SET client_min_messages = warning;
CREATE EXTENSION IF NOT EXISTS icu_parser CASCADE;
SELECT tokid, alias, description FROM ts_token_type('icu') ORDER BY tokid;
 tokid | alias  |          description           
-------+--------+--------------------------------
     2 | word   | Word, letters
    12 | blank  | Space, punctuation and symbols
    22 | number | Number
    24 | kana   | Kana word
    25 | ideo   | Ideographic word
(5 rows)

SELECT icu_parser_icu_version() ~ '^[0-9]+\.[0-9]+' AS has_icu_version;
 has_icu_version 
-----------------
 t
(1 row)

SELECT alias, dictionaries FROM ts_debug('icu_simple', 'a 1 カナ 漢字') WHERE alias <> 'blank';
 alias  | dictionaries 
--------+--------------
 word   | {simple}
 number | {simple}
 ideo   | {simple}
 ideo   | {simple}
(4 rows)

-- Stemming and stop words for Latin text, ICU segmentation for the rest
CREATE TEXT SEARCH CONFIGURATION icu_english (PARSER = icu);
ALTER TEXT SEARCH CONFIGURATION icu_english ADD MAPPING FOR word WITH english_stem;
ALTER TEXT SEARCH CONFIGURATION icu_english ADD MAPPING FOR number, kana, ideo WITH simple;
SELECT to_tsvector('icu_english', 'The running cafés of 東京駅前');
            to_tsvector             
------------------------------------
 'café':3 'run':2 '東京':5 '駅前':6
(1 row)

-- Accent folding through unaccent
CREATE EXTENSION IF NOT EXISTS unaccent;
CREATE TEXT SEARCH CONFIGURATION icu_unaccent (PARSER = icu);
ALTER TEXT SEARCH CONFIGURATION icu_unaccent ADD MAPPING FOR word, number, kana, ideo WITH unaccent, simple;
SELECT to_tsvector('icu_unaccent', 'Café Zürich São Paulo');
              to_tsvector              
---------------------------------------
 'cafe':1 'paulo':4 'sao':3 'zurich':2
(1 row)

SELECT to_tsvector('icu_unaccent', 'Café') @@ plainto_tsquery('icu_unaccent', 'cafe') AS folded;
 folded 
--------
 t
(1 row)

-- Leaving a token type unmapped drops it
ALTER TEXT SEARCH CONFIGURATION icu_english DROP MAPPING FOR number;
SELECT to_tsvector('icu_english', 'route 66 東京');
    to_tsvector    
-------------------
 'rout':1 '東京':2
(1 row)

-- Unknown token types are rejected
ALTER TEXT SEARCH CONFIGURATION icu_english ADD MAPPING FOR asciiword WITH simple;
ERROR:  token type "asciiword" does not exist
DROP TEXT SEARCH CONFIGURATION icu_english;
DROP TEXT SEARCH CONFIGURATION icu_unaccent;
-- The extension can move schemas and be dropped and recreated
CREATE SCHEMA moved;
ALTER EXTENSION icu_parser SET SCHEMA moved;
SELECT to_tsvector('moved.icu_simple', '東京駅');
   to_tsvector   
-----------------
 '東京':1 '駅':2
(1 row)

ALTER EXTENSION icu_parser SET SCHEMA public;
DROP EXTENSION icu_parser;
DROP SCHEMA moved;
CREATE EXTENSION icu_parser CASCADE;
SELECT to_tsvector('icu_simple', '東京駅');
   to_tsvector   
-----------------
 '東京':1 '駅':2
(1 row)