This PR attempts to add an exactish index for synonyms by making a CopyField for names and indexing it with the same exactish field type (which uses a lowercase keyword tokenizer). This slightly increases our database size (the gzipped tar file increased from 89G to 98G), but provides the ability to boost exact matches higher than approximate matches. This PR modifies the query to do a better job at escaping queries to send to Solr and boosting among these two types of matches. It also updates one of the tests, as some of our searching has improved as a consequence of these changes.
This PR attempts to add an exactish index for synonyms by making a CopyField for
names
and indexing it with the same exactish field type (which uses a lowercase keyword tokenizer). This slightly increases our database size (the gzipped tar file increased from 89G to 98G), but provides the ability to boost exact matches higher than approximate matches. This PR modifies the query to do a better job at escaping queries to send to Solr and boosting among these two types of matches. It also updates one of the tests, as some of our searching has improved as a consequence of these changes.