Changelog

All notable changes to the Raman Spectroscopy Analysis Application will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Documentation

  • Comprehensive ReadTheDocs documentation system

  • User guides for all major features

  • Analysis methods reference documentation

  • API documentation for developers

  • Japanese translation (in progress)

[1.0.0-alpha] - 2026-01-24

Added

Core Features

  • Complete desktop application with PySide6/Qt6 interface

  • Multi-language support (English, Japanese)

  • Project management system with workspace organization

  • Data package management for multiple datasets

  • Group-based sample organization

Preprocessing (40+ Methods)

  • Baseline Correction: AsLS, AirPLS, Polynomial, Whittaker, FABC, Butterworth High-Pass

  • Smoothing: Savitzky-Golay, Gaussian, Moving Average, Median Filter

  • Normalization: Vector, Min-Max, Area, SNV, MSC, Quantile, PQN, Rank Transform

  • Derivatives: 1st and 2nd order Savitzky-Golay

  • Feature Engineering: Peak Ratio, Wavelet Transform

  • Advanced: Convolutional Denoising Autoencoder (CDAE)

  • Pipeline System: Save, load, and share preprocessing pipelines

  • Real-time Preview: See effects before applying

Analysis Methods

  • Exploratory:

    • Principal Component Analysis (PCA) with loadings and scree plots

    • UMAP for non-linear dimensionality reduction

    • t-SNE for cluster visualization

    • Hierarchical Clustering with dendrograms

    • K-means Clustering with elbow method

  • Statistical:

    • Pairwise tests (t-test, Mann-Whitney U, Wilcoxon)

    • Multi-group comparisons (ANOVA)

    • Correlation analysis (Pearson, Spearman, Kendall)

    • Band ratio analysis with customizable ranges

    • Peak detection and identification

  • Visualization:

    • Interactive heatmaps

    • Waterfall plots

    • Overlaid spectra with group coloring

    • Peak scatter plots

    • Correlation matrices

Machine Learning

  • Algorithms:

    • Support Vector Machine (SVM)

    • Random Forest (RF)

    • XGBoost

    • Logistic Regression

    • Linear Regression

  • Validation:

    • GroupKFold cross-validation (patient-level splitting)

    • Leave-One-Patient-Out (LOPOCV)

    • Stratified K-Fold

    • Hold-out test sets

  • Evaluation:

    • ROC curves with AUC scores

    • Confusion matrices

    • Classification reports (precision, recall, F1)

    • Calibration curves

  • Interpretability:

    • Permutation importance

    • Feature importance plots mapped to wavenumbers

  • Export: Save trained models in pickle or ONNX format

Build System

  • Windows Portable: Single executable (~375 MB) with all dependencies

  • Windows Installer: NSIS-based professional installer

  • Build Scripts: Automated PowerShell build system

  • Testing: Comprehensive test suite for executables

Fixed

January 2026

  • Analysis Page Stability (2026-01-23):

    • Two-stage stop mechanism for safe cancellation

    • Improved outlier detection performance with PCA-reduced space

    • Fixed correlation heatmap tick clipping

    • Registered bundled fonts for proper EN/JA plot rendering

    • Fixed ML dropdown black popup styling

    • Improved band ratio plot embedding with proper PathPatch preservation

  • ML Evaluation Enhancements (2026-01-22):

    • Added comprehensive evaluation summary tab

    • Implemented data leakage warnings

    • Dataset-level performance metrics

  • Grouped Mode Analysis (2026-01-21):

    • Fixed PCA plots in grouped mode with proper tab organization

    • Improved runtime performance with shared component sync

  • ML UI Improvements (2026-01-20):

    • Major ML page UI refactor for better usability

    • Enhanced group management with drag-and-drop

    • Improved i18n coverage across ML interface

    • Optimized performance for large datasets

October 2025

  • Parameter Type Validation (2025-10-15):

    • Fixed FABC baseline correction integer conversion

    • Implemented two-layer type validation system

    • All 40 preprocessing methods validated (100% pass rate)

    • Zero breaking changes, full backwards compatibility

  • Preprocessing UI/UX (2025-10-08):

    • Fixed pipeline eye button crash (index out of range)

    • Fixed derivative order parameter empty field

    • Fixed feature engineering enumerate bug

    • Fixed deep learning module syntax errors

  • UI Polish (2025-10-07):

    • Input datasets layout optimization (show 3-4 items minimum)

    • Pipeline step selection visual feedback (darker background)

    • Changed pipeline add button color (blue → green)

    • Section title standardization across pages

Changed

  • Documentation Structure: Separated public docs/ from local .docs/ for better organization

  • README Organization: Removed issues/problems from README, moved to troubleshooting docs

  • Build System: Upgraded to PyInstaller 6.16.0+ for better compatibility

Security

  • Type Validation: Comprehensive parameter type checking prevents injection attacks

  • Path Validation: All file operations validate paths to prevent directory traversal

[0.1.0] - 2025-10-01

Added

  • Initial alpha release

  • Basic preprocessing pipeline

  • PCA analysis

  • Simple machine learning integration

Release Notes

Version 1.0.0-alpha

This is the first alpha release of the Raman Spectroscopy Analysis Application, developed as a final year project at the University of Toyama.

Status: Alpha - Feature complete but undergoing testing and refinement

Recommended for:

  • Research laboratories

  • Academic institutions

  • Method development and validation

Not recommended for:

  • Clinical diagnostic use (not approved)

  • Production medical systems

Known Limitations:

  1. Some preprocessing methods require further validation

  2. Deep learning features require GPU for practical use

  3. Large datasets (>5000 spectra) may require optimization

  4. Some advanced features are planned but not yet implemented

Upcoming Features (v1.1.0):

  • Additional spectral unmixing methods (NMF, ICA)

  • Enhanced batch processing capabilities

  • REST API for remote processing

  • Command-line interface

  • Additional file format support (SPC, WDF)

  • Video tutorials

  • Japanese documentation completion

  • Malay translation

Migration Guides

From v0.1.0 to v1.0.0-alpha

Breaking Changes: None - v1.0.0 is backwards compatible

New Features to Note:

  1. GroupKFold is now default for ML validation - ensures patient-level splitting

  2. SHAP values now available for all tree-based models

  3. New preprocessing methods - Review Preprocessing Guide

Recommended Actions:

  1. Update preprocessing pipelines to use new methods

  2. Re-run ML experiments with GroupKFold validation

  3. Export results with SHAP interpretability

Contributors

Core Development

  • Muhammad Helmi bin Rozain - Lead Developer, BSc Student

    • GitHub: @zerozedsc

    • University of Toyama (富山大学)

Supervision

Acknowledgments

  • University of Toyama for research facilities and support

  • Laboratory members for testing and feedback

  • Open-source community for underlying libraries:

    • PySide6/Qt - UI framework

    • RamanSPy - Raman spectroscopy tools

    • scikit-learn - Machine learning

    • pybaselines - Baseline correction

    • NumPy, SciPy, pandas - Numerical computing

    • matplotlib - Visualization

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use this software in your research, please cite:

@software{rozain2025raman,
  author = {Rozain, Muhammad Helmi bin},
  title = {Raman Spectroscopy Analysis Application: A Comprehensive Platform for Real-Time Spectral Classification},
  year = {2025},
  version = {1.0.0-alpha},
  publisher = {GitHub},
  url = {https://github.com/zerozedsc/Raman-Spectroscopy-Analysis-Application},
  institution = {University of Toyama, Laboratory for Clinical Photonics and Information Engineering}
}

Support

Note: This software is intended for research use only and is not approved for clinical diagnostic purposes.