💥 New features / Enhancements:

• Added AnnotatedPolygon annotation to items
• Added make.annotated_polygon function to plotpy.builder module
• Customization of annotation information:
  • Added info_callback argument to all annotation class constructors
  • Added set_info_callback method to all annotation classes
  • The info_callback is a function that takes the annotation object and returns a string with the information to display
  • Default info_callback is redirected to the get_infos method of the annotation object (this makes the feature backward compatible)

🛠️ Bug fixes:

• PySide6 compatibility issues:
  • Fixed deprecated call to QMouseEvent in tests/unit/utils.py
  • Added workaround for QPolygonF shape point slicing
• Issue #21 - PySide6 on Linux: segfault in test_colormap_editor.py
• Fixed sliderMoved signal connection in ColorMapEditor
• Fixed AssertionError in test_rect_zoom.py (Linux, Python 3.12, PyQt6)
• Fixed typing issues in plotpy.events module:
  • Event objects were not properly typed (QtCore.QEvent instead of QtGui.QMouseEvent)
  • Event position arguments were not properly typed (QtCore.QPoint instead of QtCore.QPointF)
• Fix NumPy FutureWarning: Format strings passed to MaskedConstant are ignored [...] when displaying masked pixel coordinates

💥 New features / Enhancements:

• Added Wheel packages for all major platforms on PyPI:
  • Windows (32/64bits), MacOS, Linux
  • Python 3.8 to 3.12

🛠️ Bug fixes:

• Fixed color theme support (dark/light mode), leveraging the new guidata V3.6 feature

ℹ️ Release V2.6.0 was a fugitive release that was replaced by V2.6.1 due to a critical bug in the segment line cross section computation for masked images.

💥 New features / Enhancements:

• Added support for color theme change at execution (relies on guidata V3.6)
• Changed strategy for default plot widget size:
  • No default size is applied to the plot widget anymore (before, the default size was 800x600 pixels)
  • Added parameter size to PlotDialog, PlotWindow classes, and make.dialog, make.window functions to set the initial size of the plot widget

🛠️ Bug fixes:

• Fixed segment line cross section computation for masked images:
  • Avoided warning message when encountering masked values in the image data
  • Replacing masked values by NaNs when computing the segment line cross section (as before, but explicitely, to avoid the warning message)

ℹ️ Release V2.5.0 was a fugitive release that was replaced by V2.5.1 due to packaging issues.

In this release, test coverage is 79%.

💥 New features / Enhancements:

• Alternative dictionary argument for plot options:

  • This new feature was introduced in the context of the cyclic import bug fix, to avoid importing the plotpy.plot module just to get the PlotOptions or BasePlotOptions classes
  • All classes (and a few functions) that used to take an options argument as a BasePlotOptions or PlotOptions instance now also accept a dictionary argument with the same keys as the BasePlotOptions or PlotOptions class attributes, and the same values as the corresponding attributes
  • This concerns the following classes and functions:
    • plotpy.plot.BasePlot
    • plotpy.plot.PlotWidget
    • plotpy.plot.PlotDialog
    • plotpy.plot.PlotWindow
    • plotpy.plot.SubPlotWidget
    • plotpy.plot.SyncPlotWindow
    • plotpy.tools.RotateCropTool
    • plotpy.widgets.fit.FitDialog
    • plotpy.widgets.fliprotate.FlipRotateDialog
    • plotpy.widgets.rotatecrop.RotateCropDialog
    • plotpy.widgets.selectdialog.SelectDialog
    • plotpy.widgets.selectdialog.select_with_shape_tool

• Added "Lock LUT range" option for image items:

  • This new option is disabled by default, which matches the previous behavior: when updating an image item data, the LUT range is automatically adjusted to the new data range (if not passed as an argument to the BaseImageItem.set_data method)
  • When enabled, the LUT range is locked and the LUT range is not adjusted when updating the image item data
  • The option is available in image parameters dialog
  • A new tool LockLUTRangeTool has been implemented to toggle the option from the plot context menu: the tool is not registered by default in the plot widget, but can be added by the host application if needed
  • See test script tests.features.test_image_data_update for an example of usage of the new option and tool

