fuzzy matching

This commit is contained in:
Luca Bilke 2023-06-17 17:43:39 +02:00
parent 4327d2a6e9
commit 158ee8a8ec
4 changed files with 255 additions and 2 deletions

View File

@ -1,7 +1,8 @@
/* See LICENSE file for copyright and license details. */
/* Default settings; can be overriden by command line. */
static int topbar = 1; /* -b option; if 0, dmenu appears at bottom */
static int topbar = 1; /*-b*/
static int fuzzy = 1; /*-F*/
static const unsigned int fgalpha = 0xffU;
static const unsigned int bgalpha = 0xffU;
/* -fn option overrides fonts[0]; default X11 font or font set */

View File

@ -21,7 +21,7 @@ FREETYPEINC = /usr/include/freetype2
# includes and libs
INCS = -I$(X11INC) -I$(FREETYPEINC)
LIBS = -L$(X11LIB) -lXrender -lX11 $(XINERAMALIBS) $(FREETYPELIBS)
LIBS = -L$(X11LIB) -lXrender -lX11 $(XINERAMALIBS) $(FREETYPELIBS) -lm
# flags
CPPFLAGS = -D_DEFAULT_SOURCE -D_BSD_SOURCE -D_XOPEN_SOURCE=700 -D_POSIX_C_SOURCE=200809L -DVERSION=\"$(VERSION)\" $(XINERAMAFLAGS)

89
dmenu.c
View File

@ -1,6 +1,7 @@
/* See LICENSE file for copyright and license details. */
#include <ctype.h>
#include <locale.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@ -33,6 +34,7 @@ struct item {
char *text;
struct item *left, *right;
int out;
double distance;
};
static char text[BUFSIZ] = "";
@ -241,9 +243,94 @@ grabkeyboard(void)
die("cannot grab keyboard");
}
int
compare_distance(const void *a, const void *b)
{
struct item *da = *(struct item **) a;
struct item *db = *(struct item **) b;
if (!db)
return 1;
if (!da)
return -1;
return da->distance == db->distance ? 0 : da->distance < db->distance ? -1 : 1;
}
void
fuzzymatch(void)
{
/* bang - we have so much memory */
struct item *it;
struct item **fuzzymatches = NULL;
char c;
int number_of_matches = 0, i, pidx, sidx, eidx;
int text_len = strlen(text), itext_len;
matches = matchend = NULL;
/* walk through all items */
for (it = items; it && it->text; it++) {
if (text_len) {
itext_len = strlen(it->text);
pidx = 0; /* pointer */
sidx = eidx = -1; /* start of match, end of match */
/* walk through item text */
for (i = 0; i < itext_len && (c = it->text[i]); i++) {
/* fuzzy match pattern */
if (!fstrncmp(&text[pidx], &c, 1)) {
if(sidx == -1)
sidx = i;
pidx++;
if (pidx == text_len) {
eidx = i;
break;
}
}
}
/* build list of matches */
if (eidx != -1) {
/* compute distance */
/* add penalty if match starts late (log(sidx+2))
* add penalty for long a match without many matching characters */
it->distance = log(sidx + 2) + (double)(eidx - sidx - text_len);
/* fprintf(stderr, "distance %s %f\n", it->text, it->distance); */
appenditem(it, &matches, &matchend);
number_of_matches++;
}
} else {
appenditem(it, &matches, &matchend);
}
}
if (number_of_matches) {
/* initialize array with matches */
if (!(fuzzymatches = realloc(fuzzymatches, number_of_matches * sizeof(struct item*))))
die("cannot realloc %u bytes:", number_of_matches * sizeof(struct item*));
for (i = 0, it = matches; it && i < number_of_matches; i++, it = it->right) {
fuzzymatches[i] = it;
}
/* sort matches according to distance */
qsort(fuzzymatches, number_of_matches, sizeof(struct item*), compare_distance);
/* rebuild list of matches */
matches = matchend = NULL;
for (i = 0, it = fuzzymatches[i]; i < number_of_matches && it && \
it->text; i++, it = fuzzymatches[i]) {
appenditem(it, &matches, &matchend);
}
free(fuzzymatches);
}
curr = sel = matches;
calcoffsets();
}
static void
match(void)
{
if (fuzzy) {
fuzzymatch();
return;
}
static char **tokv = NULL;
static int tokn = 0;
@ -769,6 +856,8 @@ main(int argc, char *argv[])
topbar = 0;
else if (!strcmp(argv[i], "-f")) /* grabs keyboard before reading stdin */
fast = 1;
else if (!strcmp(argv[i], "-F")) /* grabs keyboard before reading stdin */
fuzzy = 0;
else if (!strcmp(argv[i], "-i")) { /* case-insensitive item matching */
fstrncmp = strncasecmp;
fstrstr = cistrstr;

