API

my $fts=Search::OpenFTS->new( DBI, relfunc=>name_of_rel_function_in_psql, txttid=>NAME_OF_TXTID, prefix=>PREFIX

);

Example of relfunc:

relfunc=>q[ rank( '{0.1, 0.2, 0.4, 1.0}', $TSVECTOR, $QUERY, 0 )]

first option is weights (see tsearch documentation), Last argument defines how to normalize weight of document: 0 - no normalization (default) 1 - normalized by log(length of document) 2 - normalized by length of document

Methods

get_sql( \@ARRAY_WORD );

get_sql( $STRING );

get_sql( \$STRING );

get_sql( *, %opt );

%opt - as in the constructor (see above), plus a key dict_opt = > {}, transmitted to dictionaries

Returns parts of SQL:

($out, $condition, $order)

Here is how they can be combined in an SQL statement:

 SELECT
    $opt{txttid}$out
 FROM
    table
 WHERE
    $condition
 $order;
    

search( SEARCH_AS_IN_GET_SQL )

search( SEARCH_AS_IN_GET_SQL, %opt )

Returns the reference to the list of identifiers sorted by relevance

get_headline(

    query=>$query,
    src=>[$FH|$txt|$reftxt],
    maxlen=>$maxlen,
    maxread=>$maxread,
    otag=>$opentag,
    ctag=>$closetag,
    replace_ignore_headline=>$str_to_replace,
    dict_opt=>{})
    

Returns fragment of the document with search terms hilighted. maxread bytes reads from the document to generates headline with length maxlen.

otag,ctag denote strings used for hilighting, for example, <b>, </b>.

replace_ignore_headline - string used to replace html markups, space by default.

get_headline2(

    query=>$query,
    src=>[$FH|$txt|$reftxt],
    min_words=>$min_words,
    max_words=>$max_words,
    otag=>$opentag,
    ctag=>$closetag,)
    

Another method for getting headline. This method should be a little bit slower but more accurate. The following parameters are recognized:

min_words - minimal number of words in headline;

max_words - maximal number of words in headline;

shortword - maximal length of word that will be rejected at the end of headline;

sentence_length - the expected length of sentence to cut headline at the end of sentence;

nonword_tokens - a list of token ids (separated by spaces) which are not counted as words;

leave_tokens - a list of token ids that are not considered as words but that will appear in the output;

complex_tokens - a list of token ids that will be split on smaller tokens;

noend_tokens - token ids which are not desirable at the end of headline;

endpunct_tokens - token ids which represents end-of-sentence punctuation signs.

met_query - if set to a scalar ref stores true value if query has been found in processed text.