• Added missing set_style method to XRangeSelection class: this method is used to set the style of the range selection item from configuration options

🛠️ Bug fixes:

• Issue #19 - Fix distutils deprecation in setup.py: replaced distutils.core by setuptools in the setup.py script to avoid the deprecation warning when building the package with Python 3.10 and 3.11, and to ensure compatibility with earlier Python versions (PlotPy is already compatible with the most recent Python versions: this only concerns the build system)
• Fix cyclic import in plotpy.tools module:
  • Some tools in plotpy.tools subpackage were importing the plotpy.plot module, which was importing the plotpy.tools module, causing a cyclic import issue
  • This is now fixed by introducing new constants for axis IDs in the plotpy.constants module, and using them everywhere in the code, thus avoiding to import the plotpy.plot module just to get the axis IDs
• Fix empty label in X/Y cross section plots:
  • This is a regression introduced in V2.1.0
  • When showing the X/Y cross section plots (using the plot context menu), an empty label was displayed at the center of each of those plots
  • The label now shows "Enable a marker" as previously
• Fix historic unexpected behavior of interactive tools:
  • When triggering an interactive tool (e.g. by clicking on the corresponding toolbar button), the tool activate method was called twice, which was not expected, but was not causing any issue given the current implementation
  • However, when defining custom interactive tools, this behavior could lead to unexpected results (i.e. really executing activation actions twice)
  • This is now fixed: the activate method is called only once when triggering an interactive tool

In this release, test coverage is 79%.

🛠️ Bug fixes:

• Issue #17:
  • Debian's Python team has reported that the contour unit test was failing on arm64 architecture
  • This is the opportunity to replace the contour2d Cython extension by scikit-image's find_contours function, thus avoiding to reinvent the wheel by relying on a more robust and tested implementation
  • The contour2d Cython extension is removed from the source code
  • The contour related features remain the same, but the implementation is now based on scikit-image's find_contours function
  • The scikit-image dependency is added to the package requirements

In this release, test coverage is 79%.

🛠️ Bug fixes:

• Contrast adjustment panel:
  • A regression was introduced in V2.0.0: levels histogram was no longer removed from contrast adjustment panel when the associated image was removed from the plot
  • This is now fixed: when an image is removed, the histogram is removed as well and the contrast panel is refreshed (which was not the case even before the regression)

In this release, test coverage is 79%.

💥 New features / Enhancements:

• Contrast adjustment panel:
  • New layout: the vertical toolbar (which was constrained in a small area on the right side of the panel) is now a horizontal toolbar at the top of the panel, beside the title
  • New "Set range" button: allows the user to set manually the minimum and maximum values of the histogram range
• New Z-axis logarithmic scale feature:
  • Added new tool ZAxisLogTool to toggle the Z-axis logarithmic scale
  • The tool is registered by default in the plot widget, like the ColormapTool
  • When enabled, the active image item is displayed after applying a base-10 logarithm to its pixel values
• Curve statistics tool CurveStatsTool is now customizable:
  • When adding the tool: plot_widget.manager.add_tool(CurveStatsTool, labelfuncs=(...))
  • Or after: plot_widget.manager.get_tool(CurveStatsTool).set_labelfuncs(...)
  • The labelfuncs parameter is a list of tuples (label, func) where label is the label displayed in the statistics table, and func is a function that takes the curve data and returns the corresponding statistic value (see the documentation for more details)
• Image statistics tool ImageStatsTool is now customizable:
  • When adding the tool: plot_widget.manager.add_tool(ImageStatsTool, stats_func=...)
  • Or after: plot_widget.manager.get_tool(ImageStatsTool).set_stats_func(...)
  • The stats_func parameter is a function that takes the image item and selected rectangle coordinates, and returns a string with the statistics to display
• New SIG_AXIS_PARAMETERS_CHANGED signal emitted by BasePlot when the axes parameters are changed (e.g. when the axes are inverted, or the scale is changed)
• New "Reverse X axis" feature:
  • Added new tool ReverseXAxisTool to toggle the X-axis direction
  • The tool is registered by default in the plot widget, like its Y-axis counterpart

