[dict-beta] roget's thesaurus

[Chas W. <ch...@cm...>]

here is a diff to the 1.5 version of the dict-misc package that will
allow it to format the publicly available roget's thesaurus.

it not quite perfect yet (a few problems -- mostly related to headwords
wrapping to the next line, any ideas?)

i downloaded roget's from ftp://crl.nmsu.edu/CLR/lexica/roget-1911/t1911a.zip
but i am not sure if this is the 'cannonical' source.

[on a related side note, i have started writing a perl script to convert the
ced3 prolog database back to a more dict friendly format -- any interest
given the limbo state of the copyright of this database?]

--- dictfmt.c.000	Fri Aug 20 11:42:45 1999
+++ dictfmt.c	Sat Aug 21 07:19:47 1999
@@ -44,6 +44,7 @@
 #define PERIODIC  4
 #define HITCHCOCK 5
 #define CIA1995   6
+#define ROGET     7
 
 #define BSIZE 10240
 
@@ -214,13 +215,14 @@
    char       *s, *d;
    char       *buf;
 
-   while ((c = getopt(argc, argv, "jfephDu:s:c:")) != EOF)
+   while ((c = getopt(argc, argv, "jfephrDu:s:c:")) != EOF)
       switch (c) {
       case 'j': type = JARGON;    break;
       case 'f': type = FOLDOC;    break;
       case 'e': type = EASTON;    break;
       case 'p': type = PERIODIC;  break;
       case 'h': type = HITCHCOCK; break;
+      case 'r': type = ROGET;     break;
       case 'D': ++Debug;          break;
       case 'u': url = optarg;     break;
       case 's': sname = optarg;   break;
@@ -391,6 +393,54 @@
 		&& buffer[0] == buffer[2]
 		&& buffer[0] == buffer[3])
 	       continue;		/* Skip lines with *'s and ='s */
+	 }
+	 break;
+      case ROGET:		//chas
+	 switch (*buffer) {
+	 case ' ':
+            if (buffer[5] == '#') {
+	       header = 1;
+               s = &buffer[5];
+	       if ((s = strchr( s, ' ' ))) {
+                  while (*s == ' ') ++s;
+
+                  if (*s == '[') {
+                     for(c=0; c<strlen(s) && s[c]!=']' && s[c]!=')'; ) ++c;
+                     ++c;
+                     s = &s[c];
+                  }
+
+                  while (*s == ' ') ++s;
+
+                  for(c=0; c<strlen(s) && s[c] != ' ' && s[c] != '.'; ) ++c;
+                  d = &s[c];
+                  //d = strchr( s, '.');
+                  *d = '\0';
+#ifdef notdef
+                  fprintf(stderr, "headword -> %s\n", s);
+#endif
+	          fmt_newheadword(s, 0);
+
+                  *d = '.';
+	       }
+            }
+ 
+            /* skip lines starting with 12 blank spaces... */
+            for(c=0; c<11 && buffer[c] == ' '; ) ++c;
+            if (c == 11)
+               continue;
+
+	    break;
+         case '<':
+            continue;
+#ifdef notdef
+         case '\n':
+         case '\0':
+            header = 0;
+            break;
+         default:
+               if (!header) continue;	// dont print unless in defn
+#endif
 	 }
 	 break;
       case PERIODIC: