Home | History | Annotate | Line # | Download | only in bigram
locate.bigram.c revision 1.11 
      1 /*	$NetBSD: locate.bigram.c,v 1.11 2008/07/21 14:19:23 lukem Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1989, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * This code is derived from software contributed to Berkeley by
      8  * James A. Woods.
      9  *
     10  * Redistribution and use in source and binary forms, with or without
     11  * modification, are permitted provided that the following conditions
     12  * are met:
     13  * 1. Redistributions of source code must retain the above copyright
     14  *    notice, this list of conditions and the following disclaimer.
     15  * 2. Redistributions in binary form must reproduce the above copyright
     16  *    notice, this list of conditions and the following disclaimer in the
     17  *    documentation and/or other materials provided with the distribution.
     18  * 3. Neither the name of the University nor the names of its contributors
     19  *    may be used to endorse or promote products derived from this software
     20  *    without specific prior written permission.
     21  *
     22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     32  * SUCH DAMAGE.
     33  */
     34 
     35 #include <sys/cdefs.h>
     36 #ifndef lint
     37 __COPYRIGHT("@(#) Copyright (c) 1989, 1993\
     38  The Regents of the University of California.  All rights reserved.");
     39 #endif /* not lint */
     40 
     41 #ifndef lint
     42 #if 0
     43 static char sccsid[] = "@(#)locate.bigram.c	8.2 (Berkeley) 4/28/95";
     44 #endif
     45 __RCSID("$NetBSD: locate.bigram.c,v 1.11 2008/07/21 14:19:23 lukem Exp $");
     46 #endif /* not lint */
     47 
     48 /*
     49  *  bigram < text > bigrams
     50  *
     51  * List bigrams for 'updatedb' script.
     52  * Use 'code' to encode a file using this output.
     53  */
     54 
     55 #include <stdio.h>
     56 #include <stdlib.h>
     57 #include <string.h>
     58 #include <sys/param.h>			/* for MAXPATHLEN */
     59 
     60 int	main __P((int, char **));
     61 static int compare_bigrams __P((const void *, const void *));
     62 static void add_bigram __P((u_char, u_char));
     63 
     64 static char buf1[MAXPATHLEN] = " ";
     65 static char buf2[MAXPATHLEN];
     66 
     67 struct bigram {
     68 	int count;
     69 	u_char b1, b2;	/* needed for final sorting */
     70 };
     71 
     72 struct bigram bigrams[256 * 256];
     73 
     74 static void
     75 add_bigram(i1, i2)
     76 	u_char i1, i2;
     77 {
     78 	if (i1 != '\n' && i2 != '\n')
     79 		bigrams[(i1<<8)+i2].count++;
     80 }
     81 
     82 static int
     83 compare_bigrams(item1, item2)
     84 	const void *item1, *item2;
     85 {
     86 	const struct bigram *it1=item1, *it2=item2;
     87 
     88 	if (it1->count != it2->count)
     89 		return it2->count - it1->count;
     90 	else if (it1->b1 != it2->b1)
     91 		return it2->b1 - it1->b1;
     92 	else
     93 		return it2->b2 - it2->b2;
     94 }
     95 
     96 int
     97 main(argc, argv)
     98 	int argc;
     99 	char *argv[];
    100 {
    101   	char *cp;
    102 	char *oldpath = buf1, *path = buf2;
    103 	struct bigram *bg;
    104 	int i;
    105 
    106 	/* initialize bigram array */
    107 	memset(bigrams, 0, sizeof(bigrams));
    108 	for(i=0; i < 65536; i++) {
    109 		bigrams[i].b1 = i / 256;
    110 		bigrams[i].b2 = i % 256;
    111 	}
    112 
    113      	while ( fgets ( path, sizeof(buf2), stdin ) != NULL ) {
    114 
    115 		/* skip longest common prefix */
    116 		for ( cp = path; *cp == *oldpath; cp++, oldpath++ )
    117 			if ( *oldpath == '\0' )
    118 				break;
    119 
    120 		/*
    121 		 * output post-residue bigrams only
    122 		 */
    123 		for(; cp[0] != '\0' && cp[1] != '\0'; cp += 2)
    124 			add_bigram((u_char)cp[0], (u_char)cp[1]);
    125 
    126 		if (path == buf1)		/* swap pointers */
    127 			path = buf2, oldpath = buf1;
    128 		else
    129 			path = buf1, oldpath = buf2;
    130    	}
    131 
    132 	/* sort the bigrams by how many times it appeared and their value */
    133 	heapsort((void *)bigrams, 256 * 256, sizeof(struct bigram),
    134 			compare_bigrams);
    135 
    136 	/* write 128 most frequent bigrams out */
    137 	bg = bigrams;
    138 	for (i = 0; i < 128 && bg->count > 0; i++, bg++) {
    139 		if (bg->b1 != '\0')
    140 			fputc(bg->b1, stdout);
    141 		if (bg->b2 != '\0')
    142 			fputc(bg->b2, stdout);
    143 	}
    144 
    145 	return (0);
    146 }
    147