Breaking News
Loading...
Sunday, December 29, 2013

Stemming of a file in c++ | What is Stemming in IR

12:47 AM
Programming yes post of programming. This post may b unique over internet a Program of Stemming of File in C++, Last week while Working over a Project of Boolean Search Engine in C++. I have come up with problem and Project requirement was also to Use stemming. So i complete by project with following code of stemming.

About Stemming:

A common processing step in many IR(information retrieval) systems is called stemming. The main idea behind stemming is that users searching for information on “serving” will also be interested in documents that have information containing “served”, “r”, “serve”, and so on all worlds which is related to search query. Systems can be susceptible to errors due to poor stemming, so this is a little tricky. For example, a user interested in “information retrieval” might get a document titled “Information on Golden Retrievers” due to stemming.
this is used in search engines where we have to find documents containing relevant words the user write. this is little bit tricky but interesting the file and function i used to do stemming is a file and functions which i replied before so whenever you have to do stemming use it.

#include <stdlib.h>  /* for malloc, free */
#include <string.h>  /* for memcmp, memmove */

/* You will probably want to move the following declarations to a central
   header file.
*/

struct stemmer;

extern struct stemmer * create_stemmer(void);
extern void free_stemmer(struct stemmer * z);

extern int stem(struct stemmer * z, char * b, int k);



/* The main part of the stemming algorithm starts here.
*/

#define TRUE 1
#define FALSE 0

/* stemmer is a structure for a few local bits of data,
*/

struct stemmer {
   char * b;       /* buffer for word to be stemmed */
   int k;          /* offset to the end of the string */
   int j;          /* a general offset into the string */
};


/* Member b is a buffer holding a word to be stemmed. The letters are in
   b[0], b[1] ... ending at b[z->k]. Member k is readjusted downwards as
   the stemming progresses. Zero termination is not in fact used in the
   algorithm.

   Note that only lower case sequences are stemmed. Forcing to lower case
   should be done before stem(...) is called.


   Typical usage is:

       struct stemmer * z = create_stemmer();
       char b[] = "pencils";
       int res = stem(z, b, 6);
           /- stem the 7 characters of b[0] to b[6]. The result, res,
              will be 5 (the 's' is removed). -/
       free_stemmer(z);
*/


extern struct stemmer * create_stemmer(void)
{
    return (struct stemmer *) malloc(sizeof(struct stemmer));
    /* assume malloc succeeds */
}

extern void free_stemmer(struct stemmer * z)
{
    free(z);
}


/* cons(z, i) is TRUE <=> b[i] is a consonant. ('b' means 'z->b', but here
   and below we drop 'z->' in comments.
*/

static int cons(struct stemmer * z, int i)
{  switch (z->b[i])
   {  case 'a': case 'e': case 'i': case 'o': case 'u': return FALSE;
      case 'y': return (i == 0) ? TRUE : !cons(z, i - 1);
      default: return TRUE;
   }
}

/* m(z) measures the number of consonant sequences between 0 and j. if c is
   a consonant sequence and v a vowel sequence, and <..> indicates arbitrary
   presence,

      <c><v>       gives 0
      <c>vc<v>     gives 1
      <c>vcvc<v>   gives 2
      <c>vcvcvc<v> gives 3
      ....
*/

static int m(struct stemmer * z)
{  int n = 0;
   int i = 0;
   int j = z->j;
   while(TRUE)
   {  if (i > j) return n;
      if (! cons(z, i)) break; i++;
   }
   i++;
   while(TRUE)
   {  while(TRUE)
      {  if (i > j) return n;
            if (cons(z, i)) break;
            i++;
      }
      i++;
      n++;
      while(TRUE)
      {  if (i > j) return n;
         if (! cons(z, i)) break;
         i++;
      }
      i++;
   }
}

/* vowelinstem(z) is TRUE <=> 0,...j contains a vowel */

static int vowelinstem(struct stemmer * z)
{
   int j = z->j;
   int i; for (i = 0; i <= j; i++) if (! cons(z, i)) return TRUE;
   return FALSE;
}

/* doublec(z, j) is TRUE <=> j,(j-1) contain a double consonant. */

static int doublec(struct stemmer * z, int j)
{
   char * b = z->b;
   if (j < 1) return FALSE;
   if (b[j] != b[j - 1]) return FALSE;
   return cons(z, j);
}

/* cvc(z, i) is TRUE <=> i-2,i-1,i has the form consonant - vowel - consonant
   and also if the second c is not w,x or y. this is used when trying to
   restore an e at the end of a short word. e.g.

      cav(e), lov(e), hop(e), crim(e), but
      snow, box, tray.

