Debian Codegraph @

src/nsHebrewProber.cpp

raw
1 /* -*- Mode: C; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
2 /* ***** BEGIN LICENSE BLOCK *****
3 * Version: MPL 1.1/GPL 2.0/LGPL 2.1
4 *
5 * The contents of this file are subject to the Mozilla Public License Version
6 * 1.1 (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 * http://www.mozilla.org/MPL/
9 *
10 * Software distributed under the License is distributed on an "AS IS" basis,
11 * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
12 * for the specific language governing rights and limitations under the
13 * License.
14 *
15 * The Original Code is Mozilla Universal charset detector code.
16 *
17 * The Initial Developer of the Original Code is
18 * Shy Shalom <[email protected]>
19 * Portions created by the Initial Developer are Copyright (C) 2005
20 * the Initial Developer. All Rights Reserved.
21 *
22 * Contributor(s):
23 *
24 * Alternatively, the contents of this file may be used under the terms of
25 * either the GNU General Public License Version 2 or later (the "GPL"), or
26 * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
27 * in which case the provisions of the GPL or the LGPL are applicable instead
28 * of those above. If you wish to allow use of your version of this file only
29 * under the terms of either the GPL or the LGPL, and not to allow others to
30 * use your version of this file under the terms of the MPL, indicate your
31 * decision by deleting the provisions above and replace them with the notice
32 * and other provisions required by the GPL or the LGPL. If you do not delete
33 * the provisions above, a recipient may use your version of this file under
34 * the terms of any one of the MPL, the GPL or the LGPL.
35 *
36 * ***** END LICENSE BLOCK ***** */
37
38 #include "nsHebrewProber.h"
39 #include <stdio.h>
40
41 // windows-1255 / ISO-8859-8 code points of interest
42 #define FINAL_KAF ('\xea')
43 #define NORMAL_KAF ('\xeb')
44 #define FINAL_MEM ('\xed')
45 #define NORMAL_MEM ('\xee')
46 #define FINAL_NUN ('\xef')
47 #define NORMAL_NUN ('\xf0')
48 #define FINAL_PE ('\xf3')
49 #define NORMAL_PE ('\xf4')
50 #define FINAL_TSADI ('\xf5')
51 #define NORMAL_TSADI ('\xf6')
52
53 // Minimum Visual vs Logical final letter score difference.
54 // If the difference is below this, don't rely solely on the final letter score distance.
55 #define MIN_FINAL_CHAR_DISTANCE (5)
56
57 // Minimum Visual vs Logical model score difference.
58 // If the difference is below this, don't rely at all on the model score distance.
59 #define MIN_MODEL_DISTANCE (0.01)
60
61 #define VISUAL_HEBREW_NAME ("ISO-8859-8")
62 #define LOGICAL_HEBREW_NAME ("WINDOWS-1255")
63
64 PRBool nsHebrewProber::isFinal(char c)
65 {
66 return ((c == FINAL_KAF) || (c == FINAL_MEM) || (c == FINAL_NUN) || (c == FINAL_PE) || (c == FINAL_TSADI));
67 }
68
69 PRBool nsHebrewProber::isNonFinal(char c)
70 {
71 return ((c == NORMAL_KAF) || (c == NORMAL_MEM) || (c == NORMAL_NUN) || (c == NORMAL_PE));
72 // The normal Tsadi is not a good Non-Final letter due to words like
73 // 'lechotet' (to chat) containing an apostrophe after the tsadi. This
74 // apostrophe is converted to a space in FilterWithoutEnglishLetters causing
75 // the Non-Final tsadi to appear at an end of a word even though this is not
76 // the case in the original text.
77 // The letters Pe and Kaf rarely display a related behavior of not being a
78 // good Non-Final letter. Words like 'Pop', 'Winamp' and 'Mubarak' for
79 // example legally end with a Non-Final Pe or Kaf. However, the benefit of
80 // these letters as Non-Final letters outweighs the damage since these words
81 // are quite rare.
82 }
83
84 /** HandleData
85 * Final letter analysis for logical-visual decision.
86 * Look for evidence that the received buffer is either logical Hebrew or
87 * visual Hebrew.
88 * The following cases are checked:
89 * 1) A word longer than 1 letter, ending with a final letter. This is an
90 * indication that the text is laid out "naturally" since the final letter
91 * really appears at the end. +1 for logical score.
