US TIGER data 2021 released

[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 2bbf3ed73b4bd94c06885e8f5f5ed6e2f0daca7c..a2c6b52073ec007e052b7775a148f6159fa1239d 100644 (file)
--- a/lib-sql/tokenizer/legacy_tokenizer.sql
+++ b/lib-sql/tokenizer/legacy_tokenizer.sql
@@ -202,57 +202,6 @@ $$
 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 $$
@@ -276,16 +225,6 @@ $$
 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[]
@@ -348,26 +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
@@ -380,24 +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;