dawg.cpp

来自「一个google的OCR源码」· C++ 代码 · 共 392 行

/* -*-C-*-
 ********************************************************************************
 *
 * File:        dawg.c  (Formerly dawg.c)
 * Description:  Use a Directed Accyclic Word Graph
 * Author:       Mark Seaman, OCR Technology
 * Created:      Fri Oct 16 14:37:00 1987
 * Modified:     Wed Jul 24 16:59:16 1991 (Mark Seaman) marks@hpgrlt
 * Language:     C
 * Package:      N/A
 * Status:       Reusable Software Component
 *
 * (c) Copyright 1987, Hewlett-Packard Company.
 ** Licensed under the Apache License, Version 2.0 (the "License");
 ** you may not use this file except in compliance with the License.
 ** You may obtain a copy of the License at
 ** http://www.apache.org/licenses/LICENSE-2.0
 ** Unless required by applicable law or agreed to in writing, software
 ** distributed under the License is distributed on an "AS IS" BASIS,
 ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 ** See the License for the specific language governing permissions and
 ** limitations under the License.
 *
 *********************************************************************************/
/*----------------------------------------------------------------------
              I n c l u d e s
----------------------------------------------------------------------*/
#ifdef __MSW32__
#include <windows.h>
#else
#include <netinet/in.h>
#endif
#include "dawg.h"
#include "cutil.h"
#include "tprintf.h"
#include "freelist.h"
#include "context.h"
#include "strngs.h"
#include "emalloc.h"

/*----------------------------------------------------------------------
              V a r i a b l e s
----------------------------------------------------------------------*/
inT32 debug          = 0;
inT32 case_sensative = 1;

/*----------------------------------------------------------------------
              F u n c t i o n s
----------------------------------------------------------------------*/
/**********************************************************************
 * edge_char_of
 *
 * Return the edge that corresponds to the letter out of this node.
 **********************************************************************/
EDGE_REF edge_char_of(EDGE_ARRAY dawg,
                      NODE_REF node,
                      int character,
                      int word_end) {
  EDGE_REF   edge = node;

  if (! case_sensative) character = tolower (character);

  if (edge_occupied (dawg, edge)) {
    do {
      if ((edge_letter (dawg, edge) == character) &&
        (! word_end || end_of_word(dawg,edge)))
        return (edge);

    } edge_loop (dawg, edge);
  }

  return (NO_EDGE);
}


/**********************************************************************
 * edges_in_node
 *
 * Count the number of edges in this node in the DAWG. This includes
 * both forward and back links.
 **********************************************************************/
inT32 edges_in_node(EDGE_ARRAY dawg, NODE_REF node) {
  EDGE_REF   edge = node;

  if (edge_occupied (dawg, edge)) {
    edge_loop(dawg, edge);
    if (edge_occupied (dawg, edge) && backward_edge (dawg, edge)) {
      edge_loop(dawg, edge);
      return (edge - node);
    }
    else {
      return (edge - node);
    }
  }
  else {
    return (edge - node);
  }
}


/*
 * Initialize letter_is_okay to point to default implmentation (a main
 * program can override this).
 */
LETTER_OK_FUNC letter_is_okay = &def_letter_is_okay;

/**********************************************************************
 * def_letter_is_okay
 *
 * Default way to check this letter in light of the current state.  If
 * everything is still OK then return TRUE.
 **********************************************************************/
// TODO(tkielbus) Change the prevchar argument to make it unicode safe.
// We might want to get rid of def_letter_is_okay at some point though.
inT32 def_letter_is_okay(EDGE_ARRAY dawg,
                         NODE_REF *node,
                         inT32 char_index,
                         char prevchar,
                         const char *word,
                         inT32 word_end) {
  EDGE_REF     edge;
  STRING dummy_word(word);  // Auto-deleting string fixes memory leak.
  STRING word_single_lengths; //Lengths of single UTF-8 characters of the word.
  const char *ptr;

  for (ptr = word; *ptr != '\0';) {
    int step = UNICHAR::utf8_step(ptr);
    if (step == 0)
      return FALSE;
    word_single_lengths += step;
    ptr += step;
  }

