summaryrefslogtreecommitdiffstats
path: root/tools/linguist/shared/simtexth.cpp
blob: 31253d657efb9b96bef08ae9ac307b5db69d3bde (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
/****************************************************************************
**
** Copyright (C) 2011 Nokia Corporation and/or its subsidiary(-ies).
** All rights reserved.
** Contact: Nokia Corporation (qt-info@nokia.com)
**
** This file is part of the Qt Linguist of the Qt Toolkit.
**
** $QT_BEGIN_LICENSE:LGPL$
** GNU Lesser General Public License Usage
** This file may be used under the terms of the GNU Lesser General Public
** License version 2.1 as published by the Free Software Foundation and
** appearing in the file LICENSE.LGPL included in the packaging of this
** file. Please review the following information to ensure the GNU Lesser
** General Public License version 2.1 requirements will be met:
** http://www.gnu.org/licenses/old-licenses/lgpl-2.1.html.
**
** In addition, as a special exception, Nokia gives you certain additional
** rights. These rights are described in the Nokia Qt LGPL Exception
** version 1.1, included in the file LGPL_EXCEPTION.txt in this package.
**
** GNU General Public License Usage
** Alternatively, this file may be used under the terms of the GNU General
** Public License version 3.0 as published by the Free Software Foundation
** and appearing in the file LICENSE.GPL included in the packaging of this
** file. Please review the following information to ensure the GNU General
** Public License version 3.0 requirements will be met:
** http://www.gnu.org/copyleft/gpl.html.
**
** Other Usage
** Alternatively, this file may be used in accordance with the terms and
** conditions contained in a signed written agreement between you and Nokia.
**
**
**
**
**
** $QT_END_LICENSE$
**
****************************************************************************/

#include "simtexth.h"
#include "translator.h"

#include <QtCore/QByteArray>
#include <QtCore/QString>
#include <QtCore/QList>


QT_BEGIN_NAMESPACE

typedef QList<TranslatorMessage> TML;

/*
  How similar are two texts?  The approach used here relies on co-occurrence
  matrices and is very efficient.

  Let's see with an example: how similar are "here" and "hither"?  The
  co-occurrence matrix M for "here" is M[h,e] = 1, M[e,r] = 1, M[r,e] = 1, and 0
  elsewhere; the matrix N for "hither" is N[h,i] = 1, N[i,t] = 1, ...,
  N[h,e] = 1, N[e,r] = 1, and 0 elsewhere.  The union U of both matrices is the
  matrix U[i,j] = max { M[i,j], N[i,j] }, and the intersection V is
  V[i,j] = min { M[i,j], N[i,j] }.  The score for a pair of texts is

      score = (sum of V[i,j] over all i, j) / (sum of U[i,j] over all i, j),

  a formula suggested by Arnt Gulbrandsen.  Here we have

      score = 2 / 6,

  or one third.

  The implementation differs from this in a few details.  Most importantly,
  repetitions are ignored; for input "xxx", M[x,x] equals 1, not 2.
*/

/*
  Every character is assigned to one of 20 buckets so that the co-occurrence
  matrix requires only 20 * 20 = 400 bits, not 256 * 256 = 65536 bits or even
  more if we want the whole Unicode.  Which character falls in which bucket is
  arbitrary.

  The second half of the table is a replica of the first half, because of
  laziness.
*/
static const int indexOf[256] = {
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
//      !   "   #   $   %   &   '   (   )   *   +   ,   -   .   /
    0,  2,  6,  7,  10, 12, 15, 19, 2,  6,  7,  10, 12, 15, 19, 0,
//  0   1   2   3   4   5   6   7   8   9   :   ;   <   =   >   ?
    1,  3,  4,  5,  8,  9,  11, 13, 14, 16, 2,  6,  7,  10, 12, 15,
//  @   A   B   C   D   E   F   G   H   I   J   K   L   M   N   O
    0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  6,  10, 11, 12, 13, 14,
//  P   Q   R   S   T   U   V   W   X   Y   Z   [   \   ]   ^   _
    15, 12, 16, 17, 18, 19, 2,  10, 15, 7,  19, 2,  6,  7,  10, 0,
//  `   a   b   c   d   e   f   g   h   i   j   k   l   m   n   o
    0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  6,  10, 11, 12, 13, 14,
//  p   q   r   s   t   u   v   w   x   y   z   {   |   }   ~
    15, 12, 16, 17, 18, 19, 2,  10, 15, 7,  19, 2,  6,  7,  10, 0,

    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  2,  6,  7,  10, 12, 15, 19, 2,  6,  7,  10, 12, 15, 19, 0,
    1,  3,  4,  5,  8,  9,  11, 13, 14, 16, 2,  6,  7,  10, 12, 15,
    0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  6,  10, 11, 12, 13, 14,
    15, 12, 16, 17, 18, 19, 2,  10, 15, 7,  19, 2,  6,  7,  10, 0,
    0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  6,  10, 11, 12, 13, 14,
    15, 12, 16, 17, 18, 19, 2,  10, 15, 7,  19, 2,  6,  7,  10, 0
};

