[
https://issues.apache.org/jira/browse/SOLR-2921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13160406#comment-13160406
]
Erick Erickson commented on SOLR-2921:
--------------------------------------
Not synonyms - agreed.
Not shinglers - agreed.
Not anything that might produce multiple tokens - agreed.
Stemmers... When do stemmers produce multiple tokens? My ignorance of all the
possibilities knows no bounds. I was wondering if, in this case, stemmers
really reduced to prefix queries. Maybe it's just a bad idea altogether, I
guess it begs the question of what use adding stemmers to the mix would be. You
want to match the root, just specify the root with an asterisk and be done with
it. No need to introduce stemming into the MultiTermAwareComponent mix.
But this kind of question is exactly why I have this JIRA in place, we can
collect reasons I wouldn't think of and record them. Before I mess something up
with well-intentioned-but-wrong "help".
> Make any Filters, Tokenizers and CharFilters implement
> MultiTermAwareComponent if they should
> ---------------------------------------------------------------------------------------------
>
> Key: SOLR-2921
> URL: https://issues.apache.org/jira/browse/SOLR-2921
> Project: Solr
> Issue Type: Improvement
> Components: Schema and Analysis
> Affects Versions: 3.6, 4.0
> Environment: All
> Reporter: Erick Erickson
> Assignee: Erick Erickson
> Priority: Minor
>
> SOLR-2438 creates a new MultiTermAwareComponent interface. This allows Solr
> to automatically assemble a "multiterm" analyzer that does the right thing
> vis-a-vis transforming the individual terms of a multi-term query at query
> time. Examples are: lower casing, folding accents, etc. Currently
> (27-Nov-2011), the following classes implement MultiTermAwareComponent:
> * ASCIIFoldingFilterFactory
> * LowerCaseFilterFactory
> * LowerCaseTokenizerFactory
> * MappingCharFilterFactory
> * PersianCharFilterFactory
> When users put any of the above in their query analyzer, Solr will "do the
> right thing" at query time and the perennial question users have, "why didn't
> my wildcard query automatically lower-case (or accent fold or....) my terms?"
> will be gone. Die question die!
> But taking a quick look, for instance, at the various FilterFactories that
> exist, there are a number of possibilities that *might* be good candidates
> for implementing MultiTermAwareComponent. But I really don't understand the
> correct behavior here well enough to know whether these should implement the
> interface or not. And this doesn't include other CharFilters or Tokenizers.
> Actually implementing the interface is often trivial, see the classes above
> for examples. Note that LowerCaseTokenizerFactory returns a *Filter*, which
> is the right thing in this case.
> Here is a quick cull of the Filters that, just from their names, might be
> candidates. If anyone wants to take any of them on, that would be great. If
> all you can do is provide test cases, I could probably do the code part, just
> let me know.
> ArabicNormalizationFilterFactory
> GreekLowerCaseFilterFactory
> HindiNormalizationFilterFactory
> ICUFoldingFilterFactory
> ICUNormalizer2FilterFactory
> ICUTransformFilterFactory
> IndicNormalizationFilterFactory
> ISOLatin1AccentFilterFactory
> PersianNormalizationFilterFactory
> RussianLowerCaseFilterFactory
> TurkishLowerCaseFilterFactory
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
