QuatIca provides robust implementations of quaternion matrix operations (QR, LU, QSVD, Hessenberg form, etc), solvers, and applications.
Choose your installation method based on your needs:
Option 1: GitHub (Full Framework)
For examples, benchmarks, and Jupyter notebooks:
git clone https://github.com/vleplat/QuatIca
cd QuatIca
python -m venv quat # Create virtual environment "quat"
# Activate environment:
source quat/bin/activate # Linux/macOS
.\quat\Scripts\activate # Windows
pip install -r requirements.txt # Install dependencies
Includes:
12+ ready-to-run demos (python run_analysis.py <script>)
Validation plots and benchmark scripts
Jupyter notebooks for interactive exploration
Option 2: PyPI (Lightweight Library)
pip install quatica
Verify installation:
import quatica
print(quatica.__version__) # Should return 0.1.3+
GitHub Users
# Interactive tutorial
python run_analysis.py tutorial
# Image deblurring benchmark
python run_analysis.py deblur_benchmark
PyPI Users
from quatica import create_test_matrix, NewtonSchulzPseudoinverse
# Create random quaternion matrix
A = create_test_matrix(100, 50)
# Compute pseudoinverse
solver = NewtonSchulzPseudoinverse(gamma=1,verbose=True)
A_pinv, residuals, covariances = solver.compute(A)
Colab Demos
Test core functionality → Open
Image Completion → Open
Lorenz attractor filtering → Open
Image Deblurring → Open - Open
Pseudoinverse benchmark → Open
Q-GMRES benchmark → Open
Matrix Operations: QR, LU, SVD, eigenvalue decompositions, and more
High-Performance Solvers:
Newton-Schulz pseudoinverse (16s for 800×1000 matrices)
Q-GMRES with LU preconditioning
Applications:
Image deblurring (QSLST/NS methods)
3D Lorenz attractor signal processing
Tensor Tools preview : Order-3 quaternion tensor operations, more coming soon !
GitHub Repository: github.com/vleplat/QuatIca
Full Documentation: vleplat.github.io/QuatIca
PyPI Package: pypi.org/project/quatica
Colab Demo: Open in Google Colab