Changeset: 3a64271e8751 for MonetDB
URL: https://dev.monetdb.org/hg/MonetDB/rev/3a64271e8751
Removed Files:
sql/test/BugTracker-2020/Tests/txtsim-parallel.Bug-7016.test
Modified Files:
clients/Tests/MAL-signatures-hge.test
clients/Tests/MAL-signatures.test
monetdb5/modules/mal/txtsim.c
sql/scripts/48_txtsim.sql
sql/test/BugTracker-2016/Tests/DISTINCT_with_correlated_scalar_subquery_crashes_mserver.Bug-3920.test
sql/test/emptydb-previous-upgrade-chain-hge/Tests/upgrade.stable.out.int128
sql/test/emptydb-previous-upgrade-chain/Tests/upgrade.stable.out
sql/test/emptydb-previous-upgrade-chain/Tests/upgrade.stable.out.int128
sql/test/emptydb-previous-upgrade-hge/Tests/upgrade.stable.out.int128
sql/test/emptydb-previous-upgrade/Tests/upgrade.stable.out
sql/test/emptydb-previous-upgrade/Tests/upgrade.stable.out.int128
sql/test/emptydb-upgrade-chain-hge/Tests/upgrade.stable.out.int128
sql/test/emptydb-upgrade-chain/Tests/upgrade.stable.out
sql/test/emptydb-upgrade-chain/Tests/upgrade.stable.out.int128
sql/test/emptydb-upgrade-chain/Tests/upgrade.stable.out.ppc64.int128
sql/test/emptydb-upgrade-hge/Tests/upgrade.stable.out.int128
sql/test/emptydb-upgrade/Tests/upgrade.stable.out
sql/test/emptydb-upgrade/Tests/upgrade.stable.out.int128
sql/test/emptydb/Tests/check.stable.out
sql/test/emptydb/Tests/check.stable.out.32bit
sql/test/emptydb/Tests/check.stable.out.int128
sql/test/miscellaneous/Tests/simple_selects.test
sql/test/testdb-previous-upgrade-chain-hge/Tests/upgrade.stable.out.int128
sql/test/testdb-previous-upgrade-chain/Tests/upgrade.stable.out
sql/test/testdb-previous-upgrade-chain/Tests/upgrade.stable.out.int128
sql/test/testdb-previous-upgrade-hge/Tests/upgrade.stable.out.int128
sql/test/testdb-previous-upgrade/Tests/upgrade.stable.out
sql/test/testdb-previous-upgrade/Tests/upgrade.stable.out.int128
sql/test/testdb-upgrade-chain-hge/Tests/upgrade.stable.out.int128
sql/test/testdb-upgrade-chain/Tests/upgrade.stable.out
sql/test/testdb-upgrade-chain/Tests/upgrade.stable.out.int128
sql/test/testdb-upgrade-hge/Tests/upgrade.stable.out.int128
sql/test/testdb-upgrade/Tests/upgrade.stable.out
sql/test/testdb-upgrade/Tests/upgrade.stable.out.int128
Branch: default
Log Message:
Remove similarity functionality (DEPRECATED in Jun2023).
diffs (truncated from 1255 to 300 lines):
diff --git a/clients/Tests/MAL-signatures-hge.test
b/clients/Tests/MAL-signatures-hge.test
--- a/clients/Tests/MAL-signatures-hge.test
+++ b/clients/Tests/MAL-signatures-hge.test
@@ -35568,11 +35568,6 @@ maxlevenshtein
pattern battxtsim.maxlevenshtein(X_0:bat[:str], X_1:bat[:str], X_2:int,
X_3:int, X_4:int):bat[:bit]
BATTXTSIMmaxlevenshtein;
Same as maxlevenshtein but for BATS
-battxtsim
-similarity
-command battxtsim.similarity(X_0:bat[:str], X_1:bat[:str]):bat[:dbl]
-fstrcmp0_impl_bulk;
-(Deprecated) Normalized edit distance between two strings
baturl
extractURLHost
command baturl.extractURLHost(X_0:bat[:str], X_1:bit):bat[:str]
@@ -51179,16 +51174,6 @@ command txtsim.qgramselfjoin(X_0:bat[:oi
qgram_selfjoin;
QGram self-join on ordered(!) qgram tables and sub-ordered q-gram positions
txtsim
-similarity
-command txtsim.similarity(X_0:str, X_1:str):dbl
-fstrcmp0_impl;
-(Deprecated) Normalized edit distance between two strings
-txtsim
-similarity
-command txtsim.similarity(X_0:str, X_1:str, X_2:dbl):dbl
-fstrcmp_impl;
-(Deprecated) Normalized edit distance between two strings
-txtsim
soundex
command txtsim.soundex(X_0:str):str
soundex;
diff --git a/clients/Tests/MAL-signatures.test
b/clients/Tests/MAL-signatures.test
--- a/clients/Tests/MAL-signatures.test
+++ b/clients/Tests/MAL-signatures.test
@@ -26618,11 +26618,6 @@ maxlevenshtein
