apunkt/sqlite-vec
1
0
Fork 0
mirror of https://github.com/asg017/sqlite-vec.git synced 2026-04-25 08:46:49 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

text knn GT/GE fixes

Browse source
This commit is contained in:
Alex Garcia 2024-11-18 14:05:47 -08:00
parent 1ec1b89f60
commit df29e31ddc
3 changed files with 318 additions and 30 deletions
Download patch file Download diff file Expand all files Collapse all files

49
sqlite-vec.c
View file

@ -5968,19 +5968,19 @@ int vec0_metadata_filter_text(vec0_vtab * p, sqlite3_value * value, const void *
view = &((u8*) buffer)[i * VEC0_METADATA_TEXT_VIEW_BUFFER_LENGTH]; view = &((u8*) buffer)[i * VEC0_METADATA_TEXT_VIEW_BUFFER_LENGTH];
nPrefix = ((int*) view)[0]; nPrefix = ((int*) view)[0];
sPrefix = (char *) &view[4]; sPrefix = (char *) &view[4];
int cmpPrefix = strncmp(sPrefix, sTarget, min(nPrefix, VEC0_METADATA_TEXT_VIEW_DATA_LENGTH)); int cmpPrefix = strncmp(sPrefix, sTarget, min(min(nPrefix, VEC0_METADATA_TEXT_VIEW_DATA_LENGTH), nTarget));
// for short strings, use the prefix comparison direclty if(nPrefix < VEC0_METADATA_TEXT_VIEW_DATA_LENGTH) {
if(nPrefix <= VEC0_METADATA_TEXT_VIEW_DATA_LENGTH) { // if prefix match, check which is longer
bitmap_set(b, i, cmpPrefix > 0); if(cmpPrefix == 0) {
continue; bitmap_set(b, i, nPrefix > nTarget);
} }
else {
// for GT, only need to consult full string if EQ
if(cmpPrefix != 0) {
bitmap_set(b, i, cmpPrefix > 0); bitmap_set(b, i, cmpPrefix > 0);
}
continue; continue;
} }
// TODO(perf): may not need to compare full text in some cases
rc = vec0_get_metadata_text_long_value(p, &stmt, metadata_idx, rowids[i], &nFull, &sFull); rc = vec0_get_metadata_text_long_value(p, &stmt, metadata_idx, rowids[i], &nFull, &sFull);
if(rc != SQLITE_OK) { if(rc != SQLITE_OK) {
@ -5996,11 +5996,32 @@ int vec0_metadata_filter_text(vec0_vtab * p, sqlite3_value * value, const void *
} }
case VEC0_METADATA_OPERATOR_GE: { case VEC0_METADATA_OPERATOR_GE: {
for(int i = 0; i < size; i++) { for(int i = 0; i < size; i++) {
u8 * view = &((u8*) buffer)[i * VEC0_METADATA_TEXT_VIEW_BUFFER_LENGTH]; view = &((u8*) buffer)[i * VEC0_METADATA_TEXT_VIEW_BUFFER_LENGTH];
int n = ((int*) view)[0]; nPrefix = ((int*) view)[0];
char * s = (char *) &view[4]; sPrefix = (char *) &view[4];
if(n > VEC0_METADATA_TEXT_VIEW_DATA_LENGTH) {rc = SQLITE_ERROR;goto done;} /* TODO */ int cmpPrefix = strncmp(sPrefix, sTarget, min(min(nPrefix, VEC0_METADATA_TEXT_VIEW_DATA_LENGTH), nTarget));
bitmap_set(b, i, strncmp(s, sTarget, n) >= 0);
if(nPrefix < VEC0_METADATA_TEXT_VIEW_DATA_LENGTH) {
// if prefix match, check which is longer
if(cmpPrefix == 0) {
bitmap_set(b, i, nPrefix >= nTarget);
}
else {
bitmap_set(b, i, cmpPrefix >= 0);
}
continue;
}
// TODO(perf): may not need to compare full text in some cases
rc = vec0_get_metadata_text_long_value(p, &stmt, metadata_idx, rowids[i], &nFull, &sFull);
if(rc != SQLITE_OK) {
goto done;
}
if(nPrefix != nFull) {
rc = SQLITE_ERROR;
goto done;
}
bitmap_set(b, i, strncmp(sFull, sTarget, nFull) >= 0);
} }
break; break;
} }

