Overview

MEmilio contains several python modules offering an easy-to-use interface to the efficiently implemented C++ models and to complement these by data preparation, code generation, machine learning, or plotting functionality. Please see the individual package documentation for more details on the functionality and usage.

MEmilio Python Interface

This package provides a python interface for parts of the C++ main library, with the goal of exposing fast mathematical-epidemiological models to a bigger user base.

To the python bindings

MEmilio EpiData Package

This package provides tools to download and structure important data such as infection or derived mobility data.

To the data package

Machine Learning Surrogate Models

This package contains machine learning-based surrogate models that were trained based on the MEmilio simulation outputs.

To the intro of surrogate models

Visualization

Generalized visualization functions for MEmilio specific plots.

To the visualization

Interface Generation

Easy to use tool for helping with the creation of python bindings or interfaces to (new) C++ models.

To the generation package

Installation

Each package provides a pyproject.toml that installs the package and its dependencies with pip. The dependencies of the individual packages are denoted in their documentation. The installation can be run with the following command (from the directory containing the pyproject.toml file)

python -m pip install .

This copies the package and the required dependencies to your site-packages.

For development of code use this command instead

python -m pip install -e .[dev]

This command allows you to work on the code without having to reinstall the package after a change. It also installs all additional dependencies required for development and maintenance.

Build files for skbuild

The simulaion and generation packages use skbuild to compile python bindings or parts of the C++ library. By default, the cmake build files are put into pycode/build/memilio-{package_name} to save on time during package development. If you get unexpected cmake errors, you can try and delete the respective build directory. If you do not want to store the build files at all, you can remove the build_dir entry from the section [tool.scikit-build] in the pyproject.toml. Then skbuild will use a temporary directory instead.

Testing

Each package provides a test suite under pycode/memilio-{package_name}/tests. To run the tests, simply use the following command inside the package folder after installation:

cd tests
python -m unittest

Coverage Report

Dependencies for coverage report:

  • coverage

To get the coverage report do in the package folder

python -m coverage run -m unittest
python -m coverage report
python -m coverage xml -o coverage_python.xml
python -m coverage html -d coverage_python

Coverage report for actual master:

Coverage Report

Inspection via pylint

The following packages have to be installed to run pylint:

  • pylint

  • pylint-json2html

Run pylint with the commands in the package folder

python ../run_pylint.py
pylint-json2html -f jsonextended -o build_pylint/pylint.html < build_pylint/pylint_extended.json

From the repository root you can also target a package explicitly, for example python pycode/run_pylint.py --package-dir memilio-plot.