You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by Areek Zillur <ar...@gmail.com> on 2013/11/05 09:01:00 UTC

Suggester Version 2 Code Complete

I sent out an email couple of weeks ago about the Suggester Version 2 (
https://issues.apache.org/jira/browse/SOLR-5378) and what I was planning to
improve. The status of the patch is now code-complete, along with
documentation (javadocs) and tests.

It would be awesome if I could get some feedback/input regarding the patch.

Here are some highlights (the options,config, request and response formats
are all detailed in the jira description):

Features:
  - Simplified Configuration
  - Added distributed support for Suggesters (merge using suggestion
    weights)
  - Added dictionary pluggability support to allow easy way to manipulate
    underlying suggester input (index time), along with factories for all
available lucene dictionaries.
  - Added statistics to be displayed [including size of the underlying
    in-memory data structures created by various suggesters]

Available Dictionary Factories
  - DocumentDictionary - input from stored documents (users can specify
fields for suggestions, weight and optionally payload)
  - DocumentExpressionDictionary - same as DocumentDictionary but using
    expressions to specify suggestion weights
  - FileDictionary - input from external file
  - HighFrequencyDictionary - input from documents but where weight is
    the term frequency

Available lookup Factories (same as previous suggester version)
  - AnalyzingSuggester
  - FuzzySuggester
  - AnalyzingInfixSuggester
  - FSTLookup
  - JaspellLookup
  - TSTLookup
  - WFSTLookup

Thanks in advance,

Areek Zillur