Merge pull request #3347 from lonvia/tweak-boundary-imports

[nominatim.git] / lib-sql / tokenizer / legacy_tokenizer.sql

diff --git a/lib-sql/tokenizer/legacy_tokenizer.sql b/lib-sql/tokenizer/legacy_tokenizer.sql
index 4c22424ec74afa2504f51bbebdf9498fba2e740a..8c8f56e1c264e937bf8617ab56b5e6e9d83325e2 100644 (file)
--- a/lib-sql/tokenizer/legacy_tokenizer.sql
+++ b/lib-sql/tokenizer/legacy_tokenizer.sql
@@ -1,3 +1,10 @@
+-- SPDX-License-Identifier: GPL-2.0-only
+--
+-- This file is part of Nominatim. (https://nominatim.org)
+--
+-- Copyright (C) 2022 by the Nominatim developer community.
+-- For a full list of authors see the git log.
+

 -- Get tokens used for searching the given place.
 --
 -- These are the tokens that will be saved in the search_name table.
 -- Get tokens used for searching the given place.
 --
 -- These are the tokens that will be saved in the search_name table.


@@ -34,10 +41,73 @@ AS $$
 $$ LANGUAGE SQL IMMUTABLE STRICT;
 
 
 $$ LANGUAGE SQL IMMUTABLE STRICT;
 
 

-CREATE OR REPLACE FUNCTION token_normalized_postcode(postcode TEXT)
+CREATE OR REPLACE FUNCTION token_is_street_address(info JSONB)
+  RETURNS BOOLEAN
+AS $$
+  SELECT info->>'street' is not null or info->>'place_search' is null;
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION token_has_addr_street(info JSONB)
+  RETURNS BOOLEAN
+AS $$
+  SELECT info->>'street' is not null and info->>'street' != '{}';
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION token_has_addr_place(info JSONB)
+  RETURNS BOOLEAN
+AS $$
+  SELECT info->>'place_match' is not null;
+$$ LANGUAGE SQL IMMUTABLE;
+
+
+CREATE OR REPLACE FUNCTION token_matches_street(info JSONB, street_tokens INTEGER[])
+  RETURNS BOOLEAN
+AS $$
+  SELECT (info->>'street')::INTEGER[] && street_tokens
+$$ LANGUAGE SQL IMMUTABLE STRICT;
+
+
+CREATE OR REPLACE FUNCTION token_matches_place(info JSONB, place_tokens INTEGER[])
+  RETURNS BOOLEAN
+AS $$
+  SELECT (info->>'place_match')::INTEGER[] && place_tokens
+$$ 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;
+
+
+CREATE OR REPLACE FUNCTION token_get_address_keys(info JSONB)
+  RETURNS SETOF TEXT
+AS $$
+  SELECT * FROM jsonb_object_keys(info->'addr');
+$$ LANGUAGE SQL IMMUTABLE STRICT;
+
+
+CREATE OR REPLACE FUNCTION token_get_address_search_tokens(info JSONB, key TEXT)
+  RETURNS INTEGER[]
+AS $$
+  SELECT (info->'addr'->key->>0)::INTEGER[];
+$$ LANGUAGE SQL IMMUTABLE STRICT;
+
+
+CREATE OR REPLACE FUNCTION token_matches_address(info JSONB, key TEXT, tokens INTEGER[])
+  RETURNS BOOLEAN
+AS $$
+  SELECT (info->'addr'->key->>1)::INTEGER[] && tokens;
+$$ LANGUAGE SQL IMMUTABLE STRICT;
+
+
+CREATE OR REPLACE FUNCTION token_get_postcode(info JSONB)

   RETURNS TEXT
 AS $$
   RETURNS TEXT
 AS $$

-  SELECT CASE WHEN postcode SIMILAR TO '%(,|;)%' THEN NULL ELSE upper(trim(postcode))END;
+  SELECT info->>'postcode';

 $$ LANGUAGE SQL IMMUTABLE STRICT;
 
 
 $$ LANGUAGE SQL IMMUTABLE STRICT;
 
 


@@ -165,80 +235,6 @@ $$
 LANGUAGE plpgsql;
 
 
 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;
-
-
-CREATE OR REPLACE FUNCTION getorcreate_amenityoperator(lookup_word TEXT,
-                                                       normalized_word TEXT,
-                                                       lookup_class text,
-                                                       lookup_type text,
-                                                       op text)
-  RETURNS INTEGER
-  AS $$
-DECLARE
-  lookup_token TEXT;
-  return_word_id INTEGER;
-BEGIN
-  lookup_token := ' '||trim(lookup_word);
-  SELECT min(word_id) FROM word
-  WHERE word_token = lookup_token and word = normalized_word
-        and class = lookup_class and type = lookup_type and operator = op
-  INTO return_word_id;
-  IF return_word_id IS NULL THEN
-    return_word_id := nextval('seq_word');
-    INSERT INTO word VALUES (return_word_id, lookup_token, normalized_word,
-                             lookup_class, lookup_type, null, 0, op);
-  END IF;
-  RETURN return_word_id;
-END;
-$$
-LANGUAGE plpgsql;
-
-

 CREATE OR REPLACE FUNCTION getorcreate_name_id(lookup_word TEXT, src_word TEXT)
   RETURNS INTEGER
   AS $$
 CREATE OR REPLACE FUNCTION getorcreate_name_id(lookup_word TEXT, src_word TEXT)
   RETURNS INTEGER
   AS $$


@@ -262,16 +258,6 @@ $$
 LANGUAGE plpgsql;
 
 
 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[]
 -- Normalize a string and lookup its word ids (partial words).
 CREATE OR REPLACE FUNCTION addr_ids_from_name(lookup_word TEXT)
   RETURNS INTEGER[]


@@ -326,36 +312,6 @@ $$
 LANGUAGE plpgsql STABLE STRICT;
 
 
 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 $$
 CREATE OR REPLACE FUNCTION make_keywords(src HSTORE)
   RETURNS INTEGER[]
   AS $$


@@ -364,26 +320,21 @@ DECLARE
   s TEXT;
   w INTEGER;
   words TEXT[];
   s TEXT;
   w INTEGER;
   words TEXT[];

-  item RECORD;
+  value TEXT;

   j INTEGER;
 BEGIN
   result := '{}'::INTEGER[];
 
   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;
 
 
     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
     words := string_to_array(s, ' ');
     IF array_upper(words, 1) IS NOT NULL THEN
       FOR j IN 1..array_upper(words, 1) LOOP


@@ -396,24 +347,23 @@ BEGIN
       END LOOP;
     END IF;
 
       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 IF;

-      END LOOP;
+      END IF;

     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
       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;
         IF NOT (ARRAY[w] <@ result) THEN
           result := result || w;
         END IF;