X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/d8ed1bfc601c6eb19b2db50df466654e843f0777..834ae0a93f2f25b3231f4c47612ac1d2c170738a:/lib-sql/tokenizer/legacy_tokenizer.sql diff --git a/lib-sql/tokenizer/legacy_tokenizer.sql b/lib-sql/tokenizer/legacy_tokenizer.sql index 916ba9ae..a2c6b520 100644 --- a/lib-sql/tokenizer/legacy_tokenizer.sql +++ b/lib-sql/tokenizer/legacy_tokenizer.sql @@ -34,6 +34,50 @@ AS $$ $$ LANGUAGE SQL IMMUTABLE STRICT; +CREATE OR REPLACE FUNCTION token_addr_street_match_tokens(info JSONB) + RETURNS INTEGER[] +AS $$ + SELECT (info->>'street')::INTEGER[] +$$ LANGUAGE SQL IMMUTABLE STRICT; + + +CREATE OR REPLACE FUNCTION token_addr_place_match_tokens(info JSONB) + RETURNS INTEGER[] +AS $$ + SELECT (info->>'place_match')::INTEGER[] +$$ LANGUAGE SQL IMMUTABLE STRICT; + + +CREATE OR REPLACE FUNCTION token_addr_place_search_tokens(info JSONB) + RETURNS INTEGER[] +AS $$ + SELECT (info->>'place_search')::INTEGER[] +$$ LANGUAGE SQL IMMUTABLE STRICT; + + +DROP TYPE IF EXISTS token_addresstoken CASCADE; +CREATE TYPE token_addresstoken AS ( + key TEXT, + match_tokens INT[], + search_tokens INT[] +); + +CREATE OR REPLACE FUNCTION token_get_address_tokens(info JSONB) + RETURNS SETOF token_addresstoken +AS $$ + SELECT key, (value->>1)::int[] as match_tokens, + (value->>0)::int[] as search_tokens + FROM jsonb_each(info->'addr'); +$$ LANGUAGE SQL IMMUTABLE STRICT; + + +CREATE OR REPLACE FUNCTION token_normalized_postcode(postcode TEXT) + RETURNS TEXT +AS $$ + SELECT CASE WHEN postcode SIMILAR TO '%(,|;)%' THEN NULL ELSE upper(trim(postcode))END; +$$ LANGUAGE SQL IMMUTABLE STRICT; + + -- Return token info that should be saved permanently in the database. CREATE OR REPLACE FUNCTION token_strip_info(info JSONB) RETURNS JSONB @@ -133,100 +177,26 @@ $$ LANGUAGE plpgsql; -CREATE OR REPLACE FUNCTION getorcreate_postcode_id(postcode TEXT) - RETURNS INTEGER +CREATE OR REPLACE FUNCTION create_postcode_id(postcode TEXT) + RETURNS BOOLEAN AS $$ DECLARE + r RECORD; lookup_token TEXT; - lookup_word TEXT; return_word_id INTEGER; BEGIN - lookup_word := upper(trim(postcode)); - lookup_token := ' ' || make_standard_name(lookup_word); - SELECT min(word_id) FROM word - WHERE word_token = lookup_token and word = lookup_word + lookup_token := ' ' || make_standard_name(postcode); + FOR r IN + SELECT word_id FROM word + WHERE word_token = lookup_token and word = postcode and class='place' and type='postcode' - INTO return_word_id; - IF return_word_id IS NULL THEN - return_word_id := nextval('seq_word'); - INSERT INTO word VALUES (return_word_id, lookup_token, lookup_word, - 'place', 'postcode', null, 0); - END IF; - RETURN return_word_id; -END; -$$ -LANGUAGE plpgsql; - - -CREATE OR REPLACE FUNCTION getorcreate_country(lookup_word TEXT, - lookup_country_code varchar(2)) - RETURNS INTEGER - AS $$ -DECLARE - lookup_token TEXT; - return_word_id INTEGER; -BEGIN - lookup_token := ' '||trim(lookup_word); - SELECT min(word_id) FROM word - WHERE word_token = lookup_token and country_code=lookup_country_code - INTO return_word_id; - IF return_word_id IS NULL THEN - return_word_id := nextval('seq_word'); - INSERT INTO word VALUES (return_word_id, lookup_token, null, - null, null, lookup_country_code, 0); - END IF; - RETURN return_word_id; -END; -$$ -LANGUAGE plpgsql; - - -CREATE OR REPLACE FUNCTION getorcreate_amenity(lookup_word TEXT, normalized_word TEXT, - lookup_class text, lookup_type text) - RETURNS INTEGER - AS $$ -DECLARE - lookup_token TEXT; - return_word_id INTEGER; -BEGIN - lookup_token := ' '||trim(lookup_word); - SELECT min(word_id) FROM word - WHERE word_token = lookup_token and word = normalized_word - and class = lookup_class and type = lookup_type - INTO return_word_id; - IF return_word_id IS NULL THEN - return_word_id := nextval('seq_word'); - INSERT INTO word VALUES (return_word_id, lookup_token, normalized_word, - lookup_class, lookup_type, null, 