v4.9.0

[4.9.0] - 2026-03-11

Added

• Python 3.14 support.

• Python 3.14 free-threading support [#1454].

• Plot Annotations:

  • New labels parameter in geom_smooth() designed to display statistical summaries of the fitted model directly on the plot.
    This parameter accepts a smooth_labels() object, which provides access to model-specific variables like $R^2$, the regression equation and others.

    See: example notebook.

  • Plot tags. A tag can be specified via labs(tag=...) and styled using theme parameters [#1407].

    See: example notebook and updated plot layout scheme.

  • Plot tags customization parameters in theme():

    • plot_tag - sets the tag style via element_text().
    • plot_tag_location - specifies the area used for positioning the tag.
    • plot_tag_position - specifies the position of the tag within the selected area.
    • plot_tag_prefix - text added before the tag value.
    • plot_tag_suffix - text added after the tag value.

    See: example notebook.

• Geometries:

  • New geom_bracket(), geom_bracket_dodge() [#1114].

    See: example notebook.

  • geom_imshow():

    • Support for custom colormaps [#780].

    • New cguide parameter: use to customize the colorbar for grayscale images.

      See: example notebook.

• Color Scales:

  • New palette() method for color scales: generates a list of hex color codes that can be used with scale_color_manual() to maintain consistent colors across multiple plots [#1444].

    See: example notebook.

  • New overflow parameter in scale_color_brewer() / scale_fill_brewer(): controls how colors are generated when more colors are needed than the palette provides.
    Options: 'interpolate' ('i'), 'cycle' ('c'), 'generate' ('g').

    See: example notebook.

• Positional Scales:

  • New break_width parameter that specifies a fixed distance between axis breaks.

    See examples:

    • datetime scale
    • time (duration) scale
    • log10 scale

  • Support for axis minor ticks via axis_minor_ticks and axis_minor_ticks_length parameters in theme() [#1379].

    See: example notebook.

• gggrid(): interactive pan/zoom now propagates across subplots with shared axes (sharex/sharey) [#1413].

  See: example notebook.

Changed

• [BREAKING]: ColorBrewer palettes: changed default behavior when the requested number of colors exceeds the palette's maximum size.
  Now defaults to 'interpolate' for sequential/diverging palettes and 'generate' for qualitative palettes.
  Previously, depending on the palette type, this either resulted in duplicate colors or random additional colors.
  Use the new overflow parameter to explicitly control this behavior.

• Reduced import overhead by nearly two orders of magnitude (from ~1.2s down to ~0.02s) [#1469].

• Missing values in geom_area_ridges() create gaps in geometries instead of being interpolated over.

• Discrete color scales (Brewer, Manual) now produce a colorbar guide when used with continuous data.
  Previously they produced a legend guide regardless of the data type.

• Changes affecting users on the JVM platform:

  • Upgraded the Kotlin version to 2.2.20 (was 1.9.25).
  • New artifact for JVM Swing applications: org.jetbrains.lets-plot:lets-plot-swing.
    This artifact provides the SwingPlotPanel class, which can be used to display plots in Swing applications instead of the now-obsolete DefaultPlotPanelBatik.
    For details, see the jvm-swing-app example in the "lets-plot-mini-apps" repository.
  • [BREAKING]: Removed JavaFX artifacts.
    The org.jetbrains.lets-plot:lets-plot-jfx artifact is no longer available.
    Replace it with new org.jetbrains.lets-plot:lets-plot-swing dependency and use SwingPlotPanel instead of DefaultPlotPanelJfx.
    For details, see the jvm-javafx-app example in the "lets-plot-mini-apps" repository.
  • [BREAKING]: Removed plot-image-export module.
    The org.jetbrains.lets-plot:lets-plot-image-export artifact is no longer available.
    The PlotImageExport utility has been moved to the platf-awt module.
    The required org.jetbrains.lets-plot:platf-awt dependency is likely already present in your project.

Fixed

• Drop commons-io dependency [#1421].
• Unexpected replacement of double curly brackets with a single curly bracket [#1433].
• Upgrade to a newer version of ws [#1150].
• geom_imshow: unclear error message when mixing transparencies [#1088].
• geom_imshow and scale_y_reverse [#1210].
• Nice to be able to get a list of colors from a color scale object [#1444].
• Allow tooltips param to accept list [#1455].
• Allow grouped tooltips for plots with multiple univariate geoms [#1460].
• Fixed a regression in geom_text_repel() / geom_label_repel(): points with empty labels were incorrectly skipped
  before building the repel obstacle set, so they were not included in collision avoidance and labels could overlap
  dense point clusters.

v4.9.0rc1

Updated version v4.9.0rc1

v4.8.2

[4.8.2] - 2026-12-19

Fixed

• Hyperlinks support for lets-plot-compose.
• Colab no longer rendering lets-plot with empty setup HTML [#1436]

v4.8.2rc1

Updated version v4.8.2rc1

v4.8.1

[4.8.1] - 2025-12-01

Fixed

• saving plots as png takes a very long time [#1423].
• stat_ecdf() takes a very long time for even moderately sized datasets [#1424].
• lets-plot plots are not displayed in plots pane [PST-8170].
• inconsistencies in theme/flavor inheritance in gggrid() subplots.

v4.8.1rc1

Updated version v4.8.1rc1

v4.8.0

[4.8.0] - 2025-11-05

Added

• Geometries:

  • geom_pointdensity() [#1370].

