sqlite3/ext/fts5/test/fts5trigram.test

219 lines
5.8 KiB
Plaintext
Raw Permalink Normal View History

2023-06-27 18:34:42 +08:00
# 2020 September 30
#
# The author disclaims copyright to this source code. In place of
# a legal notice, here is a blessing:
#
# May you do good and not evil.
# May you find forgiveness for yourself and forgive others.
# May you share freely, never taking more than you give.
#
#*************************************************************************
#
# Tests for the fts5 "trigram" tokenizer.
#
source [file join [file dirname [info script]] fts5_common.tcl]
ifcapable !fts5 { finish_test ; return }
set ::testprefix fts5trigram
do_execsql_test 1.0 {
CREATE VIRTUAL TABLE t1 USING fts5(y, tokenize=trigram);
INSERT INTO t1 VALUES('abcdefghijklm');
INSERT INTO t1 VALUES('กรุงเทพมหานคร');
}
foreach {tn s res} {
1 abc "(abc)defghijklm"
2 defgh "abc(defgh)ijklm"
3 abcdefghijklm "(abcdefghijklm)"
4 กรุ "(กรุ)งเทพมหานคร"
5 งเทพมห "กรุ(งเทพมห)านคร"
6 กรุงเทพมหานคร "(กรุงเทพมหานคร)"
7 Abc "(abc)defghijklm"
8 deFgh "abc(defgh)ijklm"
9 aBcdefGhijKlm "(abcdefghijklm)"
} {
do_execsql_test 1.1.$tn {
SELECT highlight(t1, 0, '(', ')') FROM t1($s)
} $res
}
do_execsql_test 1.2.0 {
SELECT fts5_expr('ABCD', 'tokenize=trigram')
} {{"abc" + "bcd"}}
do_execsql_test 1.2.1 {
SELECT * FROM t1 WHERE y LIKE ? ESCAPE 'a'
}
foreach {tn like res} {
1 {%cDef%} 1
2 {cDef%} {}
3 {%f%} 1
4 {%f_h%} 1
5 {%f_g%} {}
6 {abc%klm} 1
7 {ABCDEFG%} 1
8 {%รุงเ%} 2
9 {%งเ%} 2
} {
do_execsql_test 1.3.$tn {
SELECT rowid FROM t1 WHERE y LIKE $like
} $res
}
#-------------------------------------------------------------------------
reset_db
do_execsql_test 2.0 {
CREATE VIRTUAL TABLE t1 USING fts5(y, tokenize="trigram case_sensitive 1");
INSERT INTO t1 VALUES('abcdefghijklm');
INSERT INTO t1 VALUES('กรุงเทพมหานคร');
}
foreach {tn s res} {
1 abc "(abc)defghijklm"
2 defgh "abc(defgh)ijklm"
3 abcdefghijklm "(abcdefghijklm)"
4 กรุ "(กรุ)งเทพมหานคร"
5 งเทพมห "กรุ(งเทพมห)านคร"
6 กรุงเทพมหานคร "(กรุงเทพมหานคร)"
7 Abc ""
8 deFgh ""
9 aBcdefGhijKlm ""
} {
do_execsql_test 2.1.$tn {
SELECT highlight(t1, 0, '(', ')') FROM t1($s)
} $res
}
foreach {tn like res} {
1 {%cDef%} 1
2 {cDef%} {}
3 {%f%} 1
4 {%f_h%} 1
5 {%f_g%} {}
6 {abc%klm} 1
7 {ABCDEFG%} 1
8 {%รุงเ%} 2
} {
do_execsql_test 2.2.$tn {
SELECT rowid FROM t1 WHERE y LIKE $like
} $res
}
foreach {tn like res} {
1 {*cdef*} 1
2 {cdef*} {}
3 {*f*} 1
4 {*f?h*} 1
5 {*f?g*} {}
6 {abc*klm} 1
7 {abcdefg*} 1
8 {*รุงเ*} 2
9 {abc[d]efg*} 1
10 {abc[]d]efg*} 1
11 {abc[^]d]efg*} {}
