stem.h

根据潜在语义分析进行查询。将文本中的特征集合做LSI变换。

/****************************************************
 *
 *  TEXT INDEXING
 *
 *  Program  : stem.h
 *  Function : Header file for stem.c
 *
 *  Original Author : C. Fox, 1990
 *  Modified By     : Choon Eng, You
 *  Year            : 1999
 *
 *****************************************************/


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

/******** Private Defines and Data Structures **********************/

#define FALSE   0
#define TRUE    1
#define EOS   '\0'

#define IsVowel(c) ('A'==(c)||'E'==(c)||'I'==(c)||'O'==(c)||'U'==(c))

typedef struct {
  int id;
  char *old_end;
  char *new_end;
  int old_offset;
  int new_offset;
  int min_root_size;
  int (*condition)();
} RuleList;

static char LAMBDA[1] = "";
static char *end;

/******** Private Function Declarations ****************************/

#ifdef __STDC__

static int WordSize(char *word);
static int ContainsVowel(char *word);
static int EndsWithCVC(char *word);
static int AddAnE(char *word);
static int RemoveAnE(char *word);
static int ReplaceEnd(char *word, RuleList *rule);

#else

static int WordSize (/* word */);
static int ContainsVowel(/* word */);
static int EndsWithCVC(/* word */);
static int AddAnE(/* word */);
static int RemoveAnE(/* word */);
static int ReplaceEnd(/* word, rule */);

#endif

/******** Initialized Private Data Structures **********************/

static RuleList step1a_rules[] =
{
   101, "SSES",     "SS",   3,  1, -1, NULL,
   102, "IES",      "I",    2,  0, -1, NULL,
   103, "SS",       "SS",   1,  1, -1, NULL,
   104, "S",        LAMBDA, 0, -1, -1, NULL,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step1b_rules[] =
{
   105, "EED",      "EE",   2,  1,  0, NULL,
   106, "ED",       LAMBDA, 1, -1, -1, ContainsVowel,
   107, "ING",      LAMBDA, 2, -1, -1, ContainsVowel,
   000, NULL,       NULL,   0,  0,  0, NULL,
};


static RuleList step1b1_rules[] =
{
   108, "AT",       "ATE",  1,  2, -1, NULL,
   109, "BL",       "BLE",  1,  2, -1, NULL,
   110, "IZ",       "IZE",  1,  2, -1, NULL,
   111, "BB",       "B",    1,  0, -1, NULL,
   112, "DD",       "D",    1,  0, -1, NULL,
   113, "FF",       "F",    1,  0, -1, NULL,
   114, "GG",       "G",    1,  0, -1, NULL,
   115, "MM",       "M",    1,  0, -1, NULL,
   116, "NN",       "N",    1,  0, -1, NULL,
   117, "PP",       "P",    1,  0, -1, NULL,
   118, "RR",       "R",    1,  0, -1, NULL,
   119, "TT",       "T",    1,  0, -1, NULL,
   120, "WW",       "W",    1,  0, -1, NULL,
   121, "XX",       "X",    1,  0, -1, NULL,
   122, LAMBDA,     "E",   -1,  0, -1, AddAnE,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step1c_rules[] =
{

   123, "Y",        "I",    0,  0, -1, ContainsVowel,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step2_rules[] =
{
   203, "ATIONAL",  "ATE",  6,  2,  0, NULL,
   204, "TIONAL",   "TION", 5,  3,  0, NULL,
   205, "ENCI",     "ENCE", 3,  3,  0, NULL,
   206, "ANCI",     "ANCE", 3,  3,  0, NULL,
   207, "IZER",     "IZE",  3,  2,  0, NULL,
   208, "ABLI",     "ABLE", 3,  3,  0, NULL,
   209, "ALLI",     "AL",   3,  1,  0, NULL,
   210, "ENTLI",    "ENT",  4,  2,  0, NULL,
   211, "ELI",      "E",    2,  0,  0, NULL,
   213, "OUSLI",    "OUS",  4,  2,  0, NULL,
   214, "IZATION",  "IZE",  6,  2,  0, NULL,
   215, "ATION",    "ATE",  4,  2,  0, NULL,
   216, "ATOR",     "ATE",  3,  2,  0, NULL,
   217, "ALISM",    "AL",   4,  1,  0, NULL,
   218, "IVENESS",  "IVE",  6,  2,  0, NULL,
   219, "FULNESS",  "FUL",  5,  2,  0, NULL,
   220, "OUSNESS",  "OUS",  6,  2,  0, NULL,
   221, "ALITI",    "AL",   4,  1,  0, NULL,
   222, "IVITI",    "IVE",  4,  2,  0, NULL,
   223, "BILITI",   "BLE",  5,  2,  0, NULL,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step3_rules[] =
{
   301, "ICATE",    "IC",   4,  1,  0, NULL,
   302, "ATIVE",    LAMBDA, 4, -1,  0, NULL,
   303, "ALIZE",    "AL",   4,  1,  0, NULL,
   304, "ICITI",    "IC",   4,  1,  0, NULL,
   305, "ICAL",     "IC",   3,  1,  0, NULL,
   308, "FUL",      LAMBDA, 2, -1,  0, NULL,
   309, "NESS",     LAMBDA, 3, -1,  0, NULL,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step4_rules[] =
{
   401, "AL",       LAMBDA, 1, -1,  1, NULL,
   402, "ANCE",     LAMBDA, 3, -1,  1, NULL,
   403, "ENCE",     LAMBDA, 3, -1,  1, NULL,
   405, "ER",       LAMBDA, 1, -1,  1, NULL,
   406, "IC",       LAMBDA, 1, -1,  1, NULL,
   407, "ABLE",     LAMBDA, 3, -1,  1, NULL,
   408, "IBLE",     LAMBDA, 3, -1,  1, NULL,
   409, "ANT",      LAMBDA, 2, -1,  1, NULL,
   410, "EMENT",    LAMBDA, 4, -1,  1, NULL,
   411, "MENT",     LAMBDA, 3, -1,  1, NULL,
   412, "ENT",      LAMBDA, 2, -1,  1, NULL,
   423, "SION",     "S",    3,  0,  1, NULL,
   424, "TION",     "T",    3,  0,  1, NULL,
   415, "OU",       LAMBDA, 1, -1,  1, NULL,
   416, "ISM",      LAMBDA, 2, -1,  1, NULL,
   417, "ATE",      LAMBDA, 2, -1,  1, NULL,
   418, "ITI",      LAMBDA, 2, -1,  1, NULL,
   419, "OUS",      LAMBDA, 2, -1,  1, NULL,
   420, "IVE",      LAMBDA, 2, -1,  1, NULL,
   421, "IZE",      LAMBDA, 2, -1,  1, NULL,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step5a_rules[] =
{
   501, "E",        LAMBDA, 0, -1,  1, NULL,
   502, "E",        LAMBDA, 0, -1, -1, RemoveAnE,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

static RuleList step5b_rules[] =
{
   503, "LL",       "L",    1,  0,  1, NULL,
   000, NULL,       NULL,   0,  0,  0, NULL,
};

/******** Public Function Declarations ************************/

int stem(word);