As a user, I want the search engine to automatically handle common spelling variations so that I can more easily find titles with variant spelling.

[rlskoeser]

Notes for testing

• The following transformations are automatically done when content is indexed and keyword search terms are entered:
  • e at end of words is removed
  • vv is converted to w
  • ll is converted to l (double L to single)
  • j is converted to i
  • v is converted to u
• The search results have a temporary section displaying search terms in context so you can get a better sense of how the search is working.
• Try searching with both versions of the variant spellings (e.g. realme and realm or methode and method, or witch and vvitch) to see how they match the book content and check the terms that match in the highlighted section.

---

E.g. If I search for witch-craft, I would like results for “vvitch-craft” to appear.
Some common variations include i=j; vv=w; u=v. See: https://quod.lib.umich.edu/t/tcp/help/search-spelling.html

[rlskoeser]

I think we should be able to do this with a Solr filter. I expect we'd need to use the same filter on the indexed content and on the search input. Maybe the pattern replace filter? https://lucene.apache.org/solr/guide/6_6/filter-descriptions.html#FilterDescriptions-PatternReplaceFilter

Might also want to incorporate the unicode folding filter we used in Derrida.

Probably worth some investigation to see if anyone else has done something like this before.

[rlskoeser]

Maybe a mapping filter would work for some of these https://lucene.apache.org/solr/guide/6_6/charfilterfactories.html#CharFilterFactories-solr.MappingCharFilterFactory

[rlskoeser]

Schema API includes options for adding and replacing field types: https://lucene.apache.org/solr/guide/6_6/schema-api.html#SchemaAPI-ModifytheSchema