92 * 2) A word longer than 1 letter, ending with a Non-Final letter. In normal
93 * Hebrew, words ending with Kaf, Mem, Nun, Pe or Tsadi, should not end with
94 * the Non-Final form of that letter. Exceptions to this rule are mentioned
95 * above in isNonFinal(). This is an indication that the text is laid out
96 * backwards. +1 for visual score
97 * 3) A word longer than 1 letter, starting with a final letter. Final letters
98 * should not appear at the beginning of a word. This is an indication that
99 * the text is laid out backwards. +1 for visual score.
100 *
101 * The visual score and logical score are accumulated throughout the text and
102 * are finally checked against each other in GetCharSetName().
103 * No checking for final letters in the middle of words is done since that case
104 * is not an indication for either Logical or Visual text.
105 *
106 * The input buffer should not contain any white spaces that are not (' ')
107 * or any low-ascii punctuation marks.
108 */
109 nsProbingState nsHebrewProber::HandleData(const char* aBuf, PRUint32 aLen)
110 {
111 // Both model probers say it's not them. No reason to continue.
112 if (GetState() == eNotMe)
113 return eNotMe;
114
115 const char *curPtr, *endPtr = aBuf+aLen;
116 char cur;
117
118 for (curPtr = (char*)aBuf; curPtr < endPtr; ++curPtr)
119 {
120 cur = *curPtr;
121 if (cur == ' ') // We stand on a space - a word just ended
122 {
123 if (mBeforePrev != ' ') // *(curPtr-2) was not a space so prev is not a 1 letter word
124 {
125 if (isFinal(mPrev)) // case (1) [-2:not space][-1:final letter][cur:space]
126 ++mFinalCharLogicalScore;
127 else if (isNonFinal(mPrev)) // case (2) [-2:not space][-1:Non-Final letter][cur:space]
128 ++mFinalCharVisualScore;
129 }
130 }
131 else // Not standing on a space
132 {
133 if ((mBeforePrev == ' ') && (isFinal(mPrev)) && (cur != ' ')) // case (3) [-2:space][-1:final letter][cur:not space]
134 ++mFinalCharVisualScore;
135 }
136 mBeforePrev = mPrev;
137 mPrev = cur;
138 }
139
140 // Forever detecting, till the end or until both model probers return eNotMe (handled above).
141 return eDetecting;
142 }
143
144 // Make the decision: is it Logical or Visual?
145 const char* nsHebrewProber::GetCharSetName()
146 {
147 // If the final letter score distance is dominant enough, rely on it.
148 PRInt32 finalsub = mFinalCharLogicalScore - mFinalCharVisualScore;
149 if (finalsub >= MIN_FINAL_CHAR_DISTANCE)
150 return LOGICAL_HEBREW_NAME;
151 if (finalsub <= -(MIN_FINAL_CHAR_DISTANCE))
152 return VISUAL_HEBREW_NAME;
153
154 // It's not dominant enough, try to rely on the model scores instead.
155 float modelsub = mLogicalProb->GetConfidence() - mVisualProb->GetConfidence();
156 if (modelsub > MIN_MODEL_DISTANCE)
157 return LOGICAL_HEBREW_NAME;
158 if (modelsub < -(MIN_MODEL_DISTANCE))
159 return VISUAL_HEBREW_NAME;
160
161 // Still no good, back to final letter distance, maybe it'll save the day.
162 if (finalsub < 0)
163 return VISUAL_HEBREW_NAME;
164
165 // (finalsub > 0 - Logical) or (don't know what to do) default to Logical.
166 return LOGICAL_HEBREW_NAME;
167 }
168
169
170 void nsHebrewProber::Reset(void)
171 {
172 mFinalCharLogicalScore = 0;
173 mFinalCharVisualScore = 0;
174
175 // mPrev and mBeforePrev are initialized to space in order to simulate a word
176 // delimiter at the beginning of the data
177 mPrev = ' ';
178 mBeforePrev = ' ';
179 }
180
181 nsProbingState nsHebrewProber::GetState(void)
182 {
183 // Remain active as long as any of the model probers are active.
184 if ((mLogicalProb->GetState() == eNotMe) && (mVisualProb->GetState() == eNotMe))
185 return eNotMe;
186 return eDetecting;
187 }
188
189 #ifdef DEBUG_chardet
190 void nsHebrewProber::DumpStatus()
191 {
192 printf(" HEB: %d - %d [Logical-Visual score]\r\n", mFinalCharLogicalScore, mFinalCharVisualScore);
193 }
194 #endif