Check consonant clusters

[chrzyki]

Command as proposed and implemented by @LinguList for checking (potentially) problematic consonant clusters.

[FredericBlum]

I have two ideas for potential additions:

• Add a summary per dataset, so we can identify the ˈworstˈ cases that need immediate attention
• Add an argument for the length of clusters to be checked. 3 could be the default, but maybe we want ot check for 4 or 5 first, to identify the most problematic cases?

Also, I am thinking that maybe the Segments would be a more relevant output than the (unstandardized) form. It seems difficult at some points to identify the problematic part in the forms. Maybe we could output both? What do you think?