🛠️ Bug fixes:

• Contrast adjustment panel:
  • Fixed histogram update issues when no image was currently selected (even if the an image was displayed and was selected before)
  • Histogram range was not updated when either the minimum or maximum value was set using the "Minimum value" or "Maximum value" buttons (which have been renamed to "Min." and "Max." in this release)
  • Histogram range was not updated when the "Set full range" button was clicked
• Image parameters: contrast range was not updated when the image Z axis bounds were changed using the "Parameters" dialog

🧹 API cleanup:

• Deprecated AnnotationParam.update_annotation method: use update_item instead
• Deprecated AxesShapeParam.update_axes method: use update_item instead
• Deprecated AxesParam.update_axes method: use update_item instead
• Deprecated ImageAxesParam.update_axes method: use update_item instead
• Deprecated LabelParam.update_label method: use update_item instead
• Deprecated MarkerParam.update_marker method: use update_item instead
• Deprecated RangeShapeParam.update_range method: use update_item instead
• Deprecated ShapeParam.update_shape method: use update_item instead

This release is mainly intended to fix the Windows binary distribution, which was not supporting NumPy 2.0.

🛠️ Bug fixes:

• Moved back conftest.py to the tests folder (was in the root folder), so that pytest can be executed with proper configuration when running the test suite from the installed package

• Removed benchmarks from automated test suite (not relevant for the end user): added plotpy-benchmarks script to run the benchmarks

In this release, test coverage is 79%.

🛠️ Bug fixes:

• Add support for NumPy 2.0:

  • Updated build system to use NumPy 2.0 on Python > 3.8 and NumPy 1 on Python 3.8
  • Use numpy.asarray instead of numpy.array(..., copy=False)
  • Use numpy.isin instead of numpy.in1d
  • Use scipy.integrate.trapezoid instead of numpy.trapz

• ColorMapManager: fix segmentation fault with PySide6 on Linux

In this release, test coverage is 79%.

🛠️ Bug fixes:

• Moved up LineCrossSection import for consistency with other cross-section panels
• Unexpected behavior regarding io.load_items and io.save_items functions:
  • Those functions were serializing and deserializing most of the parameters of the plot items, but not their visibility state
  • This is now fixed: the visibility state of the plot items is now saved and restored as expected

ℹ️ Other changes:

• Explicitely exclude NumPy V2 from the dependencies (not compatible yet)

In this release, test coverage is 79%.

Version 2.3.2 fixes a blocking issue with the colormap editor unit test introduced in version 2.3.1. The latter is a fugitive release that was not announced.

💥 New features / Enhancements:

• Colormap: added "Apply" button to the colormap manager
• Automated test suite:
  • Test coverage has been improved from 75% to 79%
  • The following features are now covered by unit tests:
    • Panning with the mouse move events
    • Zooming with the mouse wheel/move events
    • Curve statistics tool
    • Image rotation and translations via SelectTool (simulating mouse events)
    • Masked areas in images and ImageMaskTool
    • LockTrImageTool
    • Cursor tools (HCursorTool, VCursorTool, XCursorTool and HRangeTool)
    • DisplayCoordsTool: simulation of 'Alt' and 'Ctrl' keys
    • Complete coverage for MultiLineTool and FreeFormTool
    • Selection tools (RectangularSelectionTool and SelectTool)

🛠️ Bug fixes:

• Image statistics tool: fixed "No available data" message when the tool rectangular region top Y coordinate is above the image top Y coordinate
• Label items (LabelItem, LegendBoxItem, DataInfoLabel, ...) were not emitting the SIG_ITEM_MOVED signal when moved interactively (with the mouse) if the item anchor was attached to the canvas
• Colormap: fixed context menu entry update (colormap icon was updated as expected, but the colormap name was not)
• Rotate/crop dialog: added missing toolbar on plot widget
• Flip/rotate dialog: added missing toolbar on plot widget
• Fixed issue with oblique averaged cross section computation (AttributeError when clicking on the empty cross section plot)

In this release, test coverage is 75%.

