File:  [ELWIX - Embedded LightWeight unIX -] / embedaddon / sqlite3 / test / fts3atoken.test
Revision 1.1.1.1 (vendor branch): download - view: text, annotated - select for diffs - revision graph
Tue Feb 21 17:04:16 2012 UTC (12 years, 10 months ago) by misho
Branches: sqlite3, MAIN
CVS tags: v3_7_10, HEAD
sqlite3

    1: # 2007 June 21
    2: #
    3: # The author disclaims copyright to this source code.  In place of
    4: # a legal notice, here is a blessing:
    5: #
    6: #    May you do good and not evil.
    7: #    May you find forgiveness for yourself and forgive others.
    8: #    May you share freely, never taking more than you give.
    9: #
   10: #*************************************************************************
   11: # This file implements regression tests for SQLite library. The focus 
   12: # of this script is testing the pluggable tokeniser feature of the 
   13: # FTS3 module.
   14: #
   15: # $Id: fts3atoken.test,v 1.1.1.1 2012/02/21 17:04:16 misho Exp $
   16: #
   17: 
   18: set testdir [file dirname $argv0]
   19: source $testdir/tester.tcl
   20: 
   21: # If SQLITE_ENABLE_FTS3 is defined, omit this file.
   22: ifcapable !fts3 {
   23:   finish_test
   24:   return
   25: }
   26: 
   27: set ::testprefix fts3token
   28: 
   29: proc escape_string {str} {
   30:   set out ""
   31:   foreach char [split $str ""] {
   32:     scan $char %c i
   33:     if {$i<=127} {
   34:       append out $char
   35:     } else {
   36:       append out [format {\x%.4x} $i]
   37:     }
   38:   }
   39:   set out
   40: }
   41: 
   42: #--------------------------------------------------------------------------
   43: # Test cases fts3token-1.* are the warm-body test for the SQL scalar
   44: # function fts3_tokenizer(). The procedure is as follows:
   45: #
   46: #   1: Verify that there is no such fts3 tokenizer as 'blah'.
   47: #
   48: #   2: Query for the built-in tokenizer 'simple'. Insert a copy of the
   49: #      retrieved value as tokenizer 'blah'.
   50: #
   51: #   3: Test that the value returned for tokenizer 'blah' is now the
   52: #      same as that retrieved for 'simple'.
   53: #
   54: #   4: Test that it is now possible to create an fts3 table using 
   55: #      tokenizer 'blah' (it was not possible in step 1).
   56: #
   57: #   5: Test that the table created to use tokenizer 'blah' is usable.
   58: #
   59: do_test fts3token-1.1 {
   60:   catchsql {
   61:     CREATE VIRTUAL TABLE t1 USING fts3(content, tokenize blah);
   62:   }
   63: } {1 {unknown tokenizer: blah}}
   64: do_test fts3token-1.2 {
   65:   execsql {
   66:     SELECT fts3_tokenizer('blah', fts3_tokenizer('simple')) IS NULL;
   67:   }
   68: } {0}
   69: do_test fts3token-1.3 {
   70:   execsql {
   71:     SELECT fts3_tokenizer('blah') == fts3_tokenizer('simple');
   72:   }
   73: } {1}
   74: do_test fts3token-1.4 {
   75:   catchsql {
   76:     CREATE VIRTUAL TABLE t1 USING fts3(content, tokenize blah);
   77:   }
   78: } {0 {}}
   79: do_test fts3token-1.5 {
   80:   execsql {
   81:     INSERT INTO t1(content) VALUES('There was movement at the station');
   82:     INSERT INTO t1(content) VALUES('For the word has passed around');
   83:     INSERT INTO t1(content) VALUES('That the colt from ol regret had got away');
   84:     SELECT content FROM t1 WHERE content MATCH 'movement'
   85:   }
   86: } {{There was movement at the station}}
   87: 
   88: #--------------------------------------------------------------------------
   89: # Test cases fts3token-2.* test error cases in the scalar function based
   90: # API for getting and setting tokenizers.