  if (*node == NO_EDGE) {        /* Trailing punctuation */
    if (trailing_punc(dummy_word[char_index]) &&
        punctuation_ok(dummy_word.string(), word_single_lengths.string()) >= 0)
      return (TRUE);
    else
      return (FALSE);
  }
  else {
    /* Leading punctuation */
    if (*node == 0                           &&
      char_index != 0                      &&
        // TODO(tkielbus) Replace islalpha by unicode versions.
        // However the lengths information is not available at this point in the
        // code. We will probably get rid of the dictionaries at some point anyway.
        isalpha (dummy_word [char_index])          &&
      ! leading_punc (dummy_word [char_index-1]) &&
    dummy_word [char_index-1] != '-') {
      return (FALSE);
    }
  }
  /* Handle compund words */
#if 0
  if (dummy_word [char_index] == '-') {
    if (char_index>0 && !word_end
         && word [char_index-1] == '-'
         && word [char_index+1] == '-')
      return FALSE;              /*not allowed*/
    dummy_word [char_index] = (char) 0;
    if (word_in_dawg (dawg, dummy_word.string())) {
      dummy_word [char_index] = '-';
      *node = 0;
      return (TRUE);
    }
    else {
      dummy_word [char_index] = '-';
      return (FALSE);
    }
  }
#endif
  /* Check the DAWG */
  edge = edge_char_of (dawg, *node, dummy_word [char_index], word_end);

  if (edge != NO_EDGE) {         /* Normal edge in DAWG */
    if (case_sensative || case_is_okay (dummy_word, char_index)) {
                                 //next_node (dawg, edge);
      *node = next_node(dawg, edge);
      if (*node == 0)
        *node = NO_EDGE;
      return (TRUE);
    } else {
      return (FALSE);
    }
  }
  else {
    /* Leading punctuation */
    if (leading_punc (word [char_index]) &&
    (char_index == 0  ||  leading_punc (dummy_word [char_index-1]))) {
      *node = 0;
      if (punctuation_ok(word, word_single_lengths.string()) >= 0)
        return (TRUE);
      else
        return FALSE;
    }
    /* Trailing punctuation */
    if (verify_trailing_punct (dawg, &dummy_word[0], char_index)) {
      *node = NO_EDGE;
      return (TRUE);
    }

    return (FALSE);
  }
}


/**********************************************************************
 * num_forward_edges
 *
 * Count and return the number of forward edges for this node.
 **********************************************************************/
inT32 num_forward_edges(EDGE_ARRAY dawg, NODE_REF node) {
  EDGE_REF   edge = node;
  inT32        num  = 0;

  if (forward_edge (dawg, edge)) {
    do {
      num++;
    } edge_loop (dawg, edge);
  }

  return (num);
}


/**********************************************************************
 * print_dawg_node
 *
 * Print the contents of one of the nodes in the DAWG.
 **********************************************************************/
void print_dawg_node(EDGE_ARRAY dawg, NODE_REF node) {
  EDGE_REF   edge = node;
  const char       *forward_string  = "FORWARD";
  const char       *backward_string = "       ";

  const char       *last_string     = "LAST";
  const char       *not_last_string = "    ";

  const char       *eow_string      = "EOW";
  const char       *not_eow_string  = "   ";

  const char       *direction;
  const char       *is_last;
  const char       *eow;

  char       ch;

  if (edge_occupied (dawg, edge)) {
    do {
      if (forward_edge (dawg, edge)) direction = forward_string;
      else                           direction = backward_string;

      if (last_edge    (dawg, edge)) is_last   = last_string;
      else                           is_last   = not_last_string;

      if (end_of_word  (dawg, edge)) eow       = eow_string;
      else                           eow       = not_eow_string;

      ch = edge_letter (dawg, edge);
      cprintf (REFFORMAT " : next = " REFFORMAT ", char = '%c', %s %s %s\n",
        edge, next_node (dawg, edge), ch,
        direction, is_last, eow);

