Find a workaround for the implied wildcard (*) in autocomplete and searches

[kltm]

Currently, in many instances while typing for a live search in AmiGO, there is a wildcard inserted silently after the current text unless a space is added. The purpose of this is that using the current GOlr scheme, we can have both autocomplete and search working off the same index; although we have additional workarounds for the general and term autocomplete indexes.

Ideally, we should revisit how the index is constructed and consider using different filters and definition to try and eliminate the complexities that have grown up around the original schema.

For example, better use of EdgeNGramFilter, a la:

http://lucene.472066.n3.nabble.com/How-to-do-partial-beginning-matches-td781147.html
https://stackoverflow.com/questions/2780973/how-to-do-partial-beginning-matches-in-solr
https://stackoverflow.com/questions/17521699/solr-3-5-substring-match-with-edgengramfilter

[kltm]

This will likely involve a fair bit of mucking around and experimenting at the GOlr level, so may be kicked down the road a bit. This may also end up being the solution to #132.

[kltm]

It might make sense to group this in to whatever release we do for going to Solr 4.x, since we'll be tearing up some of the Solr floorboards anyways.

[kltm]

As well, it may be useful to add a "general_blob" (a la the general schema) to most of the schemas to let us add extras to more easily/quickly goose results. OR is this a hack?

[kltm]

This is fundamentally related to the search reworking in berkeleybop/bbop-js#16.