   91: #
   92: do_test fts3token-2.1 {
   93:   catchsql {
   94:     SELECT fts3_tokenizer('nosuchtokenizer');
   95:   }
   96: } {1 {unknown tokenizer: nosuchtokenizer}}
   97: 
   98: #--------------------------------------------------------------------------
   99: # Test cases fts3token-3.* test the three built-in tokenizers with a
  100: # simple input string via the built-in test function. This is as much
  101: # to test the test function as the tokenizer implementations.
  102: #
  103: do_test fts3token-3.1 {
  104:   execsql {
  105:     SELECT fts3_tokenizer_test('simple', 'I don''t see how');
  106:   }
  107: } {{0 i I 1 don don 2 t t 3 see see 4 how how}}
  108: do_test fts3token-3.2 {
  109:   execsql {
  110:     SELECT fts3_tokenizer_test('porter', 'I don''t see how');
  111:   }
  112: } {{0 i I 1 don don 2 t t 3 see see 4 how how}}
  113: ifcapable icu {
  114:   do_test fts3token-3.3 {
  115:     execsql {
  116:       SELECT fts3_tokenizer_test('icu', 'I don''t see how');
  117:     }
  118:   } {{0 i I 1 don't don't 2 see see 3 how how}}
  119: }
  120: 
  121: #--------------------------------------------------------------------------
  122: # Test cases fts3token-4.* test the ICU tokenizer. In practice, this
  123: # tokenizer only has two modes - "thai" and "everybody else". Some other
  124: # Asian languages (Lao, Khmer etc.) require the same special treatment as 
  125: # Thai, but ICU doesn't support them yet.
  126: #
  127: ifcapable icu {
  128: 
  129:   proc do_icu_test {name locale input output} {
  130:     set ::out [db eval { SELECT fts3_tokenizer_test('icu', $locale, $input) }]
  131:     do_test $name {
  132:       lindex $::out 0
  133:     } $output
  134:   }
  135:   
  136:   do_icu_test fts3token-4.1 en_US  {}   {}
  137:   do_icu_test fts3token-4.2 en_US {Test cases fts3} [list \
  138:     0 test Test 1 cases cases 2 fts3 fts3
  139:   ]
  140: 
  141:   # The following test shows that ICU is smart enough to recognise
  142:   # Thai chararacters, even when the locale is set to English/United 
  143:   # States.
  144:   #
  145:   set input "\u0e2d\u0e30\u0e44\u0e23\u0e19\u0e30\u0e04\u0e23\u0e31\u0e1a"
  146:   set output    "0 \u0e2d\u0e30\u0e44\u0e23 \u0e2d\u0e30\u0e44\u0e23 "
  147:   append output "1 \u0e19\u0e30 \u0e19\u0e30 "
  148:   append output "2 \u0e04\u0e23\u0e31\u0e1a \u0e04\u0e23\u0e31\u0e1a"
  149: 
  150:   do_icu_test fts3token-4.3 th_TH  $input $output
  151:   do_icu_test fts3token-4.4 en_US  $input $output
  152: 
  153:   # ICU handles an unknown locale by falling back to the default.
  154:   # So this is not an error.
  155:   do_icu_test fts3token-4.5 MiddleOfTheOcean  $input $output
  156: 
  157:   set    longtoken "AReallyReallyLongTokenOneThatWillSurelyRequire"
  158:   append longtoken "AReallocInTheIcuTokenizerCode"
  159: 
  160:   set    input "short tokens then "
  161:   append input $longtoken
  162:   set    output "0 short short "
  163:   append output "1 tokens tokens "
  164:   append output "2 then then "
  165:   append output "3 [string tolower $longtoken] $longtoken"
  166: 
  167:   do_icu_test fts3token-4.6 MiddleOfTheOcean  $input $output
  168:   do_icu_test fts3token-4.7 th_TH  $input $output
  169:   do_icu_test fts3token-4.8 en_US  $input $output
  170: 
  171:   do_execsql_test 5.1 {
  172:     CREATE VIRTUAL TABLE x1 USING fts3(name,TOKENIZE icu en_US);
  173:     insert into x1 (name) values (NULL);
  174:     insert into x1 (name) values (NULL);
  175:     delete from x1;
  176:   }
  177: }
  178: 
  179: 
  180: do_test fts3token-internal {
  181:   execsql { SELECT fts3_tokenizer_internal_test() }
  182: } {ok}
  183: 
  184: 
  185: finish_test
  186: 
  187: 

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>