/*
  The entry bitCount[i] (for i between 0 and 255) is the number of bits used to
  represent i in binary.
*/
static const int bitCount[256] = {
    0,  1,  1,  2,  1,  2,  2,  3,  1,  2,  2,  3,  2,  3,  3,  4,
    1,  2,  2,  3,  2,  3,  3,  4,  2,  3,  3,  4,  3,  4,  4,  5,
    1,  2,  2,  3,  2,  3,  3,  4,  2,  3,  3,  4,  3,  4,  4,  5,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    1,  2,  2,  3,  2,  3,  3,  4,  2,  3,  3,  4,  3,  4,  4,  5,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    3,  4,  4,  5,  4,  5,  5,  6,  4,  5,  5,  6,  5,  6,  6,  7,
    1,  2,  2,  3,  2,  3,  3,  4,  2,  3,  3,  4,  3,  4,  4,  5,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    3,  4,  4,  5,  4,  5,  5,  6,  4,  5,  5,  6,  5,  6,  6,  7,
    2,  3,  3,  4,  3,  4,  4,  5,  3,  4,  4,  5,  4,  5,  5,  6,
    3,  4,  4,  5,  4,  5,  5,  6,  4,  5,  5,  6,  5,  6,  6,  7,
    3,  4,  4,  5,  4,  5,  5,  6,  4,  5,  5,  6,  5,  6,  6,  7,
    4,  5,  5,  6,  5,  6,  6,  7,  5,  6,  6,  7,  6,  7,  7,  8
};

struct CoMatrix
{
    /*
      The matrix has 20 * 20 = 400 entries.  This requires 50 bytes, or 13
      words.  Some operations are performed on words for more efficiency.
    */
    union {
        quint8 b[52];
        quint32 w[13];
    };

    CoMatrix() { memset( b, 0, 52 ); }

    CoMatrix(const QString &str)
    {
        QByteArray ba = str.toUtf8();
        const char *text = ba.constData();
        char c = '\0', d;
        memset( b, 0, 52 );
        /*
          The Knuth books are not in the office only for show; they help make
          loops 30% faster and 20% as readable.
        */
        while ( (d = *text) != '\0' ) {
            setCoOccurence( c, d );
            if ( (c = *++text) != '\0' ) {
                setCoOccurence( d, c );
                text++;
            }
        }
    }

    void setCoOccurence( char c, char d ) {
        int k = indexOf[(uchar) c] + 20 * indexOf[(uchar) d];
        b[k >> 3] |= (1 << (k & 0x7));
    }

    int worth() const {
        int w = 0;
        for ( int i = 0; i < 50; i++ )
            w += bitCount[b[i]];
        return w;
    }
};

static inline CoMatrix reunion(const CoMatrix &m, const CoMatrix &n)
{
    CoMatrix p;
    for (int i = 0; i < 13; ++i)
        p.w[i] = m.w[i] | n.w[i];
    return p;
}

static inline CoMatrix intersection(const CoMatrix &m, const CoMatrix &n)
{
    CoMatrix p;
    for (int i = 0; i < 13; ++i)
        p.w[i] = m.w[i] & n.w[i];
    return p;
}

StringSimilarityMatcher::StringSimilarityMatcher(const QString &stringToMatch)
{
    m_cm = new CoMatrix(stringToMatch);
    m_length = stringToMatch.length();
}

int StringSimilarityMatcher::getSimilarityScore(const QString &strCandidate)
{
    CoMatrix cmTarget(strCandidate);
    int delta = qAbs(m_length - strCandidate.size());
    int score = ( (intersection(*m_cm, cmTarget).worth() + 1) << 10 ) /
        ( reunion(*m_cm, cmTarget).worth() + (delta << 1) + 1 );
    return score;
}

StringSimilarityMatcher::~StringSimilarityMatcher()
{
    delete m_cm;
}

/**
 * Checks how similar two strings are.
 * The return value is the score, and a higher score is more similar
 * than one with a low score.
 * Linguist considers a score over 190 to be a good match.
 * \sa StringSimilarityMatcher
 */
int getSimilarityScore(const QString &str1, const QString &str2)
{
    CoMatrix cmTarget(str2);
    CoMatrix cm(str1);
    int delta = qAbs(str1.size() - str2.size());

    int score = ( (intersection(cm, cmTarget).worth() + 1) << 10 )
        / ( reunion(cm, cmTarget).worth() + (delta << 1) + 1 );

    return score;
}

CandidateList similarTextHeuristicCandidates(const Translator *tor,
    const QString &text, int maxCandidates)
{
    QList<int> scores;
    CandidateList candidates;

    TML all = tor->translatedMessages();

    foreach (const TranslatorMessage &mtm, all) {
        if (mtm.type() == TranslatorMessage::Unfinished
            || mtm.translation().isEmpty())
            continue;

        QString s = mtm.sourceText();
        int score = getSimilarityScore(s, text);

        if (candidates.size() == maxCandidates && score > scores[maxCandidates - 1] )
            candidates.removeLast();

        if (candidates.size() < maxCandidates && score >= textSimilarityThreshold) {
            Candidate cand( s, mtm.translation() );

            int i;
            for (i = 0; i < candidates.size(); i++) {
                if (score >= scores.at(i)) {
                    if (score == scores.at(i)) {
                        if (candidates.at(i) == cand)
                            goto continue_outer_loop;
                    } else {
                        break;
                    }
                }
            }
            scores.insert(i, score);
            candidates.insert(i, cand);
        }
        continue_outer_loop:
        ;
    }
    return candidates;
}

QT_END_NAMESPACE