Changelog¶

v0.12.3¶

(2023-09-01)

Enhancements¶

Resolved Future Warnings from pandas-2.1.0

v0.12.2¶

(2023-07-21)

This is a maintenance release that prevents newer versions of mizani from breaking plotnine. This version depends on mizani<0.10.0.

v0.12.1¶

(2023-05-09)

New Features¶

A layout manager. Now you do not have to adjust spacing parameters to prevent objects around the panels from overlapping. Specifically, you can:
1. Set the legend position to "top", "left" or "bottom"
2. Use a large or multiline plot title
3. Use a large or multiline plot caption
4. Use facet_wrap with scales="free" scales="free_x" or scales="free_y"
You can now also align the plot_title, axis_title_x, axis_title_y and plot_caption with respect to the panels. Set these to "left", "right" & "center" for the horizontal flowing text. And "top", "right" & "center" for the vertical flowing text.

Also, the size of the figure is exactly determined by the theme setting. For example, this:
```
theme(figure_size=(8, 6), dpi=100)
```
will create an 800px x 600px image.
You can create a subtitle using labs and style it using the plot_subtitle parameter to theme.
```
theme(plot_subtitle=element_text(size=8))
```

Enhancements¶

ggplot object gained a new method save_helper(). It gives you access to the matplotlib figure that will be saved to file.
When plotting with an ipython interactive backend (e.g. in a jupyter notebook). The default image output is retina. You do not need to run this command.
```
%config InlineBackend.figure_format = "retina"
```
Plotnine still respects any values set the user.
In an interactive setting, after drawing an image. The size of the figure (in pixels) is printed e.g. <Figure Size: (640 x 480)>. Previously, something like <ggplot: (336175301)> was printed.

API Changes¶

__add__() and __iadd__() now accept the same types of objects. In this change __iadd__() has gained the ability to accept a list of objects.
draw() no longer accepts the argument return_ggplot and the return value is always a matplolib figure.
Themeables strip_margin(), strip_margin_x() and strip_margin_y() have been renamed to strip_align(), strip_align_x() strip_align_y() repectively.
subplots_adjust() has been deprecated. You no longer need to use theme(subplots_adjust={"right": 0.85}) and the like to make space for the legend or text around the panels. In the future, this will through an error.
Changed default font-family (san-serif) from DejaVu Sans to Helvetica.

Bug Fixes¶

Fixed bug where a discrete position scale failed when mapping an empty variable. (#647)
Fixed bug where facet_grid with a datetime column run into an exception. (#629)
Fixed bug where geom_violin with facetting and "scales = free" did not work. (#655)
Fixed bug in position_dodge2 to work with preserve="single" for geom_rect and any other geoms that accept a min-max range along the x-axis.
Fixed legend for geom_point to accurately display transparent fill colors that have been manually set. (#665)
Fix issue where infinite limits for the x or y scales lead to an exception. (#664)
Fixed geom_text and geom_label when used with string formatting so that missing values are removed. (#651)

v0.10.1¶

(2022-09-10)

This is a re-release of v0.10.0.

In the previous release the project description fileformat changed and not all references to it were corrected. PYPI read file that did not exist, got no description. This release is fixes error.

v0.10.0¶

(2022-09-10)

Bug Fixes¶

Fixed bug in stat_bin_2d where bins whose edges are zero or negative would yield the wrong plot. (#604)
Fixed bug stat_density_2d which had been broken by Matplotlib 1.6.0 (#619)
Fixed geom_text when using adjust_text for the arrows to have the same color as the text.

New Features¶

geom_ribbon acquired new parameter where which can be used shade withing given limits .
Added support for objects with a to_pandas method to convert them to a pandas dataframe. You can now pass in polars dataframes.
geom_ribbon acquired new parameter outline_type which controls how to stroke the boundaries of the shaded area. This parameter is also available to geom_area and geom_density. #605

API Changes¶

The automatic groups generated by the system and depending on the aesthetic mappings now only depend those aesthetics that have scales. This is more strict than has been, should affect very few cases of code that has been written and it should be less suprising to the users.
geom_area and geom_density now default to showing only the upper outline/boundary of the area. Use outline_type="full" for the old behaviour.

v0.9.0¶

(2022-07-01)

API Changes¶

For the calls to ggplot, the order of the positional arguments is (data, mapping) and for calls geoms and stats the order is (mapping, data). In practice, you can still provide one/two positional arguments in any order. However, the are a few cases where old code will breaks due to duplicate parameters.

