Allow ANOVA-style replotting of data in multi-group plots

[josesho]

import seaborn as sns
import dabest

# Requires internet access.
iris = sns.load_dataset("iris")

iris_db = dabest.load(iris, x="species", y="sepal_length",
                     idx=(("setosa", "versicolor"),
                          ("setosa", "virginica"),
                          ("versicolor", "virginica"))
                      )

iris_db.mean_diff

DABEST v0.2.4
=============
             
Good evening!
The current time is Tue May 28 18:30:11 2019.

The unpaired mean difference between setosa and versicolor is 0.93 [95%CI 0.76, 1.1].
The two-sided p-value of the Mann-Whitney test is 8.35e-14.

The unpaired mean difference between setosa and virginica is 1.58 [95%CI 1.38, 1.78].
The two-sided p-value of the Mann-Whitney test is 6.4e-17.

The unpaired mean difference between versicolor and virginica is 0.652 [95%CI 0.428, 0.878].
The two-sided p-value of the Mann-Whitney test is 5.87e-07.

5000 bootstrap samples were taken; the confidence interval is bias-corrected and accelerated.
The p-value(s) reported are the likelihood(s) of observing the effect size(s),
if the null hypothesis of zero difference is true.

To get the results of all valid statistical tests, use `.mean_diff.statistical_tests`

iris_db.mean_diff.plot();

The plot does not match the textual output above.

[josesho]

After much deliberation, duplicating groups across the same plot is misleading at best, and deceptive at worst. It also requires more effort to overhaul the plotting code.

The duplication of groups within idx or across tuples/lists inside idx will now be flagged up and will throw an error.