KinoSearch1::Analysis::PolyAnalyzer - multiple analyzers in series
my $analyzer = KinoSearch1::Analysis::PolyAnalyzer->new( language => 'es', ); # or... my $analyzer = KinoSearch1::Analysis::PolyAnalyzer->new( analyzers => [ $lc_normalizer, $custom_tokenizer, $snowball_stemmer, ], );
A PolyAnalyzer is a series of Analyzers -- objects which inherit from KinoSearch1::Analysis::Analyzer -- each of which will be called upon to "analyze" text in turn. You can either provide the Analyzers yourself, or you can specify a supported language, in which case a PolyAnalyzer consisting of an LCNormalizer, a Tokenizer, and a Stemmer will be generated for you.
en => English, da => Danish, de => German, es => Spanish, fi => Finnish, fr => French, it => Italian, nl => Dutch, no => Norwegian, pt => Portuguese, ru => Russian, sv => Swedish,
my $analyzer = KinoSearch1::Analysis::PolyAnalyzer->new( language => 'en', );
Construct a PolyAnalyzer object. If the parameter
analyzers is specified, it will override
language and no attempt will be made to generate a default set of Analyzers.
language - Must be an ISO code from the list of supported languages.
analyzers - Must be an arrayref. Each element in the array must inherit from KinoSearch1::Analysis::Analyzer. The order of the analyzers matters. Don't put a Stemmer before a Tokenizer (can't stem whole documents or paragraphs -- just individual words), or a Stopalizer after a Stemmer (stemmed words, e.g. "themselv", will not appear in a stoplist). In general, the sequence should be: normalize, tokenize, stopalize, stem.
Copyright 2005-2010 Marvin Humphrey
See KinoSearch1 version 1.01.