    See: example notebook.

  • Geoms with 1-to-1 statistics (such as geom_qq(), geom_sina()) preserve the mapping to original data after statistical transformation.

    See: example notebook.

  • geom_histogram(): custom bin bounds (parameter breaks) [#1382].

    See: example notebook.

• Plot Layout:

  • The legend automatically wraps to prevent overlap - up to 15 rows for vertical legends and 5 columns for horizontal ones [#1235].

    See: example notebook.

  • gggrid(): support for shared legends (parameter guides).

    See: example notebook.

• Plot Theme:

  • flavor_standard() sets the theme's default color scheme [#1277].
    
    Use to override other flavors or make defaults explicit.

    See: example notebook.

  • theme_gray() as an alias for theme_grey().

  • legend_justification parameter of theme() accepts additional string values: 'left', 'right', 'top', and 'bottom'.

    See: example notebook.

  • Support for inward axis ticks.

    See: example notebook.

• Markdown:

  • Support for target attribute for links.
  • Links now open in a new tab by default [#1397].

• ggtb(): size_zoomin and size_basis parameters for geometry scaling [#1369].

  See: example notebook.

Changed

• [BREAKING] Explicit group aesthetic now overrides default grouping behavior instead of combining with it [#1401].

  See: example notebook.

Important

Previously, setting group='variable' would group by both the explicit variable AND any discrete
aesthetics (color, shape, etc.).
Now it groups ONLY by the explicit variable, matching ggplot2 behavior.
Use group=[var1, var2, ...] to group by multiple variables explicitly,
and group=[] to disable any grouping.

• Missing values in geom_line(), geom_path(), geom_ribbon(), and geom_area() create gaps in geometries instead of being interpolated over [#818], [#1406].

  See: example notebook.

• theme: the exponent_format default value changed to 'pow' - superscript powers of 10 (was e-notation).

• The multi-layer line plot now shows tooltips for each series simultaneously, in the same way that a single-layer plot with color mapped to series does.

Fixed

• geom_pie on geospatioal plot with map_join failes to render without explicit group aesthetic.
• geom_density2d: NullPointerException when weight aesthetic contains None values [#1399].
• Tooltip shows duplicate lines when as_discrete is applied twice to the same var [#1400].
• geom_sina: incorrect shape in legend [#1403].
• geom_density2d: Incorrect processing of weighted statistics when None value occurs in the x or y column.
• facet_wrap: indescriptive error when the specified facet variable is not present in the dataset [#1409].
• Integer numbers in facet strip titles are displayed as float [#1386].
• Error when using scale_identity(aesthetic="shape") [#1212].
• ggsave: theme option face="italic" doesn't work [#1391].
• Fail early if string format is incorrect [#1410].

v4.8.0rc1

Updated version v4.8.0rc1

v4.7.3

[4.7.3] - 2025-09-12

Changed

• ggsave(): Large plot dimensions without units now require explicit unit specification.
  
  When plot size exceeds 20 without specifying units (e.g., ggsave(p, 300, 400)),
  
  we ask to specify units explicitly:
  
  ggsave(p, 300, 400, unit='px') or ggsave(p, 3, 4, unit='in').

Fixed

• ggtb() support in Swing/Batik frontend.
• Multiline support for axis labels in polar coordinates.
• When the plot size in ggsave() is specified in pixels, dpi now affects
  
  only the physical size, not the pixel dimensions as before.
• Blocking SwingUtilities.invokeAndWait() call in plot image export (AWT backend)

v4.7.2

[4.7.2] - 2025-09-01

Added

• Plot Layout:

  • New strip_spacing, strip_spacing_x, and strip_spacing_y parameters in theme() to control spacing between the facet strip (title bar) and the plot panel.
  • New panel_spacing, panel_spacing_x, and panel_spacing_y parameters in theme() to control spacing between plot panels in faceted plots, [#1380].

  See: example notebook.

• Image resolution is saved in the metadata of PNG files created with ggsave().

Changed

• plot-image-export: switched from Batik to Graphics2D API

Fixed

• ggsave(): memory leak when using geom_raster().
• Incorrect physical image size when exporting PDF with ggsave() without specifying dpi.