/* kaasearch extension: full-text search for the Karakalpak language */ \echo Use "CREATE EXTENSION pg_kaasearch" to load this file. \quit /* C functions of the text search template */ CREATE FUNCTION kaasearch_init(internal) RETURNS internal AS 'MODULE_PATHNAME' LANGUAGE C STRICT; CREATE FUNCTION kaasearch_lexize(internal, internal, internal, internal) RETURNS internal AS 'MODULE_PATHNAME' LANGUAGE C STRICT; /* Step 1: the text search template */ CREATE TEXT SEARCH TEMPLATE kaasearch_template ( INIT = kaasearch_init, LEXIZE = kaasearch_lexize ); /* Step 2: the dictionary with its stop words ($SHAREDIR/tsearch_data/kaasearch.stop). A user lexicon can be added: ALTER TEXT SEARCH DICTIONARY kaasearch_dict (Lexicon = name) reads $SHAREDIR/tsearch_data/name.kaalex */ CREATE TEXT SEARCH DICTIONARY kaasearch_dict ( TEMPLATE = kaasearch_template, StopWords = kaasearch ); /* Step 3: the parser: the default parser, with the apostrophe spelling joined first (o'zbek → ózbek; the default parser would end the word at the apostrophe) */ CREATE FUNCTION kaasearch_prsd_start(internal, integer) RETURNS internal AS 'MODULE_PATHNAME' LANGUAGE C STRICT; CREATE TEXT SEARCH PARSER kaasearch_parser ( START = kaasearch_prsd_start, GETTOKEN = pg_catalog.prsd_nexttoken, END = pg_catalog.prsd_end, HEADLINE = pg_catalog.prsd_headline, LEXTYPES = pg_catalog.prsd_lextype ); COMMENT ON TEXT SEARCH PARSER kaasearch_parser IS 'Default parser that reads the Karakalpak apostrophe spelling (o''zbek) as one word'; /* Step 4: the text search configuration */ CREATE TEXT SEARCH CONFIGURATION kaasearch ( PARSER = kaasearch_parser ); COMMENT ON TEXT SEARCH CONFIGURATION kaasearch IS 'Configuration for Karakalpak language (Latin/Cyrillic)'; /* Step 5: map word tokens to our dictionary (words with digits included: covid19, 5-mektepke) */ ALTER TEXT SEARCH CONFIGURATION kaasearch ALTER MAPPING FOR asciiword, asciihword, hword_asciipart, word, hword, hword_part, numword, numhword, hword_numpart WITH kaasearch_dict; /* Step 6: index numbers as they are (2016, 3.14, 1.2.3) */ ALTER TEXT SEARCH CONFIGURATION kaasearch ALTER MAPPING FOR int, uint, float, sfloat, version WITH simple; /* Step 7: functions for working with the stemmer directly */ CREATE FUNCTION kaa_stem(word text) RETURNS text AS 'MODULE_PATHNAME', 'kaa_stem_pg' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE; COMMENT ON FUNCTION kaa_stem(text) IS 'Stem of one Karakalpak word: kaa_stem(''mektebimizde'') = ''mektep'''; CREATE FUNCTION kaa_analyze(word text, OUT word text, OUT stem text, OUT segments text[], OUT classes text[], OUT source text) RETURNS record AS 'MODULE_PATHNAME', 'kaa_analyze_pg' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE; COMMENT ON FUNCTION kaa_analyze(text) IS 'Analysis of one Karakalpak word: normalised word, stem, segmentation, suffix classes and source'; CREATE FUNCTION kaa_normalize(txt text) RETURNS text AS 'MODULE_PATHNAME', 'kaa_normalize_pg' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE; COMMENT ON FUNCTION kaa_normalize(text) IS 'Text with every word in 2016 Latin spelling and lower case (Cyrillic, apostrophe spelling)';