💥 New features:

• Added support for colormap inversion:
  • The user can now invert the colormap of an image item:
    • From the image parameters dialog ("Invert colormap" checkbox)
    • From the plot context menu (right-click on the image item)
  • BaseImageItem.set_color_map method takes a new invert parameter (which defaults to None, meaning that the behavior is unchanged)
  • New ReverseColormapTool: registered by default in the plot widget, like the ColormapTool (add the "Invert colormap" entry in the context menu of the image)

🛠️ Bug fixes:

• ErrorBarCurveItem: fixed NumPy deprecation warning ("Conversion of an array with ndim > 0 to a scalar is deprecated [...]")

ℹ️ Other changes:

• Image plot items deserialization:
  • When an image plot item is deserialized, and needs to be reloaded from a file, the file path is adapted to the current working directory if file is not found (this is the legacy behavior).
  • An unnecessary call to ImageIOHandler.adapt_path method was removed from the RawImageItem.deserialize method: this issue has to be handled by the host application, not by the PlotPy library.
  • ImageIOHandler: removed add_change_path and adapt_path methods
• Fix typo in tests.features.test_colormap_editor module: renamed function test_colormap_manager to test_colormap_editor
• Removed unnecessary BaseImageItem.get_color_map_name method

In this release, test coverage is 75%.

New features:

• Added SIG_ITEM_PARAMETERS_CHANGED signal to BasePlot class:
  • This signal is emitted when the parameters of an item are changed using the parameters dialog, or a specific tool (e.g. the colormap selection tool, or the lock/unlock tool for image items)
  • This signal is emitted with the item as argument
  • It is often emitted before the SIG_ITEMS_CHANGED signal, which is global to all items, but not necessarily. For example, when the colormap of an image is changed, the SIG_ITEM_PARAMETERS_CHANGED signal is emitted for the image item, but the SIG_ITEMS_CHANGED signal is not emitted.
• Added new colormap presets:
  • viridis, plasma, inferno, magma, cividis
  • afmhot
  • coolwarm, bwr, seismic
  • gnuplot2, CMRmap, rainbow, turbo
• Fixed all qualitative colormaps:
  • All qualitative colormaps have been re-computed because they are not supposed to be interpolated, which was the case and made them unusable
  • The qualitative colormaps are now usable and look like the ones from Matplotlib
• Colormap manager:
  • Added a button to remove a custom colormap
  • The preset colormaps and the currently selected colormap are read-only
• Added automatic unit tests for interactive tools:
  • AnnotatedCircleTool, AnnotatedEllipseTool, AnnotatedObliqueRectangleTool, AnnotatedPointTool, AnnotatedRectangleTool, AnnotatedSegmentTool
  • AverageCrossSectionTool, CrossSectionTool, ObliqueCrossSectionTool, LineCrossSectionTool
  • EditPointTool, SelectPointsTool, SelectPointTool
  • AspectRatioTool, ImageStatsTool, SnapshotTool
  • DisplayCoordsTool, RectZoomTool
  • CircleTool, EllipseTool, FreeFormTool, MultiLineTool, ObliqueRectangleTool, PointTool, RectangleTool, SegmentTool
• Internal package reorganization: moved icons to plotpy/data/icons folder

New features:

• Added Echelon alpha function to the image parameters:
  • The Echelon alpha function is a step function, so that the alpha channel is 0 (full transparency) for the lowest value of the Lookup Table (LUT) and opaque (transparency level set by the Global alpha parameter) for the other values
  • This feature is added to the other existing alpha functions: Constant, Linear, Sigmoid, and Hyperbolic tangent

Bug fixes:

• Compatibility with PythonQwt 0.12.

Bug fixes:

• API breakage (unintentional) in V2.1.0:
  • In V2.1.0, mathutils.colormap module was renamed to mathutils.colormaps
  • Original mathutils.colormap module naming is restored in this release
• Colormap selection from the toolbar was not triggering the SIG_ITEMS_CHANGED signal (every time an item parameter is changed, the SIG_ITEMS_CHANGED signal has to be emitted by the BasePlot instance to notify the application that the plot has been modified)

In this release, test coverage is 71%.