0); - END IF; - RETURN return_word_id; -END; -$$ -LANGUAGE plpgsql; - + LOOP + RETURN false; + END LOOP; -CREATE OR REPLACE FUNCTION getorcreate_amenityoperator(lookup_word TEXT, - normalized_word TEXT, - lookup_class text, - lookup_type text, - op text) - RETURNS INTEGER - AS $$ -DECLARE - lookup_token TEXT; - return_word_id INTEGER; -BEGIN - lookup_token := ' '||trim(lookup_word); - SELECT min(word_id) FROM word - WHERE word_token = lookup_token and word = normalized_word - and class = lookup_class and type = lookup_type and operator = op - INTO return_word_id; - IF return_word_id IS NULL THEN - return_word_id := nextval('seq_word'); - INSERT INTO word VALUES (return_word_id, lookup_token, normalized_word, - lookup_class, lookup_type, null, 0, op); - END IF; - RETURN return_word_id; + INSERT INTO word VALUES (nextval('seq_word'), lookup_token, postcode, + 'place', 'postcode', null, 0); + RETURN true; END; $$ LANGUAGE plpgsql; @@ -255,16 +225,6 @@ $$ LANGUAGE plpgsql; -CREATE OR REPLACE FUNCTION getorcreate_name_id(lookup_word TEXT) - RETURNS INTEGER - AS $$ -DECLARE -BEGIN - RETURN getorcreate_name_id(lookup_word, ''); -END; -$$ -LANGUAGE plpgsql; - -- Normalize a string and lookup its word ids (partial words). CREATE OR REPLACE FUNCTION addr_ids_from_name(lookup_word TEXT) RETURNS INTEGER[] @@ -319,36 +279,6 @@ $$ LANGUAGE plpgsql STABLE STRICT; -CREATE OR REPLACE FUNCTION create_country(src HSTORE, country_code varchar(2)) - RETURNS VOID - AS $$ -DECLARE - s TEXT; - w INTEGER; - words TEXT[]; - item RECORD; - j INTEGER; -BEGIN - FOR item IN SELECT (each(src)).* LOOP - - s := make_standard_name(item.value); - w := getorcreate_country(s, country_code); - - words := regexp_split_to_array(item.value, E'[,;()]'); - IF array_upper(words, 1) != 1 THEN - FOR j IN 1..array_upper(words, 1) LOOP - s := make_standard_name(words[j]); - IF s != '' THEN - w := getorcreate_country(s, country_code); - END IF; - END LOOP; - END IF; - END LOOP; -END; -$$ -LANGUAGE plpgsql; - - CREATE OR REPLACE FUNCTION make_keywords(src HSTORE) RETURNS INTEGER[] AS $$ @@ -357,26 +287,21 @@ DECLARE s TEXT; w INTEGER; words TEXT[]; - item RECORD; + value TEXT; j INTEGER; BEGIN result := '{}'::INTEGER[]; - FOR item IN SELECT (each(src)).* LOOP - - s := make_standard_name(item.value); - w := getorcreate_name_id(s, item.value); + FOR value IN SELECT unnest(regexp_split_to_array(svals(src), E'[,;]')) LOOP + -- full name + s := make_standard_name(value); + w := getorcreate_name_id(s, value); IF not(ARRAY[w] <@ result) THEN result := result || w; END IF; - w := getorcreate_word_id(s); - - IF w IS NOT NULL AND NOT (ARRAY[w] <@ result) THEN - result := result || w; - END IF; - + -- partial single-word terms words := string_to_array(s, ' '); IF array_upper(words, 1) IS NOT NULL THEN FOR j IN 1..array_upper(words, 1) LOOP @@ -389,24 +314,23 @@ BEGIN END LOOP; END IF; - words := regexp_split_to_array(item.value, E'[,;()]'); - IF array_upper(words, 1) != 1 THEN - FOR j IN 1..array_upper(words, 1) LOOP - s := make_standard_name(words[j]); - IF s != '' THEN - w := getorcreate_word_id(s); - IF w IS NOT NULL AND NOT (ARRAY[w] <@ result) THEN - result := result || w; - END IF; + -- consider parts before an opening braket a full word as well + words := regexp_split_to_array(value, E'[(]'); + IF array_upper(words, 1) > 1 THEN + s := make_standard_name(words[1]); + IF s != '' THEN + w := getorcreate_name_id(s, words[1]); + IF w IS NOT NULL AND NOT (ARRAY[w] <@ result) THEN + result := result || w; END IF; - END LOOP; + END IF; END IF; - s := regexp_replace(item.value, '市$', ''); - IF s != item.value THEN + s := regexp_replace(value, '市$', ''); + IF s != value THEN s := make_standard_name(s); IF s != '' THEN - w := getorcreate_name_id(s, item.value); + w := getorcreate_name_id(s, value); IF NOT (ARRAY[w] <@ result) THEN result := result || w; END IF;