Author search handles initials incorrectly

There are a variety of problems with the way that search handles initials. The three most annoying are:

1. Failure to correctly tokenize initials which aren't separated by spaces.
2. Failure to consider order of initials when ranking search results
3. Inability to search for repeated initials
4. False positives (ie bad matches) for repeated initials (e.g. Daniel H. Wilson matches, and ranks higher than exact match H. H. Daniel)

The results for these two queries should be the same:
https://openlibrary.org/search/authors?q=h.a.j. (9 results)
https://openlibrary.org/search/authors?q=h.+a.+j. (254 results)

This query returns one of the worst matches first and pushes exact matches well down the list:
https://openlibrary.org/search/authors?q=e.+h.+smith

This query has an exact match but returns 50 extraneous results before getting to it (OL1661733A):
https://openlibrary.org/search/authors?q=h.+h.+daniel

### Proposal & Constraints
- [ ] Investigate whether we can make solr tokenize on `.` for authors? Do we want that?
- [x] #7371

### Related files

### Stakeholders