/TeXmacs-1.0.7.11-src/src/System/Language/hyphenate.cpp

  1
  2/******************************************************************************
  3* MODULE     : hyphenate.cpp
  4* DESCRIPTION: hyphenation by Liang's algorithm
  5* COPYRIGHT  : (C) 1999  Joris van der Hoeven
  6*******************************************************************************
  7* This software falls under the GNU general public license version 3 or later.
  8* It comes WITHOUT ANY WARRANTY WHATSOEVER. For details, see the file LICENSE
  9* in the root directory or <http://www.gnu.org/licenses/gpl-3.0.html>.
 10******************************************************************************/
 11
 12#include "file.hpp"
 13#include "hyphenate.hpp"
 14#include "analyze.hpp"
 15
 16#include <stdio.h>
 17#include <stdlib.h>
 18#include <string.h>
 19
 20typedef int SI;
 21#define MAX_SEARCH 10
 22#define MAX_BUFFER_SIZE 256
 23
 24/*
 25static bool
 26my_strncmp (char* s1, char* s2, int len) {
 27  int i;
 28  for (i=0; i<len; i++) if (s1[i]!=s2[i]) return false;
 29  return true;
 30}
 31*/
 32
 33static string
 34unpattern (string s) {
 35  int i, n= N(s);
 36  string r;
 37  for (i=0; i<n; ) {
 38    while ((i<n) && (s[i]>='0') && (s[i]<='9')) i++;
 39    if (i<n) r << s[i++];
 40  }
 41  return r;
 42}
 43
 44static string
 45hyphen_normalize (string s) {
 46  int i;
 47  string r (0);
 48  for (i=0; i<N(s); i++)
 49    if ((i+3<N(s)) && (s[i]=='^') && (s[i+1]=='^')) {
 50      r << from_hexadecimal (s (i+2, i+4));
 51      i+=3;
 52    }
 53    else r << s[i];
 54  return r;
 55}
 56
 57hashmap<string,string>
 58load_hyphen_table (string file_name) {
 59  string s;
 60  file_name= string ("hyphen.") * file_name;
 61  load_string (url ("$TEXMACS_PATH/langs/natural/hyphen", file_name), s, true);
 62  if (DEBUG_VERBOSE) cout << "TeXmacs] Loading " << file_name << "\n";
 63
 64  hashmap<string,string> H ("?");
 65  bool flag=false;
 66  int i=0, n= N(s);
 67  while (i<n) {
 68    string buffer;
 69    while ((i<n) && (s[i]!=' ') && (s[i]!='\t') && (s[i]!='\n')) {
 70      if (s[i] != '%') buffer << s[i++];
 71      else while ((i<n) && (s[i]!='\n')) i++;
 72    }
 73    if (i<n) i++;
 74    if (buffer == "}") flag=false;
 75    string norm= hyphen_normalize (buffer);
 76    //cout << norm << "\n";
 77    if (flag && (i!=0)) H (unpattern (norm))= norm;
 78    if (buffer == "\\patterns{") flag=true;
 79  }
 80
 81  // cout << file_name << " done!\n";
 82  return H;
 83}
 84
 85static string
 86lower_case (string s) {
 87  int i;
 88  string r (N(s));
 89  for (i=0; i<N(s); i++) {
 90    if ((s[i]>='A') && (s[i]<='Z'))
 91      r[i]= (char) (((int) s[i])+ ((int) 'a')- ((int) 'A'));
 92    else r[i]=s[i];
 93  }
 94  return r;
 95}
 96
 97array<int>
 98get_hyphens (string s, hashmap<string,string> H) {
 99  ASSERT (N(s) != 0, "hyphenation of empty string");
100  s= "." * lower_case (s) * ".";
101  // cout << s << "\n";
102  int i, j, k, m, len;
103  array<int> T (N(s)+1);
104  for (i=0; i<N(s)+1; i++) T[i]=0;
105  for (len=1; len < MAX_SEARCH; len++)
106    for (i=0; i<N(s)-len; i++) {
107      string r= H [s (i, i+len)];
108      if (!(r == "?")) {
109	// cout << "  " << s (i, i+len) << " => " << r << "\n";
110	for (j=0, k=0; j<=len; j++, k++) {
111	  if ((k<N(r)) && (r[k]>='0') && (r[k]<='9')) {
112	    m=((int) r[k])-((int) '0');
113	    k++;
114	  }
115	  else m=0;
116	  if (m>T[i+j]) T[i+j]=m;
117	}
118      }
119    }
120
121  array<int> penalty (N(s)-3);
122  for (i=2; i<N(s)-1; i++)
123    penalty [i-2]= (((T[i]&1)==1)? HYPH_STD: HYPH_INVALID);
124  if (N(penalty)>0) penalty[0] = penalty[N(penalty)-1] = HYPH_INVALID;
125  if (N(penalty)>1) penalty[1] = penalty[N(penalty)-2] = HYPH_INVALID;
126  if (N(penalty)>2) penalty[N(penalty)-3] = HYPH_INVALID;
127  // cout << "  -> " << penalty << "\n";
128  return penalty;
129}
130
131void
132std_hyphenate (string s, int after, string& left, string& right, int penalty) {
133  left = s (0, after+1);
134  right= s (after+1, N(s));
135  if (penalty >= HYPH_INVALID) left << string ("\\");
136  else left << string ("-");
137}