| 1 | // Copyright (C) 2016 The Qt Company Ltd. |
|---|
| 2 | // SPDX-License-Identifier: LicenseRef-Qt-Commercial OR GPL-3.0-only WITH Qt-GPL-exception-1.0 |
|---|
| 3 | |
|---|
| 4 | #ifndef SIMTEXTH_H |
|---|
| 5 | #define SIMTEXTH_H |
|---|
| 6 | |
|---|
| 7 | const int textSimilarityThreshold = 190; |
|---|
| 8 | |
|---|
| 9 | #include <QString> |
|---|
| 10 | #include <QList> |
|---|
| 11 | |
|---|
| 12 | QT_BEGIN_NAMESPACE |
|---|
| 13 | |
|---|
| 14 | class Translator; |
|---|
| 15 | |
|---|
| 16 | struct Candidate |
|---|
| 17 | { |
|---|
| 18 | Candidate() {} |
|---|
| 19 | Candidate(const QString &c, const QString &s, const QString &d, const QString &t) |
|---|
| 20 | : context(c), source(s), disambiguation(d), translation(t) |
|---|
| 21 | {} |
|---|
| 22 | |
|---|
| 23 | QString context; |
|---|
| 24 | QString source; |
|---|
| 25 | QString disambiguation; |
|---|
| 26 | QString translation; |
|---|
| 27 | }; |
|---|
| 28 | |
|---|
| 29 | inline bool operator==( const Candidate& c, const Candidate& d ) { |
|---|
| 30 | return c.translation == d.translation && c.source == d.source && c.context == d.context |
|---|
| 31 | && c.disambiguation == d.disambiguation; |
|---|
| 32 | } |
|---|
| 33 | inline bool operator!=( const Candidate& c, const Candidate& d ) { |
|---|
| 34 | return !operator==( c, d ); |
|---|
| 35 | } |
|---|
| 36 | |
|---|
| 37 | typedef QList<Candidate> CandidateList; |
|---|
| 38 | |
|---|
| 39 | struct CoMatrix |
|---|
| 40 | { |
|---|
| 41 | CoMatrix(const QString &str); |
|---|
| 42 | CoMatrix() {} |
|---|
| 43 | |
|---|
| 44 | /* |
|---|
| 45 | The matrix has 20 * 20 = 400 entries. This requires 50 bytes, or 13 |
|---|
| 46 | words. Some operations are performed on words for more efficiency. |
|---|
| 47 | */ |
|---|
| 48 | union { |
|---|
| 49 | quint8 b[52]; |
|---|
| 50 | quint32 w[13]; |
|---|
| 51 | }; |
|---|
| 52 | }; |
|---|
| 53 | |
|---|
| 54 | /** |
|---|
| 55 | * This class is more efficient for searching through a large array of candidate strings, since we only |
|---|
| 56 | * have to construct the CoMatrix for the \a stringToMatch once, |
|---|
| 57 | * after that we just call getSimilarityScore(strCandidate). |
|---|
| 58 | * \sa getSimilarityScore |
|---|
| 59 | */ |
|---|
| 60 | class StringSimilarityMatcher { |
|---|
| 61 | public: |
|---|
| 62 | StringSimilarityMatcher(const QString &stringToMatch); |
|---|
| 63 | int getSimilarityScore(const QString &strCandidate); |
|---|
| 64 | |
|---|
| 65 | private: |
|---|
| 66 | CoMatrix m_cm; |
|---|