295
tests/__snapshots__/test-metadata.ambr
View file

@ -617,6 +617,13 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[eq-bb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?",
'rows': list([
]),
})
# ---
# name: test_long_text_knn[eq-bbbb] # name: test_long_text_knn[eq-bbbb]
OrderedDict({ OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?", 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?",
@ -629,6 +636,13 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[eq-bbbbbb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?",
'rows': list([
]),
})
# ---
# name: test_long_text_knn[eq-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[eq-bbbbbbbbbbbb_aaa]
OrderedDict({ OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?", 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name = ?",
@ -662,34 +676,175 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[ge-bb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
OrderedDict({
'rowid': 3,
'name': 'bbbb',
'distance': 97.0,
}),
]),
})
# ---
# name: test_long_text_knn[ge-bbbb] # name: test_long_text_knn[ge-bbbb]
dict({ OrderedDict({
'error': 'OperationalError', 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'message': 'Could not filter metadata fields', 'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
OrderedDict({
'rowid': 3,
'name': 'bbbb',
'distance': 97.0,
}),
]),
})
# ---
# name: test_long_text_knn[ge-bbbbbb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
]),
}) })
# --- # ---
# name: test_long_text_knn[ge-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[ge-bbbbbbbbbbbb_aaa]
dict({ OrderedDict({
'error': 'OperationalError', 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'message': 'Could not filter metadata fields', 'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
]),
}) })
# --- # ---
# name: test_long_text_knn[ge-bbbbbbbbbbbb_bbb] # name: test_long_text_knn[ge-bbbbbbbbbbbb_bbb]
dict({ OrderedDict({
'error': 'OperationalError', 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'message': 'Could not filter metadata fields', 'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
]),
}) })
# --- # ---
# name: test_long_text_knn[ge-bbbbbbbbbbbb_ccc] # name: test_long_text_knn[ge-bbbbbbbbbbbb_ccc]
dict({ OrderedDict({
'error': 'OperationalError', 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'message': 'Could not filter metadata fields', 'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
]),
}) })
# --- # ---
# name: test_long_text_knn[ge-longlonglonglonglonglonglong] # name: test_long_text_knn[ge-longlonglonglonglonglonglong]
dict({ OrderedDict({
'error': 'OperationalError', 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name >= ?",
'message': 'Could not filter metadata fields', 'rows': list([
]),
})
# ---
# name: test_long_text_knn[gt-bb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name > ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
OrderedDict({
'rowid': 3,
'name': 'bbbb',
'distance': 97.0,
}),
]),
}) })
# --- # ---
# name: test_long_text_knn[gt-bbbb] # name: test_long_text_knn[gt-bbbb]
@ -714,6 +869,28 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[gt-bbbbbb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name > ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
]),
})
# ---
# name: test_long_text_knn[gt-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[gt-bbbbbbbbbbbb_aaa]
OrderedDict({ OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name > ?", 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name > ?",
@ -777,12 +954,24 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[le-bb]
dict({
'error': 'OperationalError',
'message': 'Could not filter metadata fields',
})
# ---
# name: test_long_text_knn[le-bbbb] # name: test_long_text_knn[le-bbbb]
dict({ dict({
'error': 'OperationalError', 'error': 'OperationalError',
'message': 'Could not filter metadata fields', 'message': 'Could not filter metadata fields',
}) })
# --- # ---
# name: test_long_text_knn[le-bbbbbb]
dict({
'error': 'OperationalError',
'message': 'Could not filter metadata fields',
})
# ---
# name: test_long_text_knn[le-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[le-bbbbbbbbbbbb_aaa]
dict({ dict({
'error': 'OperationalError', 'error': 'OperationalError',
@ -807,12 +996,24 @@
'message': 'Could not filter metadata fields', 'message': 'Could not filter metadata fields',
}) })
# --- # ---
# name: test_long_text_knn[lt-bb]
dict({
'error': 'OperationalError',
'message': 'Could not filter metadata fields',
})
# ---
# name: test_long_text_knn[lt-bbbb] # name: test_long_text_knn[lt-bbbb]
dict({ dict({
'error': 'OperationalError', 'error': 'OperationalError',
'message': 'Could not filter metadata fields', 'message': 'Could not filter metadata fields',
}) })
# --- # ---
# name: test_long_text_knn[lt-bbbbbb]
dict({
'error': 'OperationalError',
'message': 'Could not filter metadata fields',
})
# ---
# name: test_long_text_knn[lt-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[lt-bbbbbbbbbbbb_aaa]
dict({ dict({
'error': 'OperationalError', 'error': 'OperationalError',
@ -837,6 +1038,38 @@
'message': 'Could not filter metadata fields', 'message': 'Could not filter metadata fields',
}) })
# --- # ---
# name: test_long_text_knn[ne-bb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
OrderedDict({
'rowid': 3,
'name': 'bbbb',
'distance': 97.0,
}),
OrderedDict({
'rowid': 2,
'name': 'aaaaaaaaaaaa_aaa',
'distance': 98.0,
}),
]),
})
# ---
# name: test_long_text_knn[ne-bbbb] # name: test_long_text_knn[ne-bbbb]
OrderedDict({ OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?", 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?",
@ -869,6 +1102,38 @@
]), ]),
}) })
# --- # ---
# name: test_long_text_knn[ne-bbbbbb]
OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?",
'rows': list([
OrderedDict({
'rowid': 6,
'name': 'cccccccccccc_ccc',
'distance': 94.0,
}),
OrderedDict({
'rowid': 5,
'name': 'cccc',
'distance': 95.0,
}),
OrderedDict({
'rowid': 4,
'name': 'bbbbbbbbbbbb_bbb',
'distance': 96.0,
}),
OrderedDict({
'rowid': 3,
'name': 'bbbb',
'distance': 97.0,
}),
OrderedDict({
'rowid': 2,
'name': 'aaaaaaaaaaaa_aaa',
'distance': 98.0,
}),
]),
})
# ---
# name: test_long_text_knn[ne-bbbbbbbbbbbb_aaa] # name: test_long_text_knn[ne-bbbbbbbbbbbb_aaa]
OrderedDict({ OrderedDict({
'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?", 'sql': "select rowid, name, distance from v where vector match '[100]' and k = 5 and name != ?",

2
tests/test-metadata.py
View file

@ -147,6 +147,8 @@ def test_long_text_knn(db, snapshot):
tests = [ tests = [
"bbbb", "bbbb",
"bb",
"bbbbbb",
"bbbbbbbbbbbb_bbb", "bbbbbbbbbbbb_bbb",
"bbbbbbbbbbbb_aaa", "bbbbbbbbbbbb_aaa",
"bbbbbbbbbbbb_ccc", "bbbbbbbbbbbb_ccc",