Full text search indexing
=========================

The following FTS indexers are supported:

 * <Squat> [Plugins.FTS.Squat.txt] is Dovecot's own search index.
 * <Lucene> [Plugins.FTS.Lucene.txt] uses Lucene's C++ library.
 * <Solr> [Plugins.FTS.Solr.txt] communicates with Lucene's Solr server
   [http://lucene.apache.org/solr/]. Use this instead of fts-lucene if
   possible.

The main difference between Squat indexes and the others is that Squat provides
support for substring searches. This is required for IMAP protocol compliance,
so TEXT and BODY searches are looked up only from Squat indices. Dovecot's
non-standard X-TEXT-FAST and X-BODY-FAST SEARCH parameters can be used to look
up from non-Squat indexes.

Currently the indices are updated *only* while searching, so neither the
<LDA.txt> nor an IMAP APPEND command updates the indices immediately. This
means that if user has received a lot of mail since the last indexing (==
search operation), it may take a while to index all the mails before replying
to the search command. Dovecot sends periodic "* OK Indexed n% of the mailbox"
updates which can be caught by webmail implementations to implement a progress
bar. You could also build a cronjob to index users' mailboxes every once in a
while (by selecting each mailbox and issuing a 'SEARCH TEXT xyzzyx' command).

Using both Squat and non-Squat
------------------------------

It's possible to use both Squat and non-Squat indices at the same time, so that
TEXT/BODY are looked up from Squat indexes and X-TEXT-FAST/X-BODY-FAST are
looked up from the non-Squat index. This of course means that indices will have
to be built and updated for both, so it might not be that good idea.

---%<-------------------------------------------------------------------------
protocol imap {
..
  mail_plugins = fts fts_squat fts_solr
}
...
plugin {
  fts = squat solr
}
---%<-------------------------------------------------------------------------

(This file was created from the wiki on 2011-08-29 04:42)