12 {abc[^]XYZ]efg*} 1
} {
do_execsql_test 2.3.$tn {
SELECT rowid FROM t1 WHERE y GLOB $like
} $res
}
do_execsql_test 2.3.null.1 {
SELECT rowid FROM t1 WHERE y LIKE NULL
}
#-------------------------------------------------------------------------
reset_db
do_catchsql_test 3.1 {
CREATE VIRTUAL TABLE ttt USING fts5(c, tokenize="trigram case_sensitive 2");
} {1 {error in tokenizer constructor}}
do_catchsql_test 3.2 {
CREATE VIRTUAL TABLE ttt USING fts5(c, tokenize="trigram case_sensitive 11");
} {1 {error in tokenizer constructor}}
do_catchsql_test 3.3 {
CREATE VIRTUAL TABLE ttt USING fts5(c, "tokenize=trigram case_sensitive 1");
} {0 {}}
#-------------------------------------------------------------------------
reset_db
do_execsql_test 4.0 {
CREATE VIRTUAL TABLE t0 USING fts5(b, tokenize = "trigram");
}
do_execsql_test 4.1 {
INSERT INTO t0 VALUES (x'000b01');
}
do_execsql_test 4.2 {
INSERT INTO t0(t0) VALUES('integrity-check');
}
#-------------------------------------------------------------------------
reset_db
foreach_detail_mode $::testprefix {
foreach {ci} {0 1} {
reset_db
do_execsql_test 5.cs=$ci.0.1 "
CREATE VIRTUAL TABLE t1 USING fts5(
y, tokenize=\"trigram case_sensitive $ci\", detail=%DETAIL%
);
"
do_execsql_test 5.cs=$ci.0.2 {
INSERT INTO t1 VALUES('abcdefghijklm');
INSERT INTO t1 VALUES('กรุงเทพมหานคร');
}
foreach {tn like res} {
1 {%cDef%} 1
2 {cDef%} {}
3 {%f%} 1
4 {%f_h%} 1
5 {%f_g%} {}
6 {abc%klm} 1
7 {ABCDEFG%} 1
8 {%รุงเ%} 2
} {
do_execsql_test 5.cs=$ci.1.$tn {
SELECT rowid FROM t1 WHERE y LIKE $like
} $res
}
}
}
do_execsql_test 6.0 {
CREATE VIRTUAL TABLE ci0 USING fts5(x, tokenize="trigram");
CREATE VIRTUAL TABLE ci1 USING fts5(x, tokenize="trigram case_sensitive 1");
}
# LIKE and GLOB both work with case-insensitive tokenizers. Only GLOB works
# with case-sensitive.
do_eqp_test 6.1 {
SELECT * FROM ci0 WHERE x LIKE ?
} {VIRTUAL TABLE INDEX 0:L0}
do_eqp_test 6.2 {
SELECT * FROM ci0 WHERE x GLOB ?
} {VIRTUAL TABLE INDEX 0:G0}
do_eqp_test 6.3 {
SELECT * FROM ci1 WHERE x LIKE ?
} {{SCAN ci1 VIRTUAL TABLE INDEX 0:}}
do_eqp_test 6.4 {
SELECT * FROM ci1 WHERE x GLOB ?
} {VIRTUAL TABLE INDEX 0:G0}
reset_db
do_execsql_test 7.0 {
CREATE VIRTUAL TABLE f USING FTS5(filename, tokenize="trigram");
INSERT INTO f (rowid, filename) VALUES
(10, "giraffe.png"),
(20, "жираф.png"),
(30, "cat.png"),
(40, "кот.png"),
(50, "misic-🎵-.mp3");
}
do_execsql_test 7.1 {
SELECT rowid FROM f WHERE +filename GLOB '*ир*';
} {20}
do_execsql_test 7.2 {
SELECT rowid FROM f WHERE filename GLOB '*ир*';
} {20}
finish_test