mixture and PSIS
- [Profile Likelihood](https://munqu.github.io/commol/latest/guide/uncertainty/profiles/index.md): Profile likelihood of parameters and predictions, how to read a profile, the chi-square, Beale and bootstrap thresholds, the adaptive step, verdicts, the report and plots
- [Conformal](https://munqu.github.io/commol/latest/guide/uncertainty/conformal/index.md): Conformal prediction intervals, the guarantee, scores and pooling, split, rolling origins, jackknife+ and CV+, EnbPI, adaptive conformal, plots and reports
- [Sensitivity Analysis](https://munqu.github.io/commol/latest/guide/uncertainty/sensitivity/index.md): Forward questions with no data, input spaces, quantities of interest, propagation, PRCC, Morris screening, Sobol indices, polynomial chaos and their plots
- [Reports](https://munqu.github.io/commol/latest/guide/reports/index.md): The applied parameters of a model, and the reports of a run, of replicates, of a predictive and of an identifiability analysis
- [Threads and Parallelism](https://munqu.github.io/commol/latest/guide/parallelism/index.md): The threads and levels of the parallel work, commol.set_threads and commol.threads, RAYON_NUM_THREADS
- [Examples](https://munqu.github.io/commol/latest/guide/examples/index.md): Complete models, from rates and stratifications to other engines, a fit, its uncertainty and scenarios

## API reference

- [API Overview](https://munqu.github.io/commol/latest/api/index.md)
- [AxisPattern](https://munqu.github.io/commol/latest/api/axis-patterns/index.md)
- [AxisUnits](https://munqu.github.io/commol/latest/api/axis-units/index.md)
- [Calibration](https://munqu.github.io/commol/latest/api/calibrator/index.md)
- [Identifiability](https://munqu.github.io/commol/latest/api/identifiability/index.md)
- [ModelBuilder](https://munqu.github.io/commol/latest/api/model-builder/index.md)
- [Model](https://munqu.github.io/commol/latest/api/model/index.md)
- [Package](https://munqu.github.io/commol/latest/api/package/index.md)
- [Parallelism](https://munqu.github.io/commol/latest/api/parallelism/index.md)
- [Plotting](https://munqu.github.io/commol/latest/api/plotting/index.md)
- [Reports](https://munqu.github.io/commol/latest/api/reports/index.md)
- [Sensitivity](https://munqu.github.io/commol/latest/api/sensitivity/index.md)
- [Simulation](https://munqu.github.io/commol/latest/api/simulation/index.md)
- [Uncertainty](https://munqu.github.io/commol/latest/api/uncertainty/index.md)
- [Model Structure](https://munqu.github.io/commol/latest/advanced/model-structure/index.md): The data model of a Model

## Changelog

- [Changelog](https://munqu.github.io/commol/latest/about/changelog/index.md): Every change, breaking ones marked