New Features¶

Added geom_pointdensity, a 2D density plot that does not hide the distribution of the individual data points.
position_dodge2 gained the ability to reverse the order of the plotted groups. (#515)
Added theme_tufte, a theme that reduces the distractions around plot data.
labs can now be used to create a caption. This caption can be styled using the plot_caption parameter to theme.
```
theme(plot_caption=element_text(size=13))
```
You can now align the plot_title horizontally to the left or right.
```
theme(plot_title=element_text(ha='left'))
```
Before, the title was always centered.

Bug Fixes¶

Fixed bug in facet_grid and facet_wrap where the drop parameter was ignored. (#510)
Fixed issue with annotation_logticks where the wrong range was used for the x|y axis. (#525)
Fixed bug in geom_smooth where using a coordinate transformation without computing the standard error led to a crash. (#526)
Fixed bug where you could not specific the cap/end of a dashed line when styling a plot with element_line.
Fixed legend for geom_map so that the entries have the same stroke and points on the map. (#534)
Fixed geom_history for some cases when it fails to automatically determine the number of bins. (#543)
Fixed facet_grid and facet_wrap when mapping to lists/arrays/series so that the rows mapped this way are not mixed up. (#548)
Fixed stat_bin_2d not to error for matplotlib >= 3.5.0 so that it does not throw a Numpy error. (#567)
Fixed geom_map to be able to draw Polygons with holes in them.
Fixed geom_map to be able to draw MultiPolygons.
Fixed scale_x_discrete and scale_x_discrete to work properly with None values. (#523)
Fixed geom_point to respect not to use the fill mapping on unfilled shapes. (#100)
Fixed position_stack to work with scales that have non-linear transformations. e.g. scale_y_log10. (#580)
Fixed scales to work with limits passed as a function.
Fixed axis_ticks so you can change the colour of the axis ticks. This was probably a regression due to changes in Matplotlib.
Fixed panel_ontop so that if is set to False, then the data visualised below the grid-lines. (#585)
Fixed geom_ribbon, geom_area and geom_density to work with coord_flip. (#596)

Enhancements¶

stat_summary now has a default summary function. (#528)
stat_density now uses a fast computation path for gaussian densities with no weights. This was probably a regression.
geom_map can now draw MultiPoints, MultiLineStrings and MultiPolgons.
geom_text with adjust_text can now be used in more than one layer with good enough results.

v0.8.0¶

(2021-03-24)

API Changes¶

How you map to calculated aesthetics has changed. Use the after_stat() function. The old methods 'stat(name)' and '..name..' have been deprecated.

New Features¶

You can now map to aesthetics at three different stages. See aes, after_stat(), after_scale() and stage.
geom_violin gained the a new parameter style with which you can draw half violin (density curve on one side and flat on the other).
Added geom_raster.
geoms gained new parameter raster for the Layer. You can use it to rasterize any layer when the resulting plot is of vector format e.g. pdf.
Using the space parameter, facet_grid gained the ability to have rows and columns of panels of different sizes.

Bug Fixes¶

Fixed issue where some plots with a colorbar would fail for specific themes. (#424)
Fixed geom_map to plot MultiLineString geom types.
Fixed geom_text to allow any order of mapping and data positional arguments.
Fixed bug were the plotted image may have ignored theming that relied on some Matplotlib rcParams. (#451)
Fixed the weight aesthetic in geom_boxplot, previously ignored it is now recognised. (#438)
Fixed annotation_logticks and annotation_stripes to work without global data and aesthetics. (#469)
Fix scale_shape_discrete when print many unfilled shapes not to assign the same shapes to more than one group. (#473)
Fixed bug in stat_ellipse where the center of the ellipse assuming a multivariate t-distribution was incorrectly calculated. (#493)
Fixed calculation of ndensity in stat_bin. (#494)

Enhancements¶

Manual scales now match the values of the breaks if the breaks are given. (#445)
Using print to show a ggplot object will not show the hash (<ggplot: ...>) anymore. There is now a difference between repr(p) and str(p). (#453)
Added option to for the base_family of a theme, now you can set it once with and have it be applied to all themes. (#436)
```
from plotnine.options import set_option
set_option('base_family', 'Comic Sans MS')
```
You can now add None to a ggplot, doing so returns a copy of the the ggplot object. (#474)
Better handling of multiline facet labels. (#484)

v0.7.1¶

(2020-08-05)

Bug Fixes¶

Fixed issue where a plot has no data and the geoms have no data, but the mappings are valid. (#404)
Fixed preserve='single' in plotnine.positions.position_dodge and plotnine.positions.position_dodge2 to work for geoms that only have x aesthetic and not xmin and xmax e.g plotnine.geoms.geom_text.
Fix regression in v0.7.0 where plots with a colorbar would fail if using theme_matplotlib.