New features:

• Curve-related features:
  • New tools.SelectPointsTool to select graphically multiple points on a plot
  • New tools.EditPointTool to edit graphically the position of a point on a plot
  • New downsampling feature:
    • The user may enable it to reduce the number of points displayed on a curve (e.g. when the curve is too dense)
    • The downsampling factor is adjustable (default to 10, i.e. 1 point out of 10 is displayed)
    • The feature is disabled by default
• Image-related features:
  • New "Colormap Manager":
    • Before this release, the colormap selection was limited to presets (e.g. "gray", "jet", etc.)
    • Now, the user can select a preset, edit it, or create a new one from scratch thanks to the new "Colormap Manager" dialog
  • New line cross section feature:
    • Before this release, the cross section feature was limited to either horizontal or vertical lines, or an average cross section withing a rectangle (aligned with the axes, or oblique)
    • Now, the user can draw a line cross section with the new "Line Cross Section" tool: the intensity profile associated to the drawn segment is displayed in a dedicated plot
• Added support for gestures:
  • Zooming in/out with the a two-finger pinch gesture
  • Panning with a two-finger drag gesture

Documentation:

• Reorganized some sections
• Added sections on new features

Bug fixes:

• Fixed critical bug in oblique cross section feature (regression introduced in 2.0.0)
• Removed dependency to pytest-qt for the test suite (due to Qt6 compatibility issues)

Bug fixes:

• Issue #9 - MacOS: error: a space is required between consecutive right angle brackets (use '> >')

Bug fixes:

• Issue #3 - PlotWidget: ZeroDivisionError on resize while ignoring constraints
• Issue #4 - Average cross section: RuntimeWarning: Mean of empty slice.
• Issue #5 - Contrast panel: levels histogram is sometimes not updated
• Issue #6 - 1D Histogram items are not properly drawn
• Issue #7 - Contrast panel: histogram may contains zeros periodically due to improper bin sizes
• Issue #8 - Contrast panel: switch back to default tool after selecting min/max

Bug fixes:

• Fixed plotpy.tools.AnnotatedEllipseTool: AttributeError when finalizing the shape
• plotpy.widgets.select_with_shape_tool: added missing toolbar and options parameters
• items.XRangeSelection is now serializable, as expected

Documentation:

• plotpy.plot.SyncPlotWindow: added missing documentation
• Added more information on PlotPyStack
• New "Motivation" section explaining the reasons behind the creation of PlotPy

This version is the first release of the 2.0 series, which is distributed under the BSD 3-Clause License.

PlotPy 2.0 is a major release that brings a lot of new features and bug fixes.

When initiated in 2016, PlotPy 1.0 was the result of the merge of two projects (as well as some other changes, e.g. a redesign of the API):

• guidata, a Python library generating graphical user interfaces for easy dataset editing and display
• guiqwt, a Python library providing efficient 2D data-plotting features (curve/image visualization and related tools) for interactive computing and signal/image processing application development

With PlotPy 2.0, the guidata code base has been reextracted: PlotPy now relies on guidata as a dependency, like before the merge.

PlotPy 2.0 also integrates all the bug fixes (>30) and new features that were added to guiqwt since the merge (i.e. between 2016 and 2023).

Supported versions of Python and Qt bindings have been updated:

• Python: 3.8, 3.9, and 3.10 (3.11 should work too, but will be officially supported when dropping support for Python 3.8, to keep a 3-year support period)
• Qt bindings: PyQt5 (even if PyQt6 and PySide6 are not officially supported, efforts have been made and will continue to be made to support them)

PlotPy 2.0 is a major release because it also brings a lot of new features:

• plot.PlotWidget, plot.PlotDialog, and plot.PlotWindow: API overhaul (simple, more consistent, more flexible, more extensible - see documentation for details)

• plot.SyncPlotWindow: new class to show multiple plots in a single window, in a synchronized way (zoom, pan, etc.)

