vascusim 0.2.0

Dataset package for cardiovascular simulations

VascuSim

A Python package for working with cardiovascular simulation datasets from the BioSiMMLab.

Overview

VascuSim provides tools for processing, streaming, and analyzing vascular simulation data stored in VTU/VTP format, with efficient conversion to PyTorch Geometric data formats for graph neural network training.

Key features:

• Loading and conversion of VTU/VTP files to PyTorch Geometric data format
• Streaming functionality for efficient access to large datasets
• Support for various data sources (local files, NAS, Hugging Face)
• Comprehensive geometric processing utilities
• Visualization tools for vascular structures
• Domain decomposition for distributed processing

📚 Documentation | 📦 PyPI Package | 🐛 Issue Tracker

Installation

Basic Installation

pip install vascusim

Development Installation

git clone https://github.com/BioSiMMLab/vascusim.git
cd vascusim
pip install -e ".[dev]"

Optional Dependencies

For SMB/CIFS support (for NAS streaming):

pip install "vascusim[smb]"

For documentation building:

pip install "vascusim[docs]"

Usage Examples

Loading Vascular Data

import torch
from vascusim.data.conversion import vtu_to_pyg

# Convert a VTU file to PyTorch Geometric format
data = vtu_to_pyg("path/to/simulation.vtu")

# Access node positions and connectivity
pos = data.pos
edge_index = data.edge_index

# Access node attributes
velocity = data.node_velocity  # if available in the VTU file
pressure = data.node_pressure  # if available in the VTU file

Using the Dataset API

from vascusim.data.dataset import VascuDataset

# Create a dataset from a local directory
dataset = VascuDataset(
    source_url="path/to/data",
    cache_dir="~/.vascusim/cache",
    normalize=True
)

# Access an item
data = dataset[0]

# Use with PyTorch DataLoader
from torch_geometric.loader import DataLoader
loader = DataLoader(dataset, batch_size=4, shuffle=True)

Streaming from Remote Sources

from vascusim.data.dataset import StreamingVascuDataset

# Stream from Hugging Face dataset
hf_dataset = StreamingVascuDataset(
    source_url="huggingface.co/datasets/biosimmlab/vascu-example",
    streaming_type="hf",
    prefetch=True
)

# Stream from NAS
nas_dataset = StreamingVascuDataset(
    source_url="192.168.1.100",
    streaming_type="nas",
    username="user",
    password="pass"
)

Visualization

import matplotlib.pyplot as plt
from vascusim.utils.visualization import plot_geometry, plot_flow

# Plot geometry
fig = plot_geometry(data, show_points=True)
plt.savefig("geometry.png")

# Visualize flow fields
plot_flow(data, flow_field="velocity", color_by_magnitude=True)

Processing Vascular Geometry

from vascusim.processing.geometry import compute_curvature, extract_centerline

# Compute curvature for each node
curvature = compute_curvature(data)

# Extract centerline
centerline = extract_centerline(data, smoothing=0.1)

# Apply transformations
from vascusim.data.transforms import Normalize, AddNoise

transform = Normalize()
normalized_data = transform(data)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Run the tests (python run_test.py)
4. Commit your changes (git commit -m 'Add some amazing feature')
5. Push to the branch (git push origin feature/amazing-feature)
6. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citing

If you use VascuSim in your research, please cite:

@software{vascusim2025,
  author = {Xu, Wenzhuo},
  title = {VascuSim: Dataset package for cardiovascular simulations},
  url = {https://github.com/biosimmlab/vascusim},
  year = {2025}
}

Acknowledgements

This project is maintained by the BioSiMMLab.