v0.7.0¶

(2020-06-05)

API Changes¶

Changed the default method of caculating bandwidth for all stats that use kernel density estimation. The affected stats are stat_density, stat_ydensity, and stat_sina. These stats can now work with groups that have a single unique value.
Changed plotnine.scale.scale_colour_continuous to refer to the same scale as plotnine.scale.scale_color_continuous.
Changed plotnine.scale.scale_color_cmap so the parameter cmap_name refers to the name of the color palette and name refers to the name of the scale. (#371)

New Features¶

aes got an internal function reorder which makes it easy to change the ordering of a discrete variable according to some other variable/column.
stat_smooth can now use formulae for linear models.

Bug Fixes¶

Fixed issue where a wrong warning could be issued about changing the transform of a specialised scale. It mostly affected the timedelta scale.
Fixed plotnine.geoms.geom_violin and other geoms when used with position='dodge' not to crash when if a layer has an empty group of data.
Fixed bug in plotnine.geoms.geom_path for some cases when groups had less than 2 points. (#319)
Fixed all stats that compute kernel density estimates to work when all the data points are the same. (#317)
Fixed issue where setting the group to a string value i.e. group='string' outside aes() failed due to an error.
Fixed issue where discrete position scales could not deal with fewer limits than those present in the data. (#342)
Fixed issue with using custom tuple linetypes~ with plotnine.scales.scale_linetype_manual. (#352)
Fixed plotnine.geoms.geom_map to work with facets. (#359)
Fixed plotnine.position.jitter_dodge to work when color is used as an aesthetic. (#372)
Fixed plotnine.geoms.geom_qq to work with facets (#379)
Fixed skewed head in plotnine.geoms.arrow when drawn on facetted plot (#388)
Fixed issue with plotnine.stats.stat_density where weights could not be used with a gaussian model. (#392)
Fixed bug where guide_colorbar width and height could not be controlled by legend_key_width and legend_key_height. (#360)

Enhancements¶

You can now set the bandwidth parameter bw of stat_ydensity.
Parameters ha and va of geom_text have been converted to aesthetics. You can now map to them. (#325)
All themes (except theme_matplotlib) now do not show minor ticks. (#348)

v0.6.0¶

(2019-08-21)

API Changes¶

The draw parameter of plotnine.geoms.geom_map has been removed. Shapefiles should contain only one type of geometry and that is the geometry that is drawn.
Ordinal (Ordered categorical) columns are now mapped to ordinal scales. This creates different plots.
The default mapping for the computed aesthetic size of stat_sum has changed to 'stat(n)'. This also changes the default plot for geom_count.

New Features¶

geom_text gained the adjust_text parameter, and can now repel text.
Added annotation_logticks.
Added geom_sina
Added scales for ordinal (ordered categorical) columns.
geom_step gained the option mid for the direction parameter. The steps are taken mid-way between adjacent x values.
Added annotation_stripes.

Bug Fixes¶

Fixed bug where facetting would fail if done on a plot with annotation(s) and one of the facetting columns was also a variable in the environment.
Fixed bug where coord_flip would not flip geoms created by geom_rug (#216).
Fixed bug where plots with theme_xkcd cannot be saved twice (#199)
Fixed bug that made it impossible to map to columns with the same name as a calculated columns of the stat. (#234)
Fixed bug in geom_smooth that made it difficult to use it with stats other than stat_smooth. (#242)
Fixed bug in position_dodge where by bar plot could get thinner when facetting and useing preserve = 'single'. (#224)
Fixed bug in coord_trans where if the transformation reversed the original limits, the order in which the data was laid out remained unchanged. (#253)
Fixed bug in stat_count where float weights were rounded and lead to a wrong plot. (#260)
Fixed bug where one could not use the British spelling colour to rename a color scale. (#264)
Fixed bug in lims, xlim, and ylim where datetime and timedelta limits resulted in an error.
Fixed bug where geom_rect could not be used with coord_trans. (#256)
Fixed bug where using free scales with facetting and flipping the coordinate axes could give unexpected results. (#286)
Fixed unwanted tick along the axis for versions of Matplotlib >= 3.1.0.
Fixed geom_text not to error when using hjust and vjust. (#287)
Fixed bug where geom_abline geom_hline and geom_vline could give wrong results when used with coord_trans.
Fixed bug where layers with only infinite values would lead to an exception if they were the first layer encountered when choosing a scale.

