Skip to content

v0.13.0
Compare
Choose a tag to compare
@has2k1 has2k1 released this 20 Feb 15:33
· 82 commits to main since this release
v0.13.0
This tag was signed with the committer’s verified signature.
has2k1 Hassan Kibirige
GPG key ID: BF6C6B75366DBE25
Learn about vigilant mode.
e18acd4

API Changes

  • Requires python >= 3.9

  • Using the print or repr functions to draw and show the plot has been deprecated. Use ggplot.show().

  • The name of the calculated aesthetic of stat_function changed from y to fx.

  • stat_ecdf has gained the pad parameter. The default is set to True, which pads the domain with -inf and inf so that the ECDF does not have discontinuities at the extremes. To get the behaviour, set pad to False. (#725)

  • Removed the environment parameter from ggplot.

  • When a ggplot object is the last in a jupyter cell, the output image will not be followed by string meta information about the figure/image.

    This will happen even if the backend is set to an interactive one.

    If you set the backend to an interactive one, use show to draw the plot.

  • The default horizontal alignment for the plot title is center if it there is no subtitle. When there is a subtitle, the default is to have both aligned to the left.

  • Some parameters that control the look and feel of guide_colorbar and guide_legend have been removed. For their place, the theme parameter has been introduced and it gives better control of the look and feel.

  • Themeables legend_entry_spacing, legend_entry_spacing_x and legend_entry_spacing_y have been renamed to legend_key_spacing, legend_key_spacing_x and legend_key_spacing_y respectively

  • facet_grid now accepts two parameters, rows and cols, to specify the variables along the two dimensions of the panels.

    The previous way of using a single parameter will still work if it is a string. For cases where the value was a list, e.g.

    facet_grid(facets=["col1", "col2"])

    can be rewritten as any one of;

    facet_grid("col1", "col2")
facet_grid(["col1"], ["col2"])
facet_grid(rows="col1", cols="col2")
facet_grid(rows=["col1"], cols=["col2"])

  • The facets parameter in facet_wrap has also changed to a more straight forward specification for the column variables. It expects a single string or a list/tuple of strings. (#545)

    However, the R-style formula strings are still silently accepted.

New

  • Added symmetric logarithm transformation scales scale_x_symlog and scale_y_symlog

  • Gained themeables

    to set the plot margin on each side independently.

  • Gained themeables

    to control the x & y axis ticks length.

  • Gained themeables

    to control the x & y tick padding.

  • Some parameters in element_text can now accept lists/tuples to set the values on individual text objects. (#724)

  • Gained the option figure_format to set the format of the inline figures in an interactive session.
    e.g.

    from plotnine.options import set_option

set_option("figure_format", "svg")

    will output all subsequent figures in svg format.

  • Improved support for customizing guides/legends.

    • You can now add a frame to the colobar

    • You can now apply themes to individual guides, e.g.

      + guides(color=guide_colorbar(theme=theme_xkcd()))

      or

      + guides(color=guide_legend(theme=theme_minimal())

    • You can now place legends at more than one position around the panels. e.g.

      + guides(
    color=guide_colorbar(position="left"),
    fill=guide_legend(position="bottom"),
    size=guide_legend(position="bottom")
)

      Puts the fill and size guides at the bottom, and the color guide on the left.

    • You can easily justify the legend along the four edges of the space around the panel area.

    • You can now place the legend inside the panels with easily control its location.

  • The space parameter of facet_grid now responds to the values free, free_x and free_y to have panels whose relative width and/or height depends on the data range. (#545)

Bug Fixes

  • Fixed handling of minor breaks in scale_continuous to accept numpy arrays and when the scale has a transform, that the minor breaks are supplied in user space and not transform space. Just like the major breaks. (#685)

  • Fixed theming of axis_ticks with the size parameter. (#703)

  • Fixed space handling around axis_label, axis_text and axis_ticks when the ticks are turned off.

  • Fixed bug in geom_path where the lineend parameter was ignored. (#727)

  • Fixed bug where theme(legend_background=element_blank()) messed up the position of the legend, instead of only removing the background.

  • Fixed using facet_grid with a column named key. (#734)

  • Fixed using legend when using an identity scale and reordering the breaks. (#735)

  • Fixed drawing the upper outline of geom_ribbon. (#728)

  • Fixed issue where the gridlines overlap the panel border. (#638)

Enhancements

  • All __all__ variables are explicitly assigned to help static typecheckers infer module attributes. (#685)

  • You can now pickle the drawn matplotlib figures. (#729)

  • The facet titles in the strip_text are now better aligned are appear more centered withing the strip_background.

Assets 2