Support for Roget's thesaurus from project Gutenberg.
	
Datafiles are (temporarly) in /n/sources/steve/roget,
roget and rogetindex are the data and index respectively.
/sys/src/cmd/dict/mkroget will rebuild build these
from roget-body.rtf (with the help of rtf2txt).
	
I suggest the data and index should go in /n/sources/roget
similar to /n/sources/pgw.
	
-Steve

Reference: /n/sources/patch/applied/roget
Date: Fri Nov 19 21:00:32 CET 2004

--- /sys/src/cmd/dict/utils.c	Fri Nov 19 21:00:32 2004
+++ /sys/src/cmd/dict/utils.c	Fri Nov 19 21:00:32 2004
@@ -16,6 +16,9 @@
 	{"thesaurus",	"Collins Thesaurus",
 	 "/lib/dict/thesaurus",	"/lib/dict/thesindex",
 	 thesnextoff,	thesprintentry,	thesprintkey},
+	{"roget",		"Project Gutenberg Roget's Thesaurus",
+	 "/lib/dict/roget", "/lib/dict/rogetindex",
+	 rogetnextoff,	rogetprintentry,	rogetprintkey},
 
 	{"ce",		"Gendai Chinese->English",
 	 "/lib/dict/world/sansdata/sandic24.dat",
--- /sys/src/cmd/dict/dict.h	Fri Nov 19 21:00:32 2004
+++ /sys/src/cmd/dict/dict.h	Fri Nov 19 21:00:32 2004
@@ -130,6 +130,9 @@
 long	pgwnextoff(long);
 void	pgwprintentry(Entry,int);
 void	pgwprintkey(void);
+void	rogetprintentry(Entry, int);
+long	rogetnextoff(long);
+void	rogetprintkey(void);
 long	slangnextoff(long);
 void	slangprintentry(Entry, int);
 void	slangprintkey(void);
--- /sys/src/cmd/dict/roget.c	Thu Jan  1 00:00:00 1970
+++ /sys/src/cmd/dict/roget.c	Fri Nov 19 21:00:32 2004
@@ -0,0 +1,147 @@
+#include <u.h>
+#include <libc.h>
+#include <bio.h>
+#include <ctype.h>
+#include "dict.h"
+
+/* Roget's Thesaurus from project Gutenberg */
+
+static long Last = 0;
+
+void
+rogetprintentry(Entry e, int cmd)
+{
+	int spc;
+	char c, *p;
+
+	spc = 0;
+	p = e.start;
+
+	if(cmd == 'h'){
+		while(!isspace(*p) && p < e.end)
+			p++;
+		while(strncmp(p, " -- ", 4) != 0 && p < e.end){
+			while(isspace(*p) && p < e.end)
+				p++;
+			if (*p == '[' || *p == '{'){	
+				c = (*p == '[')? ']': '}';
+				while(*p != c && p < e.end)
+					p++;
+				p++;
+				continue;
+			}
+			if (isdigit(*p) || ispunct(*p)){
+				while(!isspace(*p) && p < e.end)
+					p++;
+				continue;
+			}
+
+
+			if (isspace(*p))
+				spc = 1;
+			else
+			if (spc){
+				outchar(' ');
+				spc = 0;
+			}
+
+			while(!isspace(*p) && p < e.end)
+				outchar(*p++);
+		}
+		return;
+	}	
+
+	while(p < e.end && !isspace(*p))
+		p++;
+	while(p < e.end && isspace(*p))
+		p++;
+
+	while (p < e.end){
+		if (p < e.end -4 && strncmp(p, " -- ", 4) == 0){	/* first line */
+			outnl(2);
+			p += 4;
+			spc = 0;
+		}
+			
+		if (p < e.end -2 && strncmp(p, "[ ", 4) == 0){		/* twiddle layout */
+			outchars(" [");
+			continue;
+		}
+
+		if (p < e.end -4 && strncmp(p, "&c (", 4) == 0){	/* usefull xref */
+			if (spc)
+				outchar(' ');
+			outchar('/');
+			while(p < e.end && *p != '(')
+				p++;
+			p++;
+			while(p < e.end && *p != ')')
+				outchar(*p++);
+			p++;
+			while(p < e.end && isspace(*p))
+				p++;
+			while(p < e.end && isdigit(*p))
+				p++;
+			outchar('/');
+			continue;
+		}
+
+		if (p < e.end -3 && strncmp(p, "&c ", 3) == 0){		/* less usefull xref */
+			while(p < e.end && !isdigit(*p))
+				p++;
+			while(p < e.end && isdigit(*p))
+				p++;
+			continue;
+		}
+
+		if (*p == '\n' && p < (e.end -1)){			/* their newlines */
+			spc = 0;
+			p++;
+			if (isspace(*p)){				/* their continuation line */
+				while (isspace(*p))
+					p++;
+				p--;
+			}
+			else{
+				outnl(2);
+			}
+		}
+		if (spc && *p != ';' && *p != '.' &&
+		    *p != ',' && !isspace(*p)){				/* drop spaces before punct */
+			spc = 0;
+			outchar(' ');
+		}
+		if (isspace(*p))
+			spc = 1;
+		else
+			outchar(*p);
+		p++;
+	}
+	outnl(0);
+}
+
+long
+rogetnextoff(long fromoff)
+{
+	int i;
+	vlong l;
+	char *p;
+
+	Bseek(bdict, fromoff, 0);
+	Brdline(bdict, '\n');
+	while ((p = Brdline(bdict, '\n')) != nil){
+		l = Blinelen(bdict);
+		if (!isdigit(*p))
+			continue;
+		for (i = 0; i < l-4; i++)
+			if (strncmp(p+i, " -- ", 4) == 0)
+				return Boffset(bdict)-l;
+	}
+	return Boffset(bdict);
+}
+
+void
+rogetprintkey(void)
+{
+	Bprint(bout, "No pronunciation key.\n");
+}
--- /sys/src/cmd/dict/mkfile	Fri Nov 19 21:00:32 2004
+++ /sys/src/cmd/dict/mkfile	Fri Nov 19 21:00:32 2004
@@ -2,7 +2,7 @@
 
 TARG=dict
 LFILES=oed.$O ahd.$O pcollins.$O pcollinsg.$O movie.$O slang.$O robert.$O\
-	world.$O jis208.$O gb2312.$O thesaurus.$O simple.$O pgw.$O
+	world.$O jis208.$O gb2312.$O thesaurus.$O simple.$O pgw.$O roget.$O
 
 OFILES=dict.$O\
 	$LFILES\
--- /sys/src/cmd/dict/mkroget	Thu Jan  1 00:00:00 1970
+++ /sys/src/cmd/dict/mkroget	Fri Nov 19 21:00:32 2004
@@ -0,0 +1,23 @@
+#!/bin/rc
+
+rtf2txt roget-body.rtf |
+	sed '
+		1,12d
+		/^100. /{
+			N
+			s/\n//
+			p
+		}
+		/^388a. /{
+			N
+			s/\n//
+			p
+		}
+	' > /lib/dict/roget
+
+mkindex -d roget | 
+	sort -u -t'	' +0f -1 +0 -1 +1n -2 |
+	sed '
+		s/[ 	]+$//g
+		s/ 	/	/g
+	' > /lib/dict/rogetindex