Optiland v0.5.7 Release Notes

v0.5.7 delivers a major architectural upgrade with new geometry capabilities, backend-agnostic optimization, and improved performance across modeling and analysis. This release focuses on extensibility, clarity, and computational efficiency while preparing the foundation for advanced features like GRIN modeling and next-generation optimization workflows.

<p align="center"> <img width="573" height="371" alt="image" src="https://github.com/user-attachments/assets/f390a6f6-e45a-4941-8118-d16ce5f982d8" /> <br /> <em>Simple example window demonstrating new NURBS functionality. Contribution by @mattemilio</em> </p>

🧩 Core Geometry & Modeling

• NURBS Geometry Support (New Major Feature)
  Introduces full Non-Uniform Rational B-Spline (NURBS) surface modeling, enabling smooth freeform and parametric designs with unprecedented flexibility.
  Includes integration with surface visualization and sag computation tools.

• Propagation Model Framework
  Added a generalized propagation model interface that supports current field propagation and will serve as the foundation for GRIN and wave-based modeling in future versions.

• Interaction Models on Surfaces
  Decoupled ray-surface interaction logic through modular interaction models, improving separation of concerns and paving the way for exotic surface behaviors (diffractive, reflective, refractive, hybrid).

• Field System Refactor
  Reworked the internal field framework to simplify extension and customization, improving maintainability and consistency across analysis modules.

  • Added Paraxial Image Height as a new field type for rapid paraxial evaluation and comparison with full raytrace results.

• MMDFT PSF Calculation
  Added a Matrix-Multiply Discrete Fourier Transform (MMDFT) method for efficient, high-accuracy point spread function computation, especially for large pupils or fine sampling grids.

• Pixel Pitch Option for Huygens PSF
  Introduced a pixel_pitch parameter for Huygens-based PSF analyses, allowing precise control over spatial sampling and detector mapping.

• AOI Operand for Optimization
  Added Angle of Incidence as a new optimization operand, enabling direct optimization of surface angles and incidence-dependent performance.

🧠 Optimization & Backend Integration

• Torch Optimization Rework (Major Update)
  Completely restructured optimization backend to allow PyTorch-driven optimization within the standard Optiland framework.

  • Supports backend-agnostic optimization problems that can run seamlessly on either native or torch-based backends.
  • Introduces a torch-like API for differentiable and gradient-based optimization workflows.
  • Includes new documentation and examples demonstrating torch-based optimization.

• Modular Variable Scaling Framework
  Refactored scaling of optimization variables into a clean modular system, improving transparency, tuning flexibility, and code clarity.

⚙️ Performance & Robustness Improvements

• Caching of Material Index and Absorption Calculations
  Added intelligent caching to material property computations, significantly improving performance during repeated evaluations and optimization.

• Wavefront Error Calculation Fixes
  Improved robustness and accuracy in wavefront error computation, addressing numerical edge cases and filtering issues.

• Forbes Geometry Improvements
  Enhanced numerical stability and handling of Forbes polynomial surfaces.

• General Bug Fixes and Cleanups
  Numerous fixes across modules, including visualization bugs, doc build issues, and cleanup of obsolete files and redundant utilities.

• pyproject.toml Compatibility Updates
  Adjusted dependencies to resolve version mismatches with numba and typing-extensions.

🖥️ Graphical User Interface

• Stability and Responsiveness Improvements
  Multiple GUI fixes addressing visual refresh issues and interactivity problems in the lens editor and analysis panels.

• Quality-of-Life Enhancements
  Better synchronization between system state and visualization components, improving reliability for multi-surface or diffractive designs.

📚 Documentation & Examples

• RTD Documentation Expansion
  Improved and expanded ReadTheDocs documentation with clearer structure, new code examples, and setup guidance.

• New Example Systems

  • Czerny–Turner Spectrometer: Added a fully functional diffractive spectrometer example.
  • Additional doc examples highlighting NURBS usage, torch optimization, and new PSF modes.

<p align="center"> <img width="534" height="525" alt="image" src="https://github.com/user-attachments/assets/3481197a-a591-4dca-bd96-fb1072ddf49c" /> <br /> <em>Czerny–Turner Spectrometer example added in v0.5.7</em> </p>

Optiland v0.5.7 reinforces the platform's foundation with powerful modeling primitives, a unified optimization framework, and improved computational efficiency. This release establishes a strong base for upcoming GRIN modeling, wave-based propagation, and deep integration of differentiable optics.

Contributors in v0.5.7: @mattemilio, @crnh, @Littie28, @rjmoerland, @drpaprika, @scottpaine, @manuelFragata, @hemkumarsrinivas