StatOxide: High-performance statistical computing in Rust with Python bindings
StatOxide is a modern, high-performance statistical computing library written in Rust, with comprehensive Python bindings via PyO3. Designed for data scientists, statisticians, and researchers who need both performance and productivity.
🚀 Features
📊 Core Data Structures
- Series: Columnar data with metadata (name, dtype, levels)
- DataFrame: Tabular data structure with column operations
- Formula: R-style formula parsing for model specification
📈 Statistical Functions
- Descriptive Statistics: Mean, variance, skewness, kurtosis, quantiles
- Probability Distributions: 12 continuous + 6 discrete distributions
- Statistical Tests: t-test, chi-square, ANOVA, correlation tests
- Correlation Measures: Pearson, Spearman, Kendall tau
🧮 Statistical Models
- Linear Models: OLS, Ridge, Lasso, Elastic Net with proper inference
- Generalized Linear Models: Logistic, Poisson, Gamma, Negative Binomial regression
- Mixed Effects Models: Linear and GLMMs with EM algorithm estimation
- Robust Statistics: M-estimators, S-estimators, MM-estimators
- Nonparametric Methods: Kernel regression, local regression, smoothing splines
📉 Time Series Analysis
- Core Structures:
TimeSerieswith datetime indexing - ARIMA Models: AR, MA, ARMA, ARIMA, SARIMA
- GARCH Models: ARCH, GARCH for volatility modeling
- Decomposition: STL, moving averages, Hodrick-Prescott filter
- Forecasting: Point forecasts, prediction intervals
🛠️ Utilities
- Linear Algebra: Matrix operations, solvers, decompositions
- Random Generation: Distributions, bootstrap, train-test split
- Data Validation: Type checking, missing value detection
- Numerical Methods: Softmax, standardization, normalization
🐍 Python API
StatOxide provides a complete Python interface through PyO3 bindings:
import statoxide
import statoxide.core as soc
import statoxide.stats as sos
# Core data structures
df = soc.DataFrame({
"x": [1.0, 2.0, 3.0, 4.0, 5.0],
"y": [2.0, 4.0, 5.0, 4.0, 5.0]
})
series = df.get_column("x")
print(f"Mean of x: {series.mean():.2f}")
print(f"Std of x: {series.std(1.0):.2f}")
# Statistical functions
print(f"Correlation: {sos.correlation(df.get_column('x').to_list(),
df.get_column('y').to_list()):.3f}")
summary = sos.descriptive_summary([1.0, 2.0, 3.0, 4.0, 5.0])
print(f"Summary: {summary}")
# Formula parsing
formula = soc.Formula("y ~ x + x^2")
print(f"Formula variables: {formula.variables()}")
# Models
import statoxide.models as som
result = som.linear_regression([[1, 1], [1, 2], [1, 3]], [5, 8, 11])
print(f"Regression coefficients: {result['coefficients']}")
# Mixed effects models
mixed_results = som.mixed_effects("y ~ x + (1 | group)", data)
print(f"Random effect variance: {mixed_results.random_variances}")
# Time series
import statoxide.tsa as sot
arima_result = sot.fit_arima([1.0, 2.0, 3.0, 4.0, 5.0], 1, 0, 1)
print(f"ARIMA AIC: {arima_result['aic']}")
# Utilities
import statoxide.utils as sou
train, test = sou.train_test_split([1.0, 2.0, 3.0, 4.0, 5.0], 0.2)
print(f"Train: {train}, Test: {test}")
🏗️ Architecture
StatOxide is organized as a multi-crate Rust workspace:
statoxide/
├── Cargo.toml # Workspace configuration
├── crates/
│ ├── so-core/ # Core data structures & formula parsing
│ ├── so-linalg/ # Linear algebra abstraction
│ ├── so-stats/ # Statistical functions & distributions
│ ├── so-models/ # Statistical models (regression, GLM, mixed effects, etc.)
│ ├── so-tsa/ # Time series analysis
│ ├── so-utils/ # Utility functions
│ └── so-python/ # Python bindings (PyO3)
├── assets/logo.png # Project logo
├── LICENSE-MIT # MIT license
└── LICENSE-APACHE-2.0 # Apache 2.0 license
📦 Installation
Prerequisites
- Rust Toolchain:
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh - Python Development Files:
- Ubuntu/Debian:
sudo apt-get install python3-dev python3.11-dev - macOS:
brew install python@3.11
- Ubuntu/Debian:
- Maturin (recommended):
pip install maturin
Building from Source
# Clone the repository
git clone https://github.com/EthanNOV56/StatOxide.git
cd StatOxide
# Build Python bindings with maturin
cd crates/so-python
maturin develop # Editable install for development
# or
maturin build --release # Build wheel for distribution
Direct Cargo Build
cd /path/to/statoxide
export PYO3_PYTHON=python3.11
cargo build --release --package so-python
The shared library will be at target/release/libso_python.so.
🧪 Testing
Rust Tests
cargo test --all
Python Tests
After installation:
python -c "import statoxide; print(statoxide.version())"
python crates/so-python/test_api.py # API demonstration
📚 Documentation
- API Reference: Run
cargo doc --all --no-deps --openfor Rust documentation - Python Docstrings: All Python functions include detailed docstrings
- Examples: See
crates/so-python/test_api.pyfor usage examples
🎯 Design Principles
- Performance: Leverage Rust's zero-cost abstractions and LLVM optimizations
- Safety: Memory safety guarantees without garbage collection
- Interoperability: Seamless Python integration with minimal overhead
- Modularity: Independent crates for clear separation of concerns
- API Consistency: Familiar interfaces inspired by R, pandas, and statsmodels
🔧 Development Status
| Module | Status | Notes |
|---|---|---|
| so-core | ✅ Complete | Data structures, formula parsing |
| so-linalg | ✅ Complete | Linear algebra abstraction |
| so-stats | ✅ Complete | Statistical functions & distributions |
| so-models | ✅ Complete | Regression, GLM, mixed effects, robust, nonparametric |
| so-tsa | ✅ Complete | ARIMA, GARCH, decomposition, forecasting |
| so-utils | ✅ Complete | Random generation, validation, numerical methods |
| so-python | ✅ Complete | Full Python bindings implemented |
📄 License
StatOxide is dual-licensed under both:
- MIT License: See LICENSE-MIT for details
- Apache License 2.0: See LICENSE-APACHE-2.0 for details
You may use StatOxide under either license at your option.
🙏 Acknowledgments
- R and statsmodels for statistical API inspiration
- pandas for DataFrame design patterns
- PyO3 team for excellent Rust-Python interop
- ndarray and faer for numerical computing foundations
🤝 Contributing
Contributions are welcome!
- Fork the repository
- Create a feature branch
- Make your changes
- Run tests:
cargo test --all - Submit a pull request
📞 Support
- Issues: GitHub Issues
- Repository: GitHub Repository
High-performance statistics meets Python productivity
Metadata
Release files for statoxide 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| statoxide-0.3.0.tar.gz | 527.7 kB | Without provenance | Details for statoxide-0.3.0.tar.gz |
Release files / statoxide-0.3.0.tar.gz
| Download URL | statoxide-0.3.0.tar.gz |
|---|---|
| Size | 527.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
98796cd81a275085d57d681989c03d51cd88ab160ce19b5237da2e8499d2dfff
|
|
BLAKE2b-256 checksum How to use checksums |
cbab73ef8da4cc7e769be4b76928e118c3d7c8f59d6e9f1a838e2c013cae0ddc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.9.7
|
Security for statoxide 0.3.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.