BibIndex: fuzzy author name tokenizer

• Introduced fuzzy author name tokenizer.
  • Get a tokenizer with b_e_t.BibIndexFuzzyNameTokenizer().
  • Call tokenizer.tokenize(name) to get a (potentially long) list of expanded forms, suitable for phrase-indexing. Strings in, lists of strings out.
  • Or, call with tokenizer.scan(name) to turn name into an idiosyncratic data structure used by tokenizer.parse_scanned. This structure is a dictionary that tags non-lastnames, lastnames, and titles.
  • You can also call tokenizer.parse_scanned(tagged stuff) to generate the expanded forms directly from tagged data.

• Includes unit tests for all of the above that cover something like 98% of the common cases. Effort has been made to test for pathological names also.
  
  (closes: #14366, #64426, #14513)