Package {rjsf}

Title: R Interface to Jump-Switch-Flow Simulation
Version: 0.1.0
Description: An R interface to Jump-Switch-Flow (JSF), a hybrid simulation framework for compartmental reaction models that combines stochastic jump dynamics at low population sizes with deterministic flow dynamics at high population sizes. Simulations are performed using the 'jsf' backend for 'Python' through 'reticulate', with structured R outputs for inspection, summary, and plotting. The underlying methodology is described in Germano et al. (2026) <doi:10.1007/s00285-026-02409-y>.
License: MIT + file LICENSE
URL: https://olivervu25.github.io/rjsf/, https://github.com/olivervu25/rjsf
BugReports: https://github.com/olivervu25/rjsf/issues
SystemRequirements: Python, Quarto command line tool
Encoding: UTF-8
RoxygenNote: 7.3.2
Imports: reticulate, S7
Suggests: quarto, testthat (≥ 3.0.0)
VignetteBuilder: quarto
Config/testthat/edition: 3
NeedsCompilation: no
Packaged: 2026-10-03 16:20:23 UTC; olivervu25
Author: Thanh Long Vu [aut, cre, cph], Domenic Germano [aut], Michael Lydeamore [aut]
Maintainer: Thanh Long Vu <longvu2000.work@gmail.com>
Repository: CRAN
Date/Publication: 2026-10-15 16:10:08 UTC

Convert a JSF output data frame to a JSFResult object

Description

Convert a JSF output data frame to a JSFResult object

Usage

as_jsf_result(
  x,
  method = "operator-splitting",
  config = list(),
  species_names = NULL
)

Arguments

x

Data frame returned by jsf_simulate() or a model-specific wrapper. It must contain a time column.

method

JSF simulation method used.

config

List of simulation configuration values.

species_names

Optional character vector of species columns to include.

Value

A JSFResult object.

Examples

df <- data.frame(
  time = c(0, 0.1, 0.2),
  S = c(990, 988, 985),
  I = c(10, 12, 15),
  R = c(0, 0, 0)
)

result <- as_jsf_result(
  df,
  config = list(model = "SIR", dt = 0.1)
)

result
summary(result)

Check whether the Python jsf package is available

Description

This function checks whether the Python jsf module can be found by reticulate in the currently active Python environment.

Usage

jsf_available()

Value

A single logical value. TRUE if jsf is available, otherwise FALSE.

Examples

if (reticulate::py_available(initialize = FALSE)) {
  jsf_available()
}

Simulate a Lotka-Volterra predator-prey model using JSF

Description

This function runs a simple Lotka-Volterra predator-prey model using the Python jsf package through reticulate.

Usage

jsf_lotka_volterra(
  prey0 = 50,
  predator0 = 10,
  prey_birth = 2,
  predation = 0.05,
  predator_death = 1.5,
  t_max = 10,
  dt = 0.01,
  switching_threshold = 30
)

Arguments

prey0

Initial prey population.

predator0

Initial predator population.

prey_birth

Prey birth rate.

predation

Predation rate.

predator_death

Predator death rate.

t_max

Final simulation time.

dt

Time step for the continuous/ODE part of JSF.

switching_threshold

Switching threshold for both species.

Value

A data frame with columns time, prey, and predator.

Examples

if (interactive() && jsf_available()) {
    result <- jsf_lotka_volterra(
    prey0 = 50,
    predator0 = 10,
    t_max = 1
  )

  head(result)
}

Simulate a compartmental model using JSF

Description

This is a general R wrapper around the Python jsf simulator. It accepts an initial state, a Python rate function, reactant and product stoichiometry matrices, and JSF configuration options.

