/*  Arabic stemmer (stem2) tring to remove common prefixes and suffixes */
/*  more conservative than stem3 (see J. Savoy, Y. Rasolofo, TREC 2002) */

static char *removeArabicSuffix2();
static char *removeArabicPrefix2();


/* For the Arabic language */

char *arabic_stemming2(word)
char *word;
{ 
   removeArabicSuffix2(word);
   removeArabicPrefix2(word);
   return(word);
}


static char *removeArabicSuffix2(word)
char *word;
{ 
int len = strlen (word)-1;

if (len > 4) {   
              /* -A{nt}    */
   if ((word[len-1]=='A') && ((word[len]=='t') || (word[len]=='n'))) {  
      word[len-1]='\0'; 
      return(word); 
      }
              /* -hA    */
   if ((word[len]=='A') && (word[len-1]=='h')) {  
      word[len-1]='\0'; 
      return(word); 
      }
              /* -Y{phn}    */
   if ((word[len-1]=='Y') && ((word[len]=='p') || (word[len]=='h') ||
                             (word[len]=='n'))) {  
      word[len-1]='\0'; 
      return(word); 
      }
              /* -wn    */
   if ((word[len-1]=='w') && (word[len]=='n')) {  
      word[len-1]='\0'; 
      return(word); 
      }
   }  /* end if len > 4 */

if (len > 3) {   
             /* -{Yph}  -> -  */
   if ((word[len]=='Y') || (word[len]=='p') || (word[len]=='h')) {
      word[len]='\0';
      return(word);  
      }
   }  /* end if len > 3 */

return(word);
}


static char *removeArabicPrefix2(word)
char *word;
{ 
int pos;
int len = strlen (word)-1;

if (len > 5) {   
              /* {fXbw}Al-- -> --    */
   if ((word[2]=='l') && (word[1]=='A') && ((word[0]=='f') || 
        (word[0]=='X') || (word[0]=='b') || (word[0]=='w'))) {  
      pos = 0;
      while (word[pos+3] != '\0') {
         word[pos] = word[pos+3];
         pos++;
         }
      word[pos]= '\0';
      return(word);
      }  
   }  /* end if len > 5 */

if (len > 4) {   
              /* Al-- -> --    */
   if ((word[0]=='A') && (word[1]=='l')) {  
      pos = 0;
      while (word[pos+2] != '\0') {
         word[pos] = word[pos+2];
         pos++;
         }
      word[pos]= '\0';
      return(word);
      }  
   }  /* end if len > 4 */

if (len > 3) {   
             /* {wY}--  -> --  */
   if ((word[0]=='w') || (word[0]=='Y')) {  
      pos = 0;
      while (word[pos] != '\0') {
         word[pos] = word[pos+1];
         pos++;
         }
      return(word);  
      }
   }  /* end if len > 3 */

return(word);
}