Pending improvements in the Silhouette Computation function

In the `silhouette_computation ` function there are some comments about improvements that should be made in the plots (see below). Do you plan to include this later? Or did you just forget to include these improvements?

```
# TODO ADD PROPER TITLE/ AXIS/ MEAN OVERALL and mean per label ?
ax1 = sns.violinplot(
y=long_df_silhouette["label"], x=long_df_silhouette["silhouette"], scale="count"
)
matplotlib.pyplot.close() # Avoid plooting in function; bad practice
silhouette_results = silhouette_in(ax1, silhouette_avg)

return silhouette_results
```

Best,
Alberto

Start by locating the `silhouette_computation` function and reviewing the shown violin-plot code and its `silhouette_in` call. The work is complete when the plot has a proper title and axis labels, includes the requested overall and per-label means if appropriate, and avoids plotting as a side effect of the function.

Written by the indexing model from the issue text.