Also removes errant "suggest" field which doesn't line up with anything in the index.
Removes 2<90% word containment from multi_match, as the layered combined_fields and overall minimum_should_match is sufficient. This is helpful for instances where patrons are trying to match against single word titles with only an author's last name.
I also added a development & feature prefix pattern for this repo, as there's also an intent to add either a stemmer or ngram analyzer later to help match roots against pluralizations and suffixes. Thought it might help with some incremental A/B testing, because that change likely to change results overall more than this one. Open to another workflow if anyone who wants to contribute has a simpler one in mind.
Allows floating prefixed queries against all fields, so {ducks AND pigeon callnums:"picture books"} is supported.
Fix the boolean operators in baseline queries, which had been disrupted by the phrase_prefix clauses implemented to improve short words.
Also removes errant "suggest" field which doesn't line up with anything in the index.
Removes 2<90% word containment from multi_match, as the layered combined_fields and overall minimum_should_match is sufficient. This is helpful for instances where patrons are trying to match against single word titles with only an author's last name.
I also added a development & feature prefix pattern for this repo, as there's also an intent to add either a stemmer or ngram analyzer later to help match roots against pluralizations and suffixes. Thought it might help with some incremental A/B testing, because that change likely to change results overall more than this one. Open to another workflow if anyone who wants to contribute has a simpler one in mind.