Version-1.30.0

Release Date : May 2026

Major Features

New BackgroundSubtractionDialog: dedicated popup containing all background subtraction controls, separating them from the main QF window for a cleaner workflow
New background_search.py engine: automated parameter search for background subtraction with optimization cache (optimization_cache.py) to reuse best configurations across images
Manual background assignment dialog (ManualBackgroundAssignmentDialog) for per-image overrides
Background configurations table in the dialog: stores multiple named configurations; best match is auto-selected during batch processing
Fold symmetry scoring (fold_symmetry.py): quantitative metric for evaluating quadrant-fold quality; integrated into QuadrantFolder processing and exported to CSV
Layer-line mask support added to QF processing
ROI (Region of Interest) support in QuadrantFolder with ROI translation for headless processing
“Save Cropped Image” option removed; only full-size folded images are saved
QF settings bindings (qf_settings_bindings.py) and defaults (qf_defaults.py) centralized for consistent configuration management
Caching refactored: proper invalidation of background subtraction states; images removed from info dict to reduce cache size

New QFAlignmentDialog: non-modal dialog reusing the AISE ImageAlignmentWidget for detecting misaligned images within a QF dataset
“Detect Image Alignment” button added to QuadrantFoldingGUI; dialog broadcasts changes back to the main window via signal
Fold symmetry score surfaced in the alignment table with configurable normalized threshold

Analytical Jacobian fast-path for GMM fitting (scipy.optimize.least_squares): significant speedup for cases with ≥ 10 peaks
Vectorised L3 batch Gaussian evaluation applied to production layerlineModel
Hull-range slicing: fit arrays are cropped to the hull range before fitting, reducing unnecessary computation
ProjectionProcessor refactored to use least_squares (TRF) solver; Gaussian and Voigt model evaluations restructured

Box Intensity Stats tool (BoxStatsTool): multi-shot rectangular ROI tool for computing region intensity statistics; boxes are editable and numbered; accessible via a dedicated popup panel
Radius tool (RadiusTool): interactive radial selection on the image
Unit selection for calibrated distance display (nm or q-space)
Center marker overlay on the image
Tab switching dynamically repositions navigation controls
Status bar formatting and coordinate display further optimized
Quadrant-folded image detection added to XRayViewer

Renamed from TotalDisplayIntensity to TotalDiffractionIntensity across the entire project
Rebuilt UI around ProcessingWorkspace for consistency with other modules
Radial range selection added via RadiusTool
Status bar enhancements

New eq_settings_bindings.py: centralized classification and loading/saving of Equator settings
Settings are now loaded from and bound to <dataset>/settings/calibration.info; headless path also sources calibration from the same file
New EQ_FittingTab: dedicated fitting settings tab extracted from the main Equator window
Equator cache size reduced: images removed from the info dict

Headless: EquatorWindowh, QuadrantFoldingh, ProjectionTracesh updated to use SettingsManager for manual center and rotation; output directory validation added
CPU limiting: headless Equator batch processing can be bounded by worker count
EquatorImage: interpolation for ignored hull values; mask threshold extended to allow negative values; improved gap interpolation mechanisms
Image Blank Dialog: integrated ImageViewerWidget for preview
Tooltips: comprehensive tooltips added across AISE, AIME, QF, PT, navigation controls, image mask tool, and center/rotation settings widgets
Cache: ImageData reduced in size by not storing raw images; cache relative-path fixes

TotalDisplayIntensity.py replaced by TotalDiffractionIntensity.py
Background subtraction functions removed from QuadrantFoldingGUI and centralized in BackgroundSubtractionDialog and background_search.py

Batch processing rewritten to use ProcessPoolExecutor for true parallel execution; proxy classes handle lightweight data transfer between GUI and worker processes
New process_one_qf_image headless function in mp_executor.py for batch-mode processing without GUI dependencies
User-initiated stop: “Stop” button cleanly cancels in-flight tasks with an indeterminate progress dialog during wind-down; prevents retries on stopped tasks
Progress bar logic refactored to use taskManager statistics as the single source of truth
Error handling improved: invalid images (all-zero/negative average fold) raise a ValueError with a specific GUI warning
FWHM calculation fixed to handle None values gracefully

XRayViewerGUI now uses ProcessingWorkspace instead of raw ImageNavigatorWidget, gaining consistent file management, display options, navigation, and calibration dialog handling shared with other modules

Startup flow refactored: workspace signals are connected before user interaction; CSV manager and first-run processing are now triggered lazily on the first imageDataReady signal instead of a forced browseFile() call at startup
Removed browseFile() method and scan timer; file loading now fully delegated to ProcessingWorkspace / navigator
Error handling improved: CancelledError and empty error strings are treated as real errors (check for 'image' key in task result instead of not task.error)
Status bar scan indicator (*) now derived from file_manager.is_scan_done() instead of a provisional flag

Status bar path label now uses QSizePolicy.Ignored horizontal policy, eliminating a runaway feedback loop where setting elided text would grow the label’s size hint, widen the window, compute a larger available width, show more text, and grow the window again
Introduced _statusPathFullText cache and _elideStatusPath() method; elision is driven by an eventFilter on the path label that fires on every Qt layout resize — no manual statusBar.width() arithmetic needed

.. note:: Version 1.30.0 is a major feature release. This version has been tested on Python 3.10 on Ubuntu 22.04.

Draw Mask: fixed draw mask tool not working in all installation environments
- New cross-environment launcher (drawmask_launcher.py): transparently handles venv, standard pip install, and PyInstaller-frozen .deb packages
- ImageMaskDialog and ImageMaskTool now use run_pyfai_drawmask() instead of calling pyFAI-drawmask via os.system()
- PyInstaller hooks for pyFAI and silx updated to collect all resources (OpenCL kernels, Qt descriptors, calibrant tables, icons) required by drawmask at runtime
- Added qtawesome as a missing dependency

Launcher – environment check: version mismatch messages clarified—warnings no longer imply a hard failure; output label changed from “Test results” to “Log excerpt”
Dependencies: requirements file simplified; setup.py is now the single source of truth for package dependencies

Total Changes: ~192 commits since v1.29.0