emmanuel / aethon Goto Github PK

There are several options for driving Solr from Ruby: rsolr, delsolr and sunspot
(which is a significant layer of syntactic sugar on top or rsolr's fairly raw interface).

Each has advantages and disadvantages:

1) rsolr provides a fairly raw (low-level) interface,
and it does not seem to be encoding-aware (w/r/t Ruby 1.9's Encoding support).
2) delsolr provides a midlevel interface, including batching of documents posted to Solr.
3) sunspot provides a very high-level interface, including DSLs for both indexing and
querying Solr. It also includes ORM interface code (esp. ActiveRecord), for making
the the task of indexing and querying data the mirrors an RDBMS more straightforward.

Why another Ruby Solr lib?

Quite simply, I don't find the internals of any of the listed options to be satisfactory.


Some might argue that the better thing would be for me to invest my energy in refining one
of the existing options to meet my goals. To them I refer to the principle of Schumpeterian
creative destruction. Alternately — evolution of a species proceeds more rapidly between
specimens than within them.