Improve Ligurian tokenization

[jeanm]

Description

• Corrected one spelling mistake in the example text.
• Improved suffixes/prefixes:
  • support elision in years (e.g. 1990 -> ’90),
  • don't split the degree symbol when it's part of a degree unit (e.g. °C),
  • don't split left apostrophes for the few words that can have elision occur on the left (’na, ’n, ’n’).
• Improved handling of special cases:
  • handle compound prepositions (e.g. a-a, co-i) in a way that doesn't break compatibility with how they're dealt with in Universal Dependencies (using NORM as described in Multi-word tokens in non-English languages #1460),
  • handle cases such as °C, and generate the correct NORM forms for cases such as ’na.
• Added tests for all of the above.
• Added a few extra stop words, including variants with the curly quote / typographic apostrophe (commonly found in Ligurian corpora).

Types of change

Enhanced support for the Ligurian language.

Checklist

• I confirm that I have the right to submit this contribution under the project's MIT license.
• I ran the tests, and all new and existing tests passed.
• My changes don't require a change to the documentation, or if they do, I've added all required information.