ODE-Designer

ODE-Designer was developed in the context of a FAPEMIG research project by postgraduate students and undergraduate students in Computer Science at the Federal University of São João Del-Rei.

The main objective of ODE-Designer is to facilitate the creation, implementation and simulation of Ordinary Differential Equation (ODE) models. The software can be applied to research and teaching-learning of Computational Modeling.

Through a visually intuitive graphical user interface (GUI), ODE-Designer allows users to conceptualize and "draw" mathematical models, while the software automatically generates the requisite implementation, simulation, and result visualization code. A distinguished feature of the software is the model's visual representation. It is designed to be intuitive and easy to use.

ODE Designer

Installation

It is recommended to use the executables available in Releases on GitHub. This link can be used to always redirect to the latest version.

Linux

The Linux distribution uses AppImages, which requires a runtime (fuse2) installable on all Linux-based distributions. It is commonly included by default by some distributions, or is already installed by another program.

[!NOTE] Although AppImages can make binaries super portable in Linux, there's still a hard dependency on your system's GLIBC, which cannot be redistributed. As it stands, the minimum supported version is 2.28. You can check your system's version with this command:

ldd --version

If your system meets this requirement, but you still get errors related to GLIBC, please open an issue and we'll look into it.

🐧 For Debian/Ubuntu/Pop_OS!/ElementaryOS 
$ sudo apt install libfuse2
🐧 For ArchLinux 
$ sudo pacman -S fuse2

Windows

The Windows distribution consists of a ZIP that can be extracted and its contents executed.

Compilation

Using Docker

You can use the provided Dockerfile to build your own AppImage with no build tools installed directly on your system. Simply run:

$ docker build -t ode-designer-appimage-builder --output=. .

After this execution, the AppImage should be available in ./ode-designer.AppImage. This AppImage, as well as those found in Releases, include Python and the scipy and matplotlib dependencies for code generation, interactive simulation and PDF export.

Manually

To compile and run the software outside of the AppImage, the Rust toolchain (nightly version) is required, as well as Python >= 3.11, and the dependencies listed in requirements.txt. Fulfilling these requirements, simply compile like any Rust project, running

$ cargo run

Or

$ cargo build --release
$ ./target/release/ode-designer-rs

If anything goes wrong, try

$ git submodule update --init --recursive
$ git submodule update --recursive --remote

Then, recompile and run.

Functionalities

Node-based editor

Image from the node editor in the software, containing the nodes 'Var', 'Const', 'grow' and 'dVar/dt', which construct the ODE dVar/dt = Var*Const

Plotting

Plotting of the ODE dVar/dt = Var*Const from times 38 to 50, for the initial values Var = 1 and Const = 2

Python code export and pdf export

# imports of scipy and numpy omitted

def initial_values() -> np.ndarray:
    Var_0 = 1.0
    return np.array((Var_0,))


def constants() -> list:
    Const = 2.0
    return [Const]


def variable_names() -> list[str]:
    return ["Var"]


def system(t: np.float64, y: np.ndarray, *constants) -> np.ndarray:
    Var, = y
    Const, = constants

    dVar_dt = Var*Const

    return np.array([dVar_dt])

# Rest of the code used to simulate and plot to PDF omitted

Extensibility via Python code

Given the following Python code:

import math

@node
def sine(x):
    return math.sin(x)


@node(format="$1 ^ $2")
def power(x, y):
    return x ** y

By importing it in the Manage Extensions menu, you can use the defined nodes as if they were native, as in the image below.

Showcase of the custom nodes 'sin' and 'pow' in the node editor

The code can be used to simulate just like native nodes.

Plotting of the previous ODEs utilizing custom nodes

Runnig an example

  • To run an example, first go to the official repository ode-designer-rs;
  • Inside the repository, open the examples directory, where you'll find several .json files.

Examples directory Files inside the directory

  • Choose one example to run and download it. Afterward, open the software.

  • In the software, click on the File option in the top menu. When the dropdown appears, select Load.

  • This action will open a window that allows you to choose a file and execute it.

File option

Loading example file

  • Once this process completes, the model will be loaded in the software, ready for use.

Parameter estimation