*/

static int cvc(struct stemmer * z, int i)
{  if (i < 2 || !cons(z, i) || cons(z, i - 1) || !cons(z, i - 2)) return FALSE;
   {  int ch = z->b[i];
      if (ch  == 'w' || ch == 'x' || ch == 'y') return FALSE;
   }
   return TRUE;
}

/* ends(z, s) is TRUE <=> 0,...k ends with the string s. */

static int ends(struct stemmer * z, char * s)
{  int length = s[0];
   char * b = z->b;
   int k = z->k;
   if (s[length] != b[k]) return FALSE; /* tiny speed-up */
   if (length > k + 1) return FALSE;
   if (memcmp(b + k - length + 1, s + 1, length) != 0) return FALSE;
   z->j = k-length;
   return TRUE;
}

/* setto(z, s) sets (j+1),...k to the characters in the string s, readjusting
   k. */

static void setto(struct stemmer * z, char * s)
{  int length = s[0];
   int j = z->j;
   memmove(z->b + j + 1, s + 1, length);
   z->k = j+length;
}

/* r(z, s) is used further down. */

static void r(struct stemmer * z, char * s) { if (m(z) > 0) setto(z, s); }

/* step1ab(z) gets rid of plurals and -ed or -ing. e.g.

       caresses  ->  caress
       ponies    ->  poni
       ties      ->  ti
       caress    ->  caress
       cats      ->  cat

       feed      ->  feed
       agreed    ->  agree
       disabled  ->  disable

       matting   ->  mat
       mating    ->  mate
       meeting   ->  meet
       milling   ->  mill
       messing   ->  mess

       meetings  ->  meet

*/

static void step1ab(struct stemmer * z)
{
   char * b = z->b;
   if (b[z->k] == 's')
   {  if (ends(z, "\04" "sses")) z->k -= 2; else
      if (ends(z, "\03" "ies")) setto(z, "\01" "i"); else
      if (b[z->k - 1] != 's') z->k--;
   }
   if (ends(z, "\03" "eed")) { if (m(z) > 0) z->k--; } else
   if ((ends(z, "\02" "ed") || ends(z, "\03" "ing")) && vowelinstem(z))
   {  z->k = z->j;
      if (ends(z, "\02" "at")) setto(z, "\03" "ate"); else
      if (ends(z, "\02" "bl")) setto(z, "\03" "ble"); else
      if (ends(z, "\02" "iz")) setto(z, "\03" "ize"); else
      if (doublec(z, z->k))
      {  z->k--;
         {  int ch = b[z->k];
            if (ch == 'l' || ch == 's' || ch == 'z') z->k++;
         }
      }
      else if (m(z) == 1 && cvc(z, z->k)) setto(z, "\01" "e");
   }
}

/* step1c(z) turns terminal y to i when there is another vowel in the stem. */

static void step1c(struct stemmer * z)
{
   if (ends(z, "\01" "y") && vowelinstem(z)) z->b[z->k] = 'i';
}


/* step2(z) maps double suffices to single ones. so -ization ( = -ize plus
   -ation) maps to -ize etc. note that the string before the suffix must give
   m(z) > 0. */

static void step2(struct stemmer * z) { switch (z->b[z->k-1])
{
   case 'a': if (ends(z, "\07" "ational")) { r(z, "\03" "ate"); break; }
             if (ends(z, "\06" "tional")) { r(z, "\04" "tion"); break; }
             break;
   case 'c': if (ends(z, "\04" "enci")) { r(z, "\04" "ence"); break; }
             if (ends(z, "\04" "anci")) { r(z, "\04" "ance"); break; }
             break;
   case 'e': if (ends(z, "\04" "izer")) { r(z, "\03" "ize"); break; }
             break;
   case 'l': if (ends(z, "\03" "bli")) { r(z, "\03" "ble"); break; } /*-DEPARTURE-*/

 /* To match the published algorithm, replace this line with
    case 'l': if (ends(z, "\04" "abli")) { r(z, "\04" "able"); break; } */

             if (ends(z, "\04" "alli")) { r(z, "\02" "al"); break; }
             if (ends(z, "\05" "entli")) { r(z, "\03" "ent"); break; }
             if (ends(z, "\03" "eli")) { r(z, "\01" "e"); break; }
             if (ends(z, "\05" "ousli")) { r(z, "\03" "ous"); break; }
             break;
   case 'o': if (ends(z, "\07" "ization")) { r(z, "\03" "ize"); break; }
             if (ends(z, "\05" "ation")) { r(z, "\03" "ate"); break; }
             if (ends(z, "\04" "ator")) { r(z, "\03" "ate"); break; }
             break;
   case 's': if (ends(z, "\05" "alism")) { r(z, "\02" "al"); break; }
             if (ends(z, "\07" "iveness")) { r(z, "\03" "ive"); break; }
             if (ends(z, "\07" "fulness")) { r(z, "\03" "ful"); break; }
             if (ends(z, "\07" "ousness")) { r(z, "\03" "ous"); break; }
             break;
   case 't': if (ends(z, "\05" "aliti")) { r(z, "\02" "al"); break; }
             if (ends(z, "\05" "iviti")) { r(z, "\03" "ive"); break; }
             if (ends(z, "\06" "biliti")) { r(z, "\03" "ble"); break; }
             break;
   case 'g': if (ends(z, "\04" "logi")) { r(z, "\03" "log"); break; } /*-DEPARTURE-*/