pattern battxtsim.maxlevenshtein(X_0:bat[:str], X_1:bat[:str], X_2:int,
X_3:int, X_4:int):bat[:bit]
BATTXTSIMmaxlevenshtein;
Same as maxlevenshtein but for BATS
-battxtsim
-similarity
-command battxtsim.similarity(X_0:bat[:str], X_1:bat[:str]):bat[:dbl]
-fstrcmp0_impl_bulk;
-(Deprecated) Normalized edit distance between two strings
baturl
extractURLHost
command baturl.extractURLHost(X_0:bat[:str], X_1:bit):bat[:str]
@@ -39504,16 +39499,6 @@ command txtsim.qgramselfjoin(X_0:bat[:oi
qgram_selfjoin;
QGram self-join on ordered(!) qgram tables and sub-ordered q-gram positions
txtsim
-similarity
-command txtsim.similarity(X_0:str, X_1:str):dbl
-fstrcmp0_impl;
-(Deprecated) Normalized edit distance between two strings
-txtsim
-similarity
-command txtsim.similarity(X_0:str, X_1:str, X_2:dbl):dbl
-fstrcmp_impl;
-(Deprecated) Normalized edit distance between two strings
-txtsim
soundex
command txtsim.soundex(X_0:str):str
soundex;
diff --git a/monetdb5/modules/mal/txtsim.c b/monetdb5/modules/mal/txtsim.c
--- a/monetdb5/modules/mal/txtsim.c
+++ b/monetdb5/modules/mal/txtsim.c
@@ -1391,484 +1391,6 @@ str_2_qgrams(bat *ret, str *val)
return MAL_SUCCEED;
}
-/* DEPRECATED (see DEPRECATED_END) */
-#define INITIAL_INT_BUFFER_LENGTH 2048
-#define CHECK_INT_BUFFER_LENGTH(BUFFER, BUFFER_LEN, NEXT_LEN, OP)
\
- do {
\
- if ((NEXT_LEN) > *BUFFER_LEN) {
\
- size_t newlen = (((NEXT_LEN) + 1023) & ~1023); /* align
to a multiple of 1024 bytes */ \
- int *newbuf = GDKmalloc(newlen);
\
- if (!newbuf)
\
- throw(MAL, OP, SQLSTATE(HY013)
MAL_MALLOC_FAIL); \
- GDKfree(*BUFFER);
\
- *BUFFER = newbuf;
\
- *BUFFER_LEN = newlen;
\
- }
\
- } while (0)
-
-struct string_data {
- /* The string to be compared. */
- const char *data;
-
- /* The length of the string to be compared. */
- int data_length;
-
- /* The number of characters inserted or deleted. */
- int edit_count;
-};
-
-struct partition {
- /* Midpoints of this partition. */
- int xmid, ymid;
-
- /* Nonzero if low half will be analyzed minimally. */
- int lo_minimal;
-
- /* Likewise for high half. */
- int hi_minimal;
-};
-
-/* NAME
- diag - find diagonal path
-
- SYNOPSIS
- int diag(int xoff, int xlim, int yoff, int ylim, int minimal,
- struct partition *part);
-
- DESCRIPTION
- Find the midpoint of the shortest edit script for a specified
- portion of the two strings.
-
- Scan from the beginnings of the strings, and simultaneously from
- the ends, doing a breadth-first search through the space of
- edit-sequence. When the two searches meet, we have found the
- midpoint of the shortest edit sequence.
-
- If MINIMAL is nonzero, find the minimal edit script regardless
- of expense. Otherwise, if the search is too expensive, use
- heuristics to stop the search and report a suboptimal answer.
-
- RETURNS
- Set PART->(XMID,YMID) to the midpoint (XMID,YMID). The diagonal
- number XMID - YMID equals the number of inserted characters
- minus the number of deleted characters (counting only characters
- before the midpoint). Return the approximate edit cost; this is
- the total number of characters inserted or deleted (counting
- only characters before the midpoint), unless a heuristic is used
- to terminate the search prematurely.
-
- Set PART->LEFT_MINIMAL to nonzero iff the minimal edit script
- for the left half of the partition is known; similarly for
- PART->RIGHT_MINIMAL.
-
- CAVEAT
- This function assumes that the first characters of the specified
- portions of the two strings do not match, and likewise that the
- last characters do not match. The caller must trim matching
- characters from the beginning and end of the portions it is
- going to specify.
