1. Skip to content

1. Installation

Use this page to install ModSSC, verify the setup, and choose the right install path. After installation, continue with Quickstart.

1.1 Supported Python versions

ModSSC requires Python 3.11 or newer. [1]

1.2 Install from PyPI

Install the core library and CLI tools from PyPI.

Choose this if you only need the packaged library and CLI. The benchmark runner lives in bench/ and is not shipped to PyPI. [2]

This package name and entry points are declared in project metadata. [1][2]

python -m pip install modssc

1.3 Install from source

Use this for development, benchmarks, or when you need the latest main branch.

This path is recommended for benchmark runs because the runner and configs are in the repository. [2][3]

git clone https://github.com/ModSSC/ModSSC
cd ModSSC
python -m pip install -e "."

1.4 Verify installation

Use the CLI check if you plan to run modssc commands, and use the import check if you embed ModSSC inside your own Python scripts. See the CLI reference for command details. [4][6]

CLI version (from the main Typer app):

modssc --version

Python import smoke test:

python -c "import modssc; print(modssc.__version__)"

The CLI version flag is implemented in src/modssc/cli/app.py, and modssc.__version__ is defined in src/modssc/__about__.py. [4][5][6]

1.5 Common pitfalls

Warning

Some datasets and methods require optional extras (for example, hf, openml, graph, inductive-torch). [1][7]

Tip

For a full-feature installation (datasets, preprocess encoders, torch backends), use the full extra: python -m pip install "modssc[full]". [1]

Sources
  1. pyproject.toml
  2. README.md
  3. bench/README.md
  4. src/modssc/cli/app.py
  5. src/modssc/__about__.py
  6. src/modssc/__init__.py
  7. src/modssc/data_loader/errors.py