API change with robustness

[trevismd]

---

This discussion was created from the release API change with robustness.

[Jamanaman]

Been using this version of statannotations for my final plots for my thesis and it's been great though I've had one issue. The test name has been displayed in both simple and full annotations for both t_test_welch and Mann-Whitney when I've set show_test_name false. My code's below to see if I've misunderstood or misused something.

plot_params = {'data': df, 'x': 'method', 'y': 'error', 'hue': 'Patient Population', 'estimator': nanmedian, 'ci': None}
fig, ax = plt.subplots()
ax = sns.barplot(ax=ax, **plot_params)
annotator = Annotator(ax, pairs, **plot_params)
annotator.configure(test='Mann-Whitney', comparisons_correction='BH', text_format='simple', show_test_name=False)
_, adjusted_pvals = annotator.apply_and_annotate()

I'm not including any further code or my plots just because it refers to patient data but if needs be I can de-identify stuff.

[trevismd]

Hi, thanks for your feedback and good question!
You're right, there's a bug there. I've fixed it and it should be included in the next release (v0.4.3) (already in the dev branch if you'd like to use it). (Also, if you want to stick to a previous version, you could try and cherry pick the commit and apply it.)

I spotted however that you intend to retrieve adjusted_pvals from apply_and_annotate() with Benjamini-Hochberg correction. You will not. Contrary to methods that work like "bonferroni", with these types of methods (BH, HB...), this package shows and returns the original p-values, but marks them as non-significant if applicable. If there is a demand for it, we could give the option to show and return corrected p-values, but it is not the case for now. See the discussion in #30 for more details.

How does that work for you ?

[Jamanaman]

Thanks for your reply @trevismd, I ended up just manually computing my p_values and using no test but I noticed that you started the PR for 0.4.3 with those changes included and just wanted to say that you're doing a great job with this project and I'm sure that I'll have plenty of use for this package going forward. Once my thesis is submitted I'll make sure to try and find some open issues or to contribute in some way to pay it back.