 /* To match the published algorithm, delete this line */

} }

/* step3(z) deals with -ic-, -full, -ness etc. similar strategy to step2. */

static void step3(struct stemmer * z) { switch (z->b[z->k])
{
   case 'e': if (ends(z, "\05" "icate")) { r(z, "\02" "ic"); break; }
             if (ends(z, "\05" "ative")) { r(z, "\00" ""); break; }
             if (ends(z, "\05" "alize")) { r(z, "\02" "al"); break; }
             break;
   case 'i': if (ends(z, "\05" "iciti")) { r(z, "\02" "ic"); break; }
             break;
   case 'l': if (ends(z, "\04" "ical")) { r(z, "\02" "ic"); break; }
             if (ends(z, "\03" "ful")) { r(z, "\00" ""); break; }
             break;
   case 's': if (ends(z, "\04" "ness")) { r(z, "\00" ""); break; }
             break;
} }

/* step4(z) takes off -ant, -ence etc., in context <c>vcvc<v>. */

static void step4(struct stemmer * z)
{  switch (z->b[z->k-1])
   {  case 'a': if (ends(z, "\02" "al")) break; return;
      case 'c': if (ends(z, "\04" "ance")) break;
                if (ends(z, "\04" "ence")) break; return;
      case 'e': if (ends(z, "\02" "er")) break; return;
      case 'i': if (ends(z, "\02" "ic")) break; return;
      case 'l': if (ends(z, "\04" "able")) break;
                if (ends(z, "\04" "ible")) break; return;
      case 'n': if (ends(z, "\03" "ant")) break;
                if (ends(z, "\05" "ement")) break;
                if (ends(z, "\04" "ment")) break;
                if (ends(z, "\03" "ent")) break; return;
      case 'o': if (ends(z, "\03" "ion") && (z->b[z->j] == 's' || z->b[z->j] == 't')) break;
                if (ends(z, "\02" "ou")) break; return;
                /* takes care of -ous */
      case 's': if (ends(z, "\03" "ism")) break; return;
      case 't': if (ends(z, "\03" "ate")) break;
                if (ends(z, "\03" "iti")) break; return;
      case 'u': if (ends(z, "\03" "ous")) break; return;
      case 'v': if (ends(z, "\03" "ive")) break; return;
      case 'z': if (ends(z, "\03" "ize")) break; return;
      default: return;
   }
   if (m(z) > 1) z->k = z->j;
}

/* step5(z) removes a final -e if m(z) > 1, and changes -ll to -l if
   m(z) > 1. */

static void step5(struct stemmer * z)
{
   char * b = z->b;
   z->j = z->k;
   if (b[z->k] == 'e')
   {  int a = m(z);
      if (a > 1 || a == 1 && !cvc(z, z->k - 1)) z->k--;
   }
   if (b[z->k] == 'l' && doublec(z, z->k) && m(z) > 1) z->k--;
}

/* In stem(z, b, k), b is a char pointer, and the string to be stemmed is
   from b[0] to b[k] inclusive.  Possibly b[k+1] == '\0', but it is not
   important. The stemmer adjusts the characters b[0] ... b[k] and returns
   the new end-point of the string, k'. Stemming never increases word
   length, so 0 <= k' <= k.
*/

extern int stem(struct stemmer * z, char * b, int k)
{
   if (k <= 1) return k; /*-DEPARTURE-*/
   z->b = b; z->k = k; /* copy the parameters into z */

   /* With this line, strings of length 1 or 2 don't go through the
      stemming process, although no mention is made of this in the
      published algorithm. Remove the line to match the published
      algorithm. */

   step1ab(z); step1c(z); step2(z); step3(z); step4(z); step5(z);
   return z->k;
}

/*--------------------stemmer definition ends here------------------------*/

#include <stdio.h>
#include <stdlib.h>      /* for malloc, free */
#include <ctype.h>       /* for isupper, islower, tolower */

static char * s;         /* buffer for words tobe stemmed */

#define INC 50           /* size units in which s is increased */
static int i_max = INC;  /* maximum offset in s */

#define LETTER(ch) (isupper(ch) || islower(ch))

void stemfile(struct stemmer * z, FILE * f)
{  while(TRUE)
   {  int ch = getc(f);
      if (ch == EOF) return;
      if (LETTER(ch))
      {  int i = 0;
         while(TRUE)
         {  if (i == i_max)
            {  i_max += INC;
               s = realloc(s, i_max + 1);
            }
            ch = tolower(ch); /* forces lower case */

            s[i] = ch; i++;
            ch = getc(f);
            if (!LETTER(ch)) { ungetc(ch,f); break; }
         }
         s[stem(z, s, i - 1) + 1] = 0;
         /* the previous line calls the stemmer and uses its result to
            zero-terminate the string in s */
         printf("%s",s);
      }
      else putchar(ch);
   }
}

int main(int argc, char * argv[])
{  int i;

   struct stemmer * z = create_stemmer();

   s = (char *) malloc(i_max + 1);
   for (i = 1; i < argc; i++)
   {  FILE * f = fopen(argv[i],"r");
      if (f == 0) { fprintf(stderr,"File %s not found\n",argv[i]); exit(1); }
      stemfile(z, f);
   }
   free(s);

   free_stemmer(z);

   return 0;
}

0 comments:

Post a Comment

 
Toggle Footer