PySNT

Python wrapper for SNT, the framework for quantification of neuronal anatomy

Technical Features

• Complete Java Bridge: Full programmatic access to SNT's neuroanatomy toolkit from Python (via JPype and ScyJava), enabling integration of advanced tracing, analysis, and visualization capabilities into Python workflows

• Consistent API: 1:1 correspondence with SNT's Java API while maintaining Python conventions, allowing developers familiar with either language to work efficiently

• Full Type Annotation: Type hints for all methods and classes, enabling static type checking with mypy and improved code reliability

• IDE Integration: Autocomplete, inline documentation, and parameter hints in VS Code, PyCharm, Spyder, and Jupyter notebooks through type stubs and docstrings

• Performant: Lazy class loading minimizes initialization overhead, loading only required SNT components on demand

• Native Python Converters: Automatic conversion between SNT objects and Python-native types:

  SNT Java Object	PySNT Python Object
  SNTChart	matplotlib Figure
  SNTGraph	NetworkX graph
  SNTTable	xarray Dataset; pandas DataFrame
  ImagePlus	xarray; numpy (via ScyJava)
  Collections	lists/dicts, etc. (via ScyJava)

Limitations

• Pre-release: PySNT 5.0 is under active development and targets SNT 5.0.x: Expect bugs and API changes until the first stable release

• Early Stage Software: Limited community validation. Please report anything unexpected

Getting Started for Users

See Install instructions.

Getting Started for Developers

Environment Setup

# 1. Clone the repository
git clone https://github.com/morphonets/pysnt.git
cd pysnt

# 2. Create and activate a virtual environment (Python 3.11+, JDK 21 available)
python -m venv .venv
source .venv/bin/activate          # Windows: .venv\Scripts\activate

# 3. Install PySNT in development mode, with all optional and dev dependencies
pip install -e ".[dev]"

# 4. Verify the setup
python -c "import pysnt; print(pysnt.version())"
python -m pytest tests -q

# 5. Have a look at ./dev/README.md for common workflows

Prefer conda/mamba? mamba env create -f environment-dev.yml && mamba activate pysnt-dev && pip install -e .[dev] does the same. Select the resulting interpreter in your IDE (see the install docs).

Project Structure

pysnt/
├── src/pysnt/           # main package source (one subpackage per curated SNT package)
├── tests/               # test suite (JVM-dependent tests skip themselves if SNT cannot be resolved)
├── docs/                # documentation source (Sphinx/MyST) and tutorial notebooks
├── dev/                 # development utilities and templates (see dev/README.md)
├── dev/scripts/         # SNT sync, stub generation, and documentation scripts
├── .github/workflows/   # CI: tests, lowest-dependency tests, docs build
├── pyproject.toml       # packaging metadata, dependency floors, and extras
├── requirements*.txt    # pip equivalents (core, optional features, development)
└── environment*.yml     # conda-compatible equivalents (min, full, development)

Dependencies

pyproject.toml contains all details: core dependencies are listed under dependencies, and optional features under extras (display, gui, java, info, all, dev, docs). The requirements*.txt and environment*.yml files mirror it; The floors are verified by the "lowest declared dependencies" CI job.

Versioning and SNT

PySNT's major.minor version matches the SNT release it targets (PySNT 5.1.x targets SNT 5.1.x); the patch number is independent, so PySNT can ship fixes without a new SNT release. By default, pysnt.initialize() downloads SNT from Maven at DEFAULT_SNT_VERSION (see src/pysnt/_endpoints.py). Users can override it with initialize(snt_version=...) or SNT_VERSION, or use a local Fiji instead. PySNT warns when the loaded SNT differs in major.minor. See dev/README.md for the update workflow.

Need Help?

• 📖 Documentation: pysnt.readthedocs.io
• 💬 Forum: forum.image.sc/tag/snt
• 🐛 Issues: GitHub Issues