Usage

jsf_simulate(
  x0,
  rates,
  reactant,
  product,
  do_disc,
  t_max,
  dt,
  switching_threshold,
  tau_threshold = NULL,
  tau_epsilon = 0.03,
  critical_threshold = 10,
  tau_debug = FALSE,
  tau_debug_every = 20,
  tau_debug_max = 30,
  species_names = NULL,
  enforce_do = NULL,
  method = "operator-splitting",
  return_type = c("data.frame", "JSFResult")
)

Arguments

x0

Numeric vector of initial states.

rates

Python callable rate function, usually created with reticulate::py_eval().

reactant

Matrix where each row gives the reactants for one reaction.

product

Matrix where each row gives the products for one reaction.

do_disc

Integer vector indicating which species may be treated discretely/stochastically.

t_max

Final simulation time.

dt

Time step for the continuous/ODE part of JSF.

switching_threshold

Switching threshold. Can be a scalar or one value per species.

tau_threshold

Threshold controlling when species are handled by the tau-leaping component. May be NULL, a scalar, or one value per species. If NULL, defaults to switching_threshold.

tau_epsilon

Positive numeric scalar specifying the tau-leaping epsilon parameter passed to the Python JSF backend.

critical_threshold

Non-negative numeric scalar specifying the critical-reaction threshold passed to the Python JSF backend.

tau_debug

Logical. If TRUE, enables diagnostic output from the tau-leaping simulator.

tau_debug_every

Integer controlling how often tau-leaping diagnostic information is printed.

tau_debug_max

Integer giving the maximum number of tau-leaping diagnostic messages to print.

species_names

Optional character vector of species names.

enforce_do

Integer vector controlling enforced discrete behaviour. Defaults to zero for all species.

method

JSF simulation method. Supported values include "exact", "operator-splitting", and "tau-operator-splitting".

return_type

Output type. Either "data.frame" or "JSFResult".

Value

If return_type = "data.frame", a data frame with a time column and one column per species. If return_type = "JSFResult", a JSFResult object.

Examples

if (interactive() && jsf_available()) {
  rates_lv <- reticulate::py_eval(
    "lambda x, t: [2.0 * x[0], 1.5 * x[1], 0.05 * x[0] * x[1]]"
  )

  result <- jsf_simulate(
    x0 = c(prey = 50, predator = 10),
    rates = rates_lv,
    reactant = matrix(
      c(
        1, 0,
        0, 1,
        1, 1
      ),
      ncol = 2,
      byrow = TRUE
    ),
    product = matrix(
      c(
        2, 0,
        0, 0,
        0, 2
      ),
      ncol = 2,
      byrow = TRUE
    ),
    do_disc = c(1, 1),
    t_max = 1,
    dt = 0.01,
    switching_threshold = c(30, 30),
    species_names = c("prey", "predator"),
    return_type = "JSFResult"
  )

  result
  summary(result)
  plot(result)
}

Simulate an SIR epidemic model using JSF

Description

This function runs a simple SIR epidemic model using the Python jsf package through reticulate.

Usage

jsf_sir(
  N = 1000,
  I0 = 10,
  beta = 0.5,
  gamma = 0.1,
  t_max = 60,
  dt = 0.1,
  switching_threshold_I = 10
)

Arguments

N

Total population size.

I0

Initial infectious population.

beta

Infection rate.

gamma

Recovery rate.

t_max

Final simulation time.

dt

Time step for the continuous/ODE part of JSF.

switching_threshold_I

Switching threshold for the infectious compartment.

Value

A data frame with columns time, S, I, and R.

Examples

if (interactive() && jsf_available()) {
  result <- jsf_sir(
    N = 1000,
    I0 = 10,
    beta = 0.5,
    gamma = 0.1,
    t_max = 1,
    dt = 0.1,
    switching_threshold_I = 10
  )

  head(result)
}

Create a JSFResult object

Description

Create a JSFResult object

Usage

new_jsf_result(
  time,
  state,
  species_names = names(state),
  method = "operator-splitting",
  config = list()
)

Arguments

time

Numeric vector of simulation times.

state

Data frame of state trajectories.

species_names

Character vector of species names. Defaults to the column names of state.

method

JSF simulation method used.

config

List of simulation configuration values.

Value

A JSFResult object.

Examples

state <- data.frame(
  prey = c(50, 51, 52),
  predator = c(10, 9, 9)
)

result <- new_jsf_result(
  time = c(0, 0.1, 0.2),
  state = state,
  config = list(dt = 0.1)
)

result
summary(result)

mirror server hosted at Truenetwork, Russian Federation.