Enhancements¶

Legends are now plotted in a predictable order which dedends on how the plot is constructed.
The spokes drawn by geom_spoke can now have a fixed angle.
Aesthetics that share a scale (e.g. color and fill can have the same scale) get different guides if mapped to different columns.
When the transform of a specialised (one that is not and identity scale) continuous scale is altered, the user is warned about a possible error in what they expect. (#254, #255)
The method_args parameter in stat_smooth can now differentiate between arguments for initialising and those for fitting the smoothing model.
position_nudge can now deal with more geoms e.g. geom_boxplot.
The limits parameter of scale_x_discrete and
scale_y_discrete can now be a function.
The width of the boxplot can now be set irrespective of the stat.
The mid-point color of scale_color_distiller now matches that of the trainned data.
The way in which layers are created has been refactored to give packages that that extend plotnine more flexibility in manipulating the layers.
You can now specify one sided limits for coordinates. e.g. coord_cartesian(limits=(None, 10)).
All the themeables have been lifted into the definition of theme so they can be suggested autocomplete.

v0.5.1¶

(2018-10-17)

Bug Fixes¶

Changed the dependency for mizani to v0.5.2. This fixes an issue where facetting may create plots with missing items. (#210)

v0.5.0¶

(2018-10-16)

API Changes¶

Plotnine 0.5.0 only supports Python 3.5 and higher
geopandas has been removed as a requirement for installation. Users of geom_map will have to install it separately. (#178)

Bug Fixes¶

Fixed issue where with the subplots_adjust themeable could not be used to set the wspace and hspace Matplotlib subplot parameters. (#185)
Fixed in stat_bin where setting custom limits for the scale leads to an error. (#189)
Fixed issue interactive plots where the x & y coordinates of the mouse do not show. (#187)
Fixed bug in geom_abline where passing the mapping as a keyword parameter lead to a wrong plot. (#196)
Fixed issue where minor_breaks for tranformed scaled would have to be given in the transformed coordinates. Know they are given the data coordinates just like the major breaks.

Enhancements¶

For all geoms, with coord_cartesian float('inf') or np.inf are interpreted as the boundary of the plot panel.
Discrete scales now show missing data (None and nan). This behaviour is controlled by the new na_translate option.
The minor_breaks parameter for continuous scales can now be given as an integer. An integer is taken to controll the number of minor breaks between any set of major breaks.

v0.4.0¶

2018-01-08

API Changes¶

Calculated aesthetics are accessed using the stat() function. The old method (double dots ..name..) still works.
stat_qq calculates slightly different points for the theoretical quantiles.
The scales (when set to free, free_x or free_y') parameter of facet_grid and facet_wrap assigns the same scale across the rows and columns.

New Features¶

Added geom_qq_line and stat_qq_line, for lines through Q-Q plots.
Added geom_density_2d and geom_stat_2d.
Added stat_ellipse.
Added geom_map.
Plotnine learned to respect plydata groups.
Added stat_hull.
Added save_as_pdf_pages().

Bug Fixes¶

Fixed issue where colorbars may chop off the colors at the limits of a scale.
Fixed issue with creating fixed mappings to datetime and timedelta type values.(#88)
Fixed scale_x_datetime and scale_y_datetime to handle the intercepts along the axes (#97).
Fixed stat_bin and stat_bin_2d to properly handle the breaks parameter when used with a transforming scale.
Fixed issue with x and y scales where the name of the scale was ignored when determining the axis titles. Now, the name parameter is specified, it is used as the title. (#105)
Fixed bug in discrete scales where a column could not be mapped to integer values. (#108)
Make it possible to hide the legend with theme(legend_position='none'). (#119)
Fixed issue in stat_summary_bin where some input values gave an error. (#123)
Fixed geom_ribbon to sort data before plotting. (#127)
Fixed IndexError in facet_grid when row/column variable has 1 unique value. (#129)
Fixed facet_grid when scale='free', scale='free_x' or scale='free_y', the panels share axes along the row or column.
Fixed geom_boxplot so that user can create a boxplot by specifying all required aesthetics. (#136)
Fixed geom_violin to work when some groups are empty. (#131)
Fixed continuous scales to accept minor=None (#120)
Fixed bug for discrete position scales, where drop=False did not drop unused categories (#139)
Fixed bug in stat_ydensity that caused an exception when a panel had no data. (#147)
Fixed bug in coord_trans where coordinate transformation and facetting could fail with a KeyError. (#151)
Fixed bug that lead to a TypeError when aesthetic mappings to could be recognised as being groupable. It was easy to stumble on this bug when using geom_density. (#165)
Fixed bug in facet_wrap where some combination of parameters lead to unexpected panel arrangements. (#163)
Fixed bug where the legend text of colorbars could not be themed. (#171)

v0.3.0¶

(2017-11-08)