      if (edge - node > MAX_NODE_EDGES_DISPLAY) return;
    } edge_loop (dawg, edge);

    if (edge_occupied (dawg, edge) && backward_edge (dawg, edge)) {
      do {
        if (forward_edge (dawg, edge)) direction = forward_string;
        else                           direction = backward_string;

        if (last_edge    (dawg, edge)) is_last   = last_string;
        else                           is_last   = not_last_string;

        if (end_of_word  (dawg, edge)) eow       = eow_string;
        else                           eow       = not_eow_string;

        ch = edge_letter (dawg, edge);
        cprintf (REFFORMAT " : next = " REFFORMAT ", char = '%c', %s %s %s\n",
          edge, next_node (dawg, edge), ch,
          direction, is_last, eow);

        if (edge - node > MAX_NODE_EDGES_DISPLAY) return;
      } edge_loop (dawg, edge);
    }
  }
  else {
    cprintf (REFFORMAT " : no edges in this node\n", node);
  }
  cprintf("\n");
}


/**********************************************************************
 * read_squished_dawg
 *
 * Read the DAWG from a file and return it. Must be freed with memfree.
 **********************************************************************/
EDGE_ARRAY read_squished_dawg(const char *filename) {
  FILE       *file;
  EDGE_REF   edge;
  inT32      num_edges = 0;
  inT32      node_count = 0;

  if (debug) print_string ("read_debug");

  #ifdef __UNIX__
  file = open_file (filename, "r");
  #else
  file = open_file (filename, "rb");
  #endif
  fread (&num_edges,  sizeof (inT32), 1, file);
  num_edges = ntohl(num_edges);
  if (num_edges > MAX_NUM_EDGES_IN_SQUISHED_DAWG_FILE || num_edges < 0) {
    tprintf("(ENDIAN)Error: trying to read a DAWG '%s' that contains "
            "%d edges while the maximum is %d.\n",
            filename, num_edges, MAX_NUM_EDGES_IN_SQUISHED_DAWG_FILE);
    exit(1);
  }

  uinT32 *dawg_32 = (uinT32*) Emalloc(num_edges * sizeof (uinT32));
  fread(&dawg_32[0], sizeof (uinT32), num_edges, file);
  fclose(file);
  EDGE_ARRAY dawg = (EDGE_ARRAY) memalloc(sizeof(EDGE_RECORD) * num_edges);

  for (edge = 0; edge < num_edges; ++edge)
    dawg[edge] = ntohl(dawg_32[edge]);

  Efree(dawg_32);

  for (edge = 0; edge < num_edges; ++edge)
    if (last_edge (dawg, edge)) node_count++;

  return dawg;
}


/**********************************************************************
 * verify_trailing_punct
 *
 * Make sure that there is a valid transition from the word core to a
 * string of trailing puntuation.  TRUE is returned if everything is
 * OK.
 **********************************************************************/
inT32 verify_trailing_punct(EDGE_ARRAY dawg, char *word, inT32 char_index) {
  char       last_char;
  char       *first_char;

  if (trailing_punc (word [char_index])) {

    last_char = word [char_index];
    word [char_index] = (char) 0;

    for (first_char = word; leading_punc (first_char[0]); first_char++);

    if (word_in_dawg (dawg, first_char)) {
      word [char_index] = last_char;
      return (TRUE);
    }
    word [char_index] = last_char;
  }
  return (FALSE);
}


/**********************************************************************
 * word_in_dawg
 *
 * Test to see if the word can be found in the DAWG.
 **********************************************************************/
inT32 word_in_dawg(EDGE_ARRAY dawg, const char *string) {
  NODE_REF   node = 0;
  inT32        i;
  inT32         length;

  length=strlen(string);
  if (length==0)
    return FALSE;
  for (i=0; i<length; i++) {
    if (debug > 1) {
      print_dawg_node(dawg, node);
      new_line();
    }
    if (! letter_is_okay (dawg, &node, i, '\0', string, (string[i+1]==0))) {
      return (FALSE);
    }
  }

  return (TRUE);
}