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

[pepsiv]

Sorry to bother you again.

Firstly, GraphCompass is a wonderful job. But when I try to create a boxplot of WWL with significance showing, a new issue appears.

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[23], [line 5](vscode-notebook-cell:?execution_count=23&line=5)
      1 # Note: a smaller Wasserstein distance indicates a higher similarity between the two graphs, 
      2 # while a larger distance indicates less similarity.
----> [5](vscode-notebook-cell:?execution_count=23&line=5) gc.pl.wlkernel.compare_conditions(
      6     adata=adata,
      7     library_key=library_key,
      8     condition_key=condition_key,
      9     control_group=control_group,
     10     metric_key=metric_key,
     11     method=method,
     12     figsize=(3,5),
     13     dpi=100,
     14     add_sign=True,
     15     #save="figures/visium_wwlkerenl.pdf"
     16 )

File ~/anaconda3/envs/yzx_GraphCompass_3.11/lib/python3.11/site-packages/graphcompass/pl/_WLkernel.py:143, in compare_conditions(adata, library_key, condition_key, control_group, metric_key, method, fig, ax, return_ax, figsize, dpi, color, palette, add_sign, save, **kwargs)
    140                 pairs.append((contrasts[i], contrasts[j]))
    142         from statannot import add_stat_annotation
--> [143](https://vscode-remote+ssh-002dremote-002bd04.vscode-resource.vscode-cdn.net/home/lixiangyu/yzx/GraphCompass/notebooks/wlkernel/~/anaconda3/envs/yzx_GraphCompass_3.11/lib/python3.11/site-packages/graphcompass/pl/_WLkernel.py:143)         add_stat_annotation(data=df_for_plot, x="contrast", y="vals",
    144                             ax=ax,
    145                             box_pairs=pairs,
    146                             test='t-test_ind', text_format='star', loc='outside', verbose=2, comparisons_correction=None)
...
    440         estimator=np.mean, ci=95, n_boot=1000, units=None,
    441         orient=None, color=None, palette=None, saturation=.75,
    442         errcolor=".26", errwidth=None, capsize=None, dodge=True)

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

It seems that there is a conflict between statannot and seaborn>=0.13.0. However, scanpy must need a seaborn>=0.13.0 version.

I have read the code of statannot and seaborn, ''statannot: box_plotter = sns.categorical._BoxPlotter'' is conflict to ''seaborn: barplot''.

---

I will appreciate it if you can provide a new method to add significance.

Best wishes!