/* Pro un explication del scopo de iste programma, vide
   https://rudhar.com/lingtics/intrlnga/cgi-grep/fria/intro-ia.htm
   e specificamente
   https://rudhar.com/lingtics/intrlnga/cgi-grep/fria/intro-ia.htm#frhtmtxt
 */

#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <ctype.h>

#include "repundrl.h"

static char line1[4096];
static char lemma[512];

char *DT    = "<dt>";
char *SDTDD = "</dt> <dd>";
char *SDT   = "</dt>";
char *DD    = "<dd>";
char *DL    = "<dl>";
char *SDL   = "</dl>";

int main (void)
{
   FILE *fpi = stdin, *fpo = stdout, *flog = stderr;
   int inprepost = 1;
   long linecnt = 0, blanks = 0, dts = 0, dtss = 0, dds = 0;
   long prepost = 0, unrecog = 0, invalid = 0;
   char *s, *s1, *s2;
   int lemmalen;

   while (fgets(line1, sizeof line1, fpi))
   {
      linecnt++;

      if (s = strstr(line1, DL), s && s == line1)
      {
         inprepost = 0;
         prepost++;
         continue;
      }
      else if (s = strstr(line1, SDL), s && s == line1)
      {
         inprepost = 1;
      }

      if (inprepost)
      {
         prepost++;
      }
      else
      {
         if (line1[0] == '\n' && line1[1] == '\0')
         {
            /* Skip empty lines */
            blanks++;
         }
         else if (s1 = strstr(line1, DT), s1 && s1 == line1)
         {
            if (s2 = strstr(line1, SDTDD), s2)
            {
               char *b = s2 + strlen(SDTDD);

               dts++;

               lemmalen = s2 - line1 - strlen(DT);
               if (lemmalen >= (sizeof lemma) - 1)
                  lemmalen = (sizeof lemma) - 1;

               /* Save the lemma (between <dt> and </dt>) in a buffer for
                  later use. */
               sprintf(lemma, "%.*s", lemmalen, s1 + strlen(DT));

               RepeatWordsWithUnderlining(b, (sizeof line1) - (b - line1));

               /* 2., 3. etc, will be preceded by comma-space, so do that for
                  a numbered sense 1. as well. */
               if (isdigit(b[0]))
                  fprintf(fpo, "%s, %s", lemma, b);
               else
                  fprintf(fpo, "%s %s", lemma, b);
            }
            else if (s2 = strstr(line1, SDT), s2)
            {
               dts++;

               /* Sometimes all the beef is in following <dd> lines. No nee
                  to mention those. */
               lemmalen = s2 - line1 - strlen(DT);
               if (lemmalen >= (sizeof lemma) - 1)
                  lemmalen = (sizeof lemma) - 1;
               sprintf(lemma, "%.*s", lemmalen, s1 + strlen(DT));
               /* fprintf(fpo, "%s\n", lemma); */
            }
            else
            {
               invalid++;
               fprintf(flog, "Invalid line\n%s", line1);
            }
         }
         else if (strncmp(line1, DD, strlen(DD)) == 0)
         {
            dds++;
            RepeatWordsWithUnderlining(line1, (sizeof line1));
            fprintf(fpo, "%s, %s", lemma, line1 + strlen(DD));
         }
         else
         {
            unrecog++;
            fprintf(flog, "Unrecognised line\n%s", line1);
         }
      }
   }

   fprintf(flog, "\n");
   fprintf(flog, "Lines read    : %6ld\n", (long)linecnt);
   fprintf(flog, "\n");
   fprintf(flog, "Pre-post HTML : %6ld\n", (long)prepost);
   fprintf(flog, "Blanks skipped: %6ld\n", (long)blanks);
   fprintf(flog, "Dt's handled  : %6ld\n", (long)dts);
   fprintf(flog, "Dt shorts     : %6ld\n", (long)dtss);
   fprintf(flog, "Dd's handled  : %6ld\n", (long)dds);
   fprintf(flog, "Unrecognised  : %6ld\n", (long)unrecog);
   fprintf(flog, "Invalid lines : %6ld\n", (long)invalid);

   return 0;
}