API Changes¶

geom_smooth gained an extra parameter legend_fill_ratio that control the area of the legend that is filled to indicate confidence intervals. (#32)
plotnine.ggplot.save() gained an extra parameter verbose. It no longer guesses when to print information and when not to.
plotnine.ggplot.draw() gained an extra parameter return_ggplot.
If the minor_breaks parameter of scales is a callable, it now expects one argument, the limits. Previously it accepted breaks and limits.

New Features¶

Added PlotnineAnimation for animations.
Added watermark for watermarks.
Added datetime scales for alpha, colour, fill and size aesthetics

Enhancements¶

Changed parameter settings for stat_smooth.
1. Default span=0.75 instead of 2/3
2. When using loess smoothing, the control parameter surface is only set to the value 'direct' if predictions will be made outside the data range.
Better control of scale limits. You can now specify individual limits of a scale.
```
scale_y_continuous(limits=(0, None))
xlim(None, 100)
```
You can also use expand_limits()
Low and high scale limits can now be expanded separately with different factors multiplicative and additive factors.
The layer parameter show_legend can now accept a dict for finer grained control of which aesthetics to exclude in the legend.
Infinite values are removed before statistical computations stats (#40).

stats also gained new parameter na_rm, that controls whether missing values are removed before statistical computations.
qplot() can now use the name and a Pandas series to label the scales of the aesthetics.
You can now put stuff to add to a ggplot object into a list and add that that instead. No need to wrap the list around the internal class Layers.
```
lst = [geom_point(), geom_line()]
g = ggplot(df, aes('x', 'y'))
print(g + lst)
```
Using a list allows you to bundle up objects. It can be convenient when creating some complicated plots. See the Periodic Table Example.
You can now use a dict (with manual scales) to map data values to aesthetics (#169).
You can now specify infinite coordinates with plotnine.geoms.geom_rect (#166)

Bug Fixes¶

Fixed bug where facetting led to a reordering of the data. This would manifest as a bug for geoms where order was important. (#26)
Fix bug where facetting by a column whose name (eg. class) is a python keyword resulted in an exception. (#28)
Fix bug where y-axis scaling was calculated from the xlim argument.
Fix bug where initialising geoms from stats, and positions from geoms, when passed as classes (e.g. stat_smooth(geom=geom_point), would fail.
Fixed bug in plotnine.ggplot.save() where specifying the width and height would mess up the strip_text and spacing for the facetted plots. (#44).
Fixed bug in geom_abline, geom_hline and geom_vline where facetting on a column that is not mapped to an aesthetic fails. (#48)
Fixed bug in geom_text, the fontstyle parameter was being ignored.
Fixed bug where boolean data was mapped to the same value on the coordinate axis. (#57)
Fixed bug in facet_grid where the scales sometimes has no effect. (#58)
Fixed bug in stat_boxplot where setting the width parameter caused an exception.

v0.2.1¶

(2017-06-22)

Fixed bug where manually setting the aesthetic fill=None or fill='None' could lead to a black fill instead of an empty fill.
Fixed bug where computed aesthetics could not be used in larger statements. (#7)
Fixed bug in stat_summary where the you got an exception for some types of the x aesthetic values.
Fixed bug where ggplot(data=df) resulted in an exception.
Fixed missing axis ticks and labels for facet_wrap when the scales are allowed to vary (e.g scales='free') between the panels.
Fixed bug in stat_density where changing the x limits lead to an exception (#22)

v0.2.0¶

(2017-05-18)

Fixed bug in scale_x_discrete and scale_y_discrete where if they were instantiated with parameter limits that is either a numpy array or a pandas series, plotting would fail with a ValueError.
Fixed exceptions when using pandas.pivot_table() for Pandas v0.20.0. The API was fixed.
Fixed issues where lines/paths with segments that all belonged in the same group had joins that in some cases were "butted".

API Changes¶

geom_text now uses ha and va as parameter names for the horizontal and vertical alignment. This is what matplotlib users expect. The previous names hjust and vjust are silently accepted.
Layers() can now be used to bundle up geoms and stats. This makes it easy to reuse geoms and stats or organise them in sensible bundles when making complex plots.

v0.1.0¶

(2017-04-25)

First public release