scientific-computing-system

scientific-computing-system (CDS) is a pure-Python computational science platform designed for numerical methods, modeling, validation, uncertainty quantification, and reproducible scientific workflows. The project emphasizes readability, portability, and inspectability over raw performance, making it suitable for education, prototyping, and environments where compiled dependencies are undesirable. The current package line (2.0.0) maintains a zero-runtime-dependency core while providing optional adapters to established scientific libraries (NumPy, SciPy, statsmodels, scikit-learn, SymPy, Z3) for cross-validation and specialized solvers. The architecture is organized into five layers: compute (quantum, signals, math, ODE/PDE, integration), analysis (statistics, probability, ML, modeling, sensitivity), assurance (validation, uncertainty, units, provenance), orchestration (workflow, optional tools), and data (data analysis, streaming I/O, knowledge). Key capabilities include: - Quantum computing: single- and multi-qubit circuits, Bell/GHZ states, entanglement utilities - Signal processing: DFT/FFT, convolution, filters, STFT, wavelet operations - Numerical methods: linear algebra, decompositions, quadrature, stiff and symplectic ODE solvers - Optimization: gradient/Newton/Adam methods, Nelder-Mead, simulated annealing, constrained search - Statistics and probability: inference, regression, statistical tests, distributions, Bayesian utilities - Machine learning: classical estimators, PCA, k-means, kNN, preprocessing, validation - Scientific domains: physical constants, genetics helpers, fractals, information theory - Data analysis: tabular analysis, normalization, visualization, streaming I/O with optional HDF5/NetCDF - Scientific assurance: dimensional analysis, uncertainty propagation, sensitivity analysis, cross-method validation, provenance tracking, approval-gated workflow orchestration The platform distinguishes itself by separating computed results from the evidence supporting them, enabling inspectable, reproducible, and falsifiable scientific workflows. Installation profiles range from the pure-Python core to full scientific backends, research data I/O, plotting, and dashboard support.