View File

@ -0,0 +1,163 @@
From 94353eb52055927d9079f3d9e33da1c954abf386 Mon Sep 17 00:00:00 2001
From: aleks <aleks.stier@icloud.com>
Date: Wed, 26 Jun 2019 13:25:10 +0200
Subject: [PATCH] Add support for fuzzy-matching
---
config.def.h | 1 +
config.mk | 2 +-
dmenu.c | 89 ++++++++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 91 insertions(+), 1 deletion(-)
diff --git a/config.def.h b/config.def.h
index 1edb647..51612b9 100644
--- a/config.def.h
+++ b/config.def.h
@@ -2,6 +2,7 @@
/* Default settings; can be overriden by command line. */
static int topbar = 1; /* -b option; if 0, dmenu appears at bottom */
+static int fuzzy = 1; /* -F option; if 0, dmenu doesn't use fuzzy matching */
/* -fn option overrides fonts[0]; default X11 font or font set */
static const char *fonts[] = {
"monospace:size=10"
diff --git a/config.mk b/config.mk
index 0929b4a..d14309a 100644
--- a/config.mk
+++ b/config.mk
@@ -20,7 +20,7 @@ FREETYPEINC = /usr/include/freetype2
# includes and libs
INCS = -I$(X11INC) -I$(FREETYPEINC)
-LIBS = -L$(X11LIB) -lX11 $(XINERAMALIBS) $(FREETYPELIBS)
+LIBS = -L$(X11LIB) -lX11 $(XINERAMALIBS) $(FREETYPELIBS) -lm
# flags
CPPFLAGS = -D_DEFAULT_SOURCE -D_BSD_SOURCE -D_XOPEN_SOURCE=700 -D_POSIX_C_SOURCE=200809L -DVERSION=\"$(VERSION)\" $(XINERAMAFLAGS)
diff --git a/dmenu.c b/dmenu.c
index 6b8f51b..96ddc98 100644
--- a/dmenu.c
+++ b/dmenu.c
@@ -1,6 +1,7 @@
/* See LICENSE file for copyright and license details. */
#include <ctype.h>
#include <locale.h>
+#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
@@ -32,6 +33,7 @@ struct item {
char *text;
struct item *left, *right;
int out;
+ double distance;
};
static char text[BUFSIZ] = "";
@@ -210,9 +212,94 @@ grabkeyboard(void)
die("cannot grab keyboard");
}
+int
+compare_distance(const void *a, const void *b)
+{
+ struct item *da = *(struct item **) a;
+ struct item *db = *(struct item **) b;
+
+ if (!db)
+ return 1;
+ if (!da)
+ return -1;
+
+ return da->distance == db->distance ? 0 : da->distance < db->distance ? -1 : 1;
+}
+
+void
+fuzzymatch(void)
+{
+ /* bang - we have so much memory */
+ struct item *it;
+ struct item **fuzzymatches = NULL;
+ char c;
+ int number_of_matches = 0, i, pidx, sidx, eidx;
+ int text_len = strlen(text), itext_len;
+
+ matches = matchend = NULL;
+
+ /* walk through all items */
+ for (it = items; it && it->text; it++) {
+ if (text_len) {
+ itext_len = strlen(it->text);
+ pidx = 0; /* pointer */
+ sidx = eidx = -1; /* start of match, end of match */
+ /* walk through item text */
+ for (i = 0; i < itext_len && (c = it->text[i]); i++) {
+ /* fuzzy match pattern */
+ if (!fstrncmp(&text[pidx], &c, 1)) {
+ if(sidx == -1)
+ sidx = i;
+ pidx++;
+ if (pidx == text_len) {
+ eidx = i;
+ break;
+ }
+ }
+ }
+ /* build list of matches */
+ if (eidx != -1) {
+ /* compute distance */
+ /* add penalty if match starts late (log(sidx+2))
+ * add penalty for long a match without many matching characters */
+ it->distance = log(sidx + 2) + (double)(eidx - sidx - text_len);
+ /* fprintf(stderr, "distance %s %f\n", it->text, it->distance); */
+ appenditem(it, &matches, &matchend);
+ number_of_matches++;
+ }
+ } else {
+ appenditem(it, &matches, &matchend);
+ }
+ }
+
+ if (number_of_matches) {
+ /* initialize array with matches */
+ if (!(fuzzymatches = realloc(fuzzymatches, number_of_matches * sizeof(struct item*))))
+ die("cannot realloc %u bytes:", number_of_matches * sizeof(struct item*));
+ for (i = 0, it = matches; it && i < number_of_matches; i++, it = it->right) {
+ fuzzymatches[i] = it;
+ }
+ /* sort matches according to distance */
+ qsort(fuzzymatches, number_of_matches, sizeof(struct item*), compare_distance);
+ /* rebuild list of matches */
+ matches = matchend = NULL;
+ for (i = 0, it = fuzzymatches[i]; i < number_of_matches && it && \
+ it->text; i++, it = fuzzymatches[i]) {
+ appenditem(it, &matches, &matchend);
+ }
+ free(fuzzymatches);
+ }
+ curr = sel = matches;
+ calcoffsets();
+}
+
static void
match(void)
{
+ if (fuzzy) {
+ fuzzymatch();
+ return;
+ }
static char **tokv = NULL;
static int tokn = 0;
@@ -702,6 +789,8 @@ main(int argc, char *argv[])
topbar = 0;
else if (!strcmp(argv[i], "-f")) /* grabs keyboard before reading stdin */
fast = 1;
+ else if (!strcmp(argv[i], "-F")) /* grabs keyboard before reading stdin */
+ fuzzy = 0;
else if (!strcmp(argv[i], "-i")) { /* case-insensitive item matching */
fstrncmp = strncasecmp;
fstrstr = cistrstr;
--
2.22.0