• widgets.selectdialog.SelectDialog: a dialog box to select items using a shape tool (segment, rectangle or custom)

• Image lookup table (LUT):

  • Initially, the LUT alpha channel was either constant (input parameter alpha was a float between 0 and 1) or linearly dependent on the image pixel values (when the alpha_mask parameter was enabled)
  • Now, the LUT may be either constant (same as before) or dependent on the image pixel values but not only linearly: the LUT alpha channel may follow a linear, a sigmoid or an hyperbolic tangent function (see the new alpha_function parameter). The old alpha_mask parameter was removed

• Image pixels are now centered on their coordinates:

  • This means that the pixel at row i and column j is centered on the point (j, i) (before, the top-left corner of the pixel at row i and column j was centered on the point (j, i))
  • This convention is more consistent with the way images are displayed in other scientific image processing tools
  • This is one of the benefits of porting back guiqwt changes since the merge (i.e. between 2016 and 2023)

• New SVG-based shapes:

  • items.RectangleSVGShape: rectangle shape based on SVG data or file
  • items.SquareSVGShape: square shape based on SVG data or file
  • items.CircleSVGShape: circle shape based on SVG data or file

• builder.PlotBuilder:

  • Renamed PlotBuilder (originally guiqwt.builder.PlotItemBuilder)
  • Builder instance is still available using from plotpy.builder import make
  • Plot widget creation is now supported:
    • make.widget() for plot.PlotWidget
    • make.dialog() for plot.PlotDialog
    • make.window() for plot.PlotWindow
  • Added support for more plot items:
    • make.contours() for generating a list of items.ContourItem objects
    • make.annotated_point() for items.AnnotatedPoint
    • make.polygon() for items.PolygonShape
    • make.svg() for items.RectangleSVGShape, items.SquareSVGShape, and items.CircleSVGShape

• Added JSON serialization support for all plot items (curve, image, etc.)

• Brand new documentation, based on Sphinx with links to other projects API, examples and tutorials (e.g. on development related topics).

• Black code formatting on all Python files

• New automated test suite:

  • Automatic execution: --unattended command line option (Qt loop is bypassed)
  • Test suite based on pytest, supporting pytest-cov for coverage testing, pytest-xvfb for headless testing, and pytest-qt for Qt testing
  • Added support for Coverage: test coverage improved up to 70%

• Added typing annotations on (almost) all Python files

• Distribution: switched to pyproject.toml (still relying on setuptools and setup.py for building Cython/C++ extensions)

• Added code quality configuration files:

  • .pylintrc: pylint configuration file
  • .isort.cfg: isort configuration file
  • .coveragerc: coverage configuration file

• Added Visual Studio Code configuration files:

  • .vscode/settings.json: Python interpreter, code formatting, etc.
  • .vscode/tasks.json: build, test, etc.
  • .vscode/launch.json: run current file, run tests, etc.

PlotPy 2.0 also brings a lot of bug fixes and improvements:

• Handled all Cython/C++ extensions compilation warnings
• Fixed all NumPy deprecation issues (e.g. numpy.matrix)
• Fixed (annotated) circle/ellipse item creation/test
• Fixed all documentation build warnings
• Fixed regressions introduced by PlotPy V1 on top of guiqwt:
  • Global references for the Debian package management
  • Major aspect ratio issues:
    • When resizing the plot widget (images were downsized indefinitely)
    • When auto-scaling the plot widget (images were not displayed entirely)
  • TrImageItem rotation algorithm
  • Oblique cross-section test
  • About dialog, version informations
• Ported all guiqwt bug fixes since the merge (i.e. between 2016 and 2023):
  • Added support for Visual Studio 2015 and earlier
  • Speeding-up image alpha channel calculation
  • Optimized colormap icon caching
  • X-axis direction and auto-scale
  • Added load test (with a very large number of plot widgets)
  • Coordinates inversion in EllipseShape
  • ValueError with levels histogram
  • Various fixes regarding plot item creation, cross-section features, PyQt5 support, DICOM support, TIFF support, etc.
  • Etc.