-
- If we return the "wrong" partitions, the worst this can do is
- cause suboptimal diff output. It cannot cause incorrect diff
- output. */
-
-static inline int
-diag(int xoff, int xlim, int yoff, int ylim, int minimal, struct partition
*part, int too_expensive, struct string_data *string, int *fdiag, int *bdiag)
-{
- int *const fd = fdiag; /* Give the compiler a chance. */
- int *const bd = bdiag; /* Additional help for the compiler. */
- const char *const xv = string[0].data; /* Still more help for the
compiler. */
- const char *const yv = string[1].data; /* And more and more . . . */
- const int dmin = xoff - ylim; /* Minimum valid diagonal. */
- const int dmax = xlim - yoff; /* Maximum valid diagonal. */
- const int fmid = xoff - yoff; /* Center diagonal of top-down search.
*/
- const int bmid = xlim - ylim; /* Center diagonal of bottom-up search.
*/
- int fmin = fmid;
- int fmax = fmid; /* Limits of top-down search. */
- int bmin = bmid;
- int bmax = bmid; /* Limits of bottom-up search. */
- int c; /* Cost. */
- int odd = (fmid - bmid) & 1;
-
- /*
- * True if southeast corner is on an odd diagonal with respect
- * to the northwest.
- */
- fd[fmid] = xoff;
- bd[bmid] = xlim;
- for (c = 1;; ++c) {
- int d; /* Active diagonal. */
-
- /* Extend the top-down search by an edit step in each diagonal.
*/
- if (fmin > dmin)
- fd[--fmin - 1] = -1;
- else
- ++fmin;
- if (fmax < dmax)
- fd[++fmax + 1] = -1;
- else
- --fmax;
- for (d = fmax; d >= fmin; d -= 2) {
- int x;
- int y;
- int tlo;
- int thi;
-
- tlo = fd[d - 1], thi = fd[d + 1];
-
- if (tlo >= thi)
- x = tlo + 1;
- else
- x = thi;
- y = x - d;
- while (x < xlim && y < ylim && xv[x] == yv[y]) {
- ++x;
- ++y;
- }
- fd[d] = x;
- if (odd && bmin <= d && d <= bmax && bd[d] <= x) {
- part->xmid = x;
- part->ymid = y;
- part->lo_minimal = part->hi_minimal = 1;
- return 2 * c - 1;
- }
- }
- /* Similarly extend the bottom-up search. */
- if (bmin > dmin)
- bd[--bmin - 1] = INT_MAX;
- else
- ++bmin;
- if (bmax < dmax)
- bd[++bmax + 1] = INT_MAX;
- else
- --bmax;
- for (d = bmax; d >= bmin; d -= 2) {
- int x;
- int y;
- int tlo;
- int thi;
-
- tlo = bd[d - 1], thi = bd[d + 1];
- if (tlo < thi)
- x = tlo;
- else
- x = thi - 1;
- y = x - d;
- while (x > xoff && y > yoff && xv[x - 1] == yv[y - 1]) {
- --x;
- --y;
- }
- bd[d] = x;
- if (!odd && fmin <= d && d <= fmax && x <= fd[d]) {
- part->xmid = x;
- part->ymid = y;
- part->lo_minimal = part->hi_minimal = 1;
- return 2 * c;
- }
- }
-
- if (minimal)
- continue;
-
- /* Heuristic: if we've gone well beyond the call of duty, give
up
- and report halfway between our best results so far. */
- if (c >= too_expensive) {
- int fxybest;
- int fxbest;
- int bxybest;
- int bxbest;
-
- /* Pacify `gcc -Wall'. */
- fxbest = 0;
- bxbest = 0;
-
- /* Find forward diagonal that maximizes X + Y. */
- fxybest = -1;
- for (d = fmax; d >= fmin; d -= 2) {
- int x;
- int y;
-
- x = fd[d] < xlim ? fd[d] : xlim;
- y = x - d;
-
- if (ylim < y) {
- x = ylim + d;
- y = ylim;
- }
- if (fxybest < x + y) {
- fxybest = x + y;
- fxbest = x;
- }
- }
- /* Find backward diagonal that minimizes X + Y. */
- bxybest = INT_MAX;
- for (d = bmax; d >= bmin; d -= 2) {
- int x;
- int y;
-
- x = xoff > bd[d] ? xoff : bd[d];
- y = x - d;
-
- if (y < yoff) {
- x = yoff + d;
- y = yoff;
- }
- if (x + y < bxybest) {
- bxybest = x + y;
- bxbest = x;
- }
- }
- /* Use the better of the two diagonals. */
- if ((xlim + ylim) - bxybest < fxybest - (xoff + yoff)) {
_______________________________________________
checkin-list mailing list -- checkin-list@monetdb.org
To unsubscribe send an email to checkin-list-le...@monetdb.org
