v0.60 not compatible with seaborn v0.13

[xinwei-sher]

When using statannotations v0.60 with seaborn v0.13, it would fail with this error:
AttributeError: module 'seaborn.categorical' has no attribute '_Violin(Box...)Plotter'

It only works with seaborn v0.11.
Any fix for this would be very much appreciated!

[moqri]

Similar error here:

AttributeError: module 'seaborn.categorical' has no attribute '_BoxPlotter'

Thanks for the updates

[pfluec]

Same issue here:

AttributeError: module 'seaborn.categorical' has no attribute '_BoxPlotter'

[wenyuhaokikika]

Same issue here:

code

from statannotations.Annotator import Annotator
x = "color"
y = "price"
hue = "cut"
hue_order=['Ideal', 'Premium', 'Good', 'Very Good', 'Fair']
order = ["E", "I", "J"]
pairs=[
    (("E", "Ideal"), ("E", "Very Good")),
    (("E", "Ideal"), ("E", "Premium")),
    (("E", "Ideal"), ("E", "Good")),
    (("I", "Ideal"), ("I", "Premium")),
    (("I", "Ideal"), ("I", "Good")),
    (("J", "Ideal"), ("J", "Premium")),
    (("J", "Ideal"), ("J", "Good")),
    (("E", "Good"), ("I", "Ideal")),
    (("I", "Premium"), ("J", "Ideal")),
    ]
ax = sns.boxplot(data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)
annot = Annotator(ax, pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)
annot.configure(test='Mann-Whitney', verbose=2)
annot.apply_test()
annot.annotate()
plt.legend(loc='upper left', bbox_to_anchor=(1.03, 1))
plt.show()

raise Exception

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[282], line 19
      7 pairs=[
      8     (("E", "Ideal"), ("E", "Very Good")),
      9     (("E", "Ideal"), ("E", "Premium")),
   (...)
     16     (("I", "Premium"), ("J", "Ideal")),
     17     ]
     18 ax = sns.boxplot(data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)
---> 19 annot = Annotator(ax, pairs, data=df, x=x, y=y, order=order, hue=hue, hue_order=hue_order)
     20 annot.configure(test='Mann-Whitney', verbose=2)
     21 annot.apply_test()

File ~/anaconda3/envs/pyG/lib/python3.8/site-packages/statannotations/Annotator.py:108, in Annotator.__init__(self, ax, pairs, plot, data, x, y, hue, order, hue_order, engine, verbose, **plot_params)
    106     self._plotter = None
    107 else:
--> 108     self._plotter = self._get_plotter(engine, ax, pairs, plot, data,
    109                                       x, y, hue, order, hue_order,
    110                                       verbose=verbose, **plot_params)
    112 self._test = None
    113 self.perform_stat_test = None

File ~/anaconda3/envs/pyG/lib/python3.8/site-packages/statannotations/Annotator.py:784, in Annotator._get_plotter(engine, *args, **kwargs)
    782 if engine_plotter is None:
    783     raise NotImplementedError(f"{engine} engine not implemented.")
--> 784 return engine_plotter(*args, **kwargs)

File ~/anaconda3/envs/pyG/lib/python3.8/site-packages/statannotations/_Plotter.py:87, in _SeabornPlotter.__init__(self, ax, pairs, plot, data, x, y, hue, order, hue_order, verbose, **plot_params)
     85 self.check_plot_is_implemented(plot)
     86 self.plot = plot
---> 87 self.plotter = self._get_plotter(plot, x, y, hue, data, order,
     88                                  hue_order, **plot_params)
     90 self.group_names, self.labels = self._get_group_names_and_labels()
     91 self.groups_positions = _GroupsPositions(self.plotter,
     92                                          self.group_names)

