A Python framework for formulating, simulating, and estimating information dynamics on networks.

This repository implements two complementary approaches:

• an equation-based approach, in which information-theoretic quantities evolve through explicit dynamical equations;
• a model-based approach, in which those quantities are estimated from data generated by underlying dynamical models.

Information_dynamics/
├── Core/            # Core classes for networks, information-dynamic equations, estimators, and model basics
├── on_Equations/    # Equation-based examples and demonstrations
├── on_Model/        # Model-based simulations and estimation workflows
├── Utils/           # Utility scripts for plotting and postprocessing
├── Data/            # Saved results, figures, and related outputs
└── README.md

This repository includes data and scripts used to reproduce results reported in the manuscript.

• Model 004: Gene regulatory network (GRN) simulation
• Model 005: Three-node oscillatory system

python -m on_Model.004_ABN_for_GRN.main

python -m Utils.plot_results \
    --dir on_Model/004_ABN_for_GRN/Temporal_Results/ \
    --link B4 B5 --paper

python -m Utils.plot_results \
    --dir on_Model/004_ABN_for_GRN/Temporal_Results/ \
    --node B1 --paper

• Results are saved in:

  • on_Model/.../Temporal_Results/
  • Data/ (if exported)

• Figures used in the manuscript are generated directly from these outputs.

• All simulations are deterministic given fixed initial conditions.
• Random seeds (if used) should be set explicitly in model scripts.
• The estimator configuration (k, metric, etc.) is defined in Core/Estimators.

The core library of the repository.

It contains:

• network data structures
• information-dynamic update rules
• estimator classes
• model base classes
• supporting utilities for simulation and analysis

Examples in which the information variables themselves evolve according to dynamical equations.

This directory is intended for studying:

• intrinsic information dynamics
• the effect of control parameters
• stationary solutions
• blocked-flow conditions
• network-topology dependence

Examples in which raw system trajectories are first generated from explicit models, and the information variables are then estimated from those data.

This directory is intended for studying:

• model-based information dynamics
• estimator behavior
• comparisons between simulated mechanisms and information-theoretic observables

Utility scripts for plotting and postprocessing results.

Typical uses include:

• plotting temporal results
• generating paper-style figures
• comparing outputs across runs

Saved outputs from simulations and analyses.

This may include:

• temporal results
• processed data
• figure source files
• exported plots

From the repository root:

python -m on_Equations.001_A_Single_Cycle.main

or

python -m on_Equations.002_Cycle_and_Source.main

From the repository root:

python -m on_Model.004_ABN_for_GRN.main

or

python -m on_Model.005_Three_Nodes_GRN.main

Example commands from the repository root:

python -m Utils.plot_results --dir on_Model/004_ABN_for_GRN/Temporal_Results/ --link B4 B5 --paper

python -m Utils.plot_results --dir on_Model/004_ABN_for_GRN/Temporal_Results/ --node B1 --paper

Adjust the directory and arguments depending on the model and output you want to visualize.

This repository is intended for developing and testing a unified framework for information dynamics on networks.

Its goals include:

• deriving and studying dynamical equations for information variables
• comparing equation-based and model-based approaches
• examining the role of topology, control parameters, and constraints
• exploring applications to biological and other complex systems

Install dependencies with:

pip install -r requirements.txt

• Run all commands from the repository root.
• Some examples save outputs into subdirectories under Data/ or model-specific result folders.
• The repository is organized so that the Core/ directory provides the shared framework used by both on_Equations/ and on_Model/.

Please check the repository files for license and data-use details if they are provided separately.