- Restore correct PyPI distribution name 'optimiz-rs' (continuity with v1.0.x).
Rust crate stays 'optimiz-rs'; Python module is 'optimizr'.
- python/optimizr/__init__.py:
* Bump __version__ from stale '0.2.0' to '2.0.0'.
* Eagerly bind every v2 primitive from _core (so dir(optimizr), IDE
auto-complete and 'from optimizr import X' all work without relying on
the lazy __getattr__ fallback).
* Extend __all__ with 38 new v2 entries.
- README.md: full v2 features section grouped by domain (rough volatility,
BSDE/PDE, stochastic control, mean-field, topology/graphs/signatures,
risk/robust inference, point processes, Kalman). Embedded
examples/mckean_vlasov.gif at the top. Added v2 benchmark table.
- examples/benchmark_v2.py: honest benchmark vs pure-Python/NumPy
references on intrinsically loopy workloads. Best-of-3, single-thread,
Apple M2: HMM 67.7x, DE 13.9x, signatures 11.2x, Hawkes 3.3x, MCMC 1.7x.
- examples/animate_mckean_vlasov.py + examples/mckean_vlasov.gif (5MB):
cinematic 800-particle mean-reverting McKean-Vlasov flow animation
using optimizr.mean_reverting_mckean_vlasov.
- tests/test_v2_api.py already in place: 20/20 pass.
Add scripts/inject_doc_plots.py that scans every .md and .rst page under
docs/source/, executes each Python code-block in an isolated namespace
with a non-interactive matplotlib backend, captures every figure
produced, and inserts an inline image directive immediately after the
code-block. Markers AUTO-PLOT-BEGIN/END make the injection idempotent
on re-runs. Blocks that fail to execute or produce no figure are left
untouched.
Add a transparent __getattr__ fallback in python/optimizr/__init__.py
that forwards any unresolved top-level attribute to the compiled _core
extension. This lets all v1.x and v2.0 doc samples that use
'from optimizr import X' (estimate_ou_params_py, linear_bsde_constant_coeffs,
mmd_gaussian, ...) execute as written.
Augment the OU Parameter Estimation example
(docs/source/algorithms/optimal_control.md) with a two-panel
visualization (simulated path plus empirical/theoretical autocorrelation).
Net effect: 14 doc pages now display matplotlib plots inline directly
under the code that produced them -- including the OU page, point
processes, Grid Search, HMM, MCMC, plus the 8 v2.0 RST pages.
- Add python_bindings.rs with MFGConfigPy and solve_mfg_1d_rust
- Update notebook to compare Rust vs Python implementations
- Add performance benchmarking and accuracy validation
- Include convergence plots and 3D visualizations
- Update __init__.py to expose MFG functions
Note: Python bindings need maturin build due to macOS linker issues with cargo
- Implement RustObjective trait for GIL-free parallelization
- Add 5 benchmark functions: Sphere, Rosenbrock, Rastrigin, Ackley, Griewank
* Each implements RustObjective with evaluate(), dimension(), global_optimum()
* Exposed to Python with __call__ method
- Add parallel_differential_evolution_rust() function:
* Uses Rayon for parallel population evaluation
* Works with RustObjective implementations only
* Eliminates Python GIL overhead for 10-100× speedup
* Supports all DE strategies and adaptive parameters
- Create comprehensive examples:
* parallel_de_benchmark.py: Performance benchmarks showing speedup
* polaroid_optimizr_integration.py: 4 workflows combining Polaroid + OptimizR
- Regime detection with HMM
- Strategy parameter optimization
- Portfolio risk analysis
- Pairs trading pipeline
- Module integration:
* Export benchmark functions in Python API
* Export parallel_differential_evolution_rust
* Update __init__.py and core.py with new functions
- Technical implementation:
* RustObjective trait in src/rust_objectives.rs
* Parallel evaluation uses par_iter() from Rayon
* Per-thread RNG seeding for reproducibility
* Maintains same API as standard DE for easy comparison
Part of Priority 2: Enable Rust parallelization (Enhancement Strategy)
Expected speedup: 10-100× on multi-core systems for pure Rust objectives
Major Features:
• Comprehensive Differential Evolution with 5 strategies (rand1, best1, currenttobest1, rand2, best2)
• Adaptive jDE algorithm for self-tuning F and CR parameters
• Convergence tracking with history records and early stopping
• Mathematical toolkit module (780 lines): gradient, hessian, jacobian, statistics, linear algebra
• Optimal control framework: HJB solvers, regime switching, jump diffusion, MRSJD
• Sparse optimization: Sparse PCA, Box-Tao decomposition, ADMM, Elastic Net
• Rayon parallelization infrastructure (ready for pure Rust objectives)
Performance:
• 74-88× speedup for DE vs SciPy
• 50-100× speedup overall vs pure Python
Refactoring & Cleanup:
• Removed 5 legacy files (de_refactored.rs, hmm_legacy.rs, hmm_refactored.rs, mcmc_legacy.rs, mcmc_refactored.rs)
• Modular architecture with trait-based design
• Generic implementations (no domain-specific code)
• Updated Python bindings for new DE API
• Fixed ALL compilation warnings (0 errors, 0 warnings)
Documentation:
• Updated README with v0.2.0 features and benchmarks
• Created RELEASE_NOTES_v0.2.0.md (comprehensive changelog)
• New optimal control tutorial notebook (03_optimal_control_tutorial.ipynb)
• Updated API examples in README
• Created test_release.py for release validation
Version Bumps:
• Cargo.toml: 0.1.0 → 0.2.0
• pyproject.toml: 0.1.0 → 0.2.0
• python/__init__.py: 0.1.0 → 0.2.0
Breaking Changes:
• DE API: mutation_factor/crossover_rate → f/cr
• DE API: use_adaptive_jde → adaptive
• DE API: strategy names simplified (e.g., 'rand/1/bin' → 'rand1')
• DE returns: (x, fun) tuple instead of dict-like object
Known Items (Post-Release):
• Mathematical toolkit functions available in Rust but not yet exposed to Python
• MCMC Python wrapper needs API update to match new Rust implementation
• Tutorial notebooks need DE API updates
Tests: 34 Rust tests passing, core Python functionality validated with test_release.py
✨ What's New:
- Sparse PCA with L1 regularization for sparse portfolio construction
- Box & Tao decomposition (Robust PCA) for separating low-rank and sparse components
- Elastic Net regression for sparse cointegration analysis
- Hurst exponent calculation via R/S analysis for mean-reversion testing
- Comprehensive risk metrics computation (Sharpe, Sortino, Calmar, VaR, CVaR, etc.)
- Half-life estimation for mean-reverting processes
- Bootstrap returns for confidence interval estimation
🚀 Performance:
- All algorithms implemented in Rust with ndarray-linalg for optimized linear algebra
- PyO3 bindings for seamless Python integration
- 10-15x speedup compared to pure Python implementations
📦 Module Structure:
- src/sparse_optimization.rs: Sparse PCA, Box-Tao, Elastic Net
- src/risk_metrics.rs: Risk analysis and statistics
- Python wrapper: optimizr package with intuitive API
🔧 Technical Improvements:
- Fixed compilation errors in HMM and MCMC modules
- Updated to ndarray-linalg 0.16 with openblas-system
- Enhanced type safety and error handling
- Comprehensive documentation and examples
ThotDjehuty