MB-19243: Detect fuzziness automatically based on term length

[CascadingRadium]

• The following queries can now automatically detect fuzziness based on the length of the terms:
  • Match Query
  • Fuzzy Query
  • Match-Phrase Query
  • Multi-Phrase Query
  • Phrase Query
• In these queries, each term (whether in a multi-term query like Match or Phrase, or in a single-term query like Fuzzy) can have its own edit distance based on its length. The edit distance is calculated as follows:
  • For terms with 1 or 2 characters: edit distance = 0 (exact match)
  • For terms with 3, 4, or 5 characters: edit distance = 1 (fuzzy match)
  • For terms with more than 5 characters: edit distance = 2 (fuzzy match)
• This feature can be enabled using the <query>.SetAutoFuzziness(<bool>) API.
• Additionally, we've extended the functionality to query JSON parsing. You can specify fuzziness as either "auto" or a static value in the JSON query. Both formats are valid:

1. With auto fuzziness:

{
  "match" : "lorem",
  "field" : "bleve"
  "fuzziness" : "auto"
}

2. With static fuzziness:

{
  "match" : "lorem",
  "field" : "bleve"
  "fuzziness" : 2
}

When unmarshalled, the query will correctly apply the chosen fuzziness method.

• Fixed a bug where the code incorrectly returned an error message saying fuzziness exceeds maximum when using a fuzzy searcher with fuzziness = 0. Instead, a term searcher is now returned in this case.