File ~/anaconda3/envs/pyG/lib/python3.8/site-packages/statannotations/_Plotter.py:119, in _SeabornPlotter._get_plotter(self, plot, x, y, hue, data, order, hue_order, **plot_params)
    116 self.fix_and_warn(dodge, hue, plot)
    118 if plot == 'boxplot':
--> 119     plotter = sns.categorical._BoxPlotter(
    120 
    121         x, y, hue, data, order, hue_order,
    122         orient=plot_params.get("orient"),
    123         width=plot_params.get("width", 0.8),
    124         dodge=True,
    125         fliersize=plot_params.get("fliersize", 5),
    126         linewidth=plot_params.get("linewidth"),
    127         saturation=.75, color=None, palette=None)
    129 elif plot == 'swarmplot':
    130     plotter = sns.categorical._SwarmPlotter(
    131         x, y, hue, data, order, hue_order,
    132         orient=plot_params.get("orient"),
    133         dodge=True, color=None, palette=None)

AttributeError: module 'seaborn.categorical' has no attribute '_BoxPlotter'

[wenyuhaokikika]

It need older version

import statannotations,matplotlib
sns.__version__,statannotations.__version__,matplotlib.__version__
>>> ('0.11.2', '0.5.0', '3.7.1')

and pip install statannotations==0.5.0,It works fine using the above code~~~

[mahlzahn]

@trevismd Any plans to keep this project going? Thanks for the work!

[xinwei-sher]

This is a much-needed package! Please update so it can be used with the latest Seaborn. Thank you!

[elide-b]

You could use the package starbars. You give the pairs and their p-value, and it draws it for you! It also has an option not to show non-significant p-value bars.

You can find the starbars documentation here.
Hope it helps! :)

Disclaimer: I'm the author of the package.

[wenyuhaokikika]

You could use the package starbars. You give the pairs and their p-value, and it draws it for you! It also has an option not to show non-significant p-value bars.

You can find the starbars documentation here. Hope it helps! :)

Disclaimer: I'm the author of the package.

How to label significance for a boxplot containing hue? How to use different inspection methods（ttest or ...）?

[elide-b]

Hey, sorry for the late reply!

I hadn't thought about the hue case, so thank you for pointing it out to me! I'll try to find a solution for that too.

Regarding the statistical methods, for now I haven't included it in the function, so you'd have to find the p-values with your chosen statistics separetely and then save the results in a tuple with the variables' names (var1, var2, pvalue).
Here's an example:

values_dict = {
    'A': [1, 2, 3, 4, 5],
    'B': [2, 3, 4, 5, 6],
    'C': [3, 4, 5, 6, 7]
}

annotations = []
keys = list(values_dict.keys())

# Perform t-test for each pair of variables
for i in range(len(keys)):
    for j in range(i + 1, len(keys)):
        key1 = keys[i]
        key2 = keys[j]
        values1 = values_dict[key1]
        values2 = values_dict[key2]

        t_stat, p_value = stats.ttest_ind(values1, values2)
        annotations.append((key1, key2, p_value))

values = [values_dict[key] for key in keys]

# Create a boxplot
plt.boxplot(values, tick_labels=keys, patch_artist=True)
starbars.draw_annotation(annotations)
plt.show()

Hope it helps and feel free to ask for more :)

[elide-b]

You could use the package starbars. You give the pairs and their p-value, and it draws it for you! It also has an option not to show non-significant p-value bars.
You can find the starbars documentation here. Hope it helps! :)
Disclaimer: I'm the author of the package.

How to label significance for a boxplot containing hue? How to use different inspection methods（ttest or ...）?

heyhey! Just checking in to say that the new 1.3.0 version of starbars allows you to use hues as labels. You'd just need to specify the hue label and group in a further tuple such as: ((hue_label1, group1), (hue_label2, group2), pvalue).

Check out the new documentation for more info and I'm always available for more help ✨

[xinwei-sher]

You could use the package starbars. You give the pairs and their p-value, and it draws it for you! It also has an option not to show non-significant p-value bars.
You can find the starbars documentation here. Hope it helps! :)
Disclaimer: I'm the author of the package.

How to label significance for a boxplot containing hue? How to use different inspection methods（ttest or ...）?

heyhey! Just checking in to say that the new 1.3.0 version of starbars allows you to use hues as labels. You'd just need to specify the hue label and group in a further tuple such as: ((hue_label1, group1), (hue_label2, group2), pvalue).

Check out the new documentation for more info and I'm always available for more help ✨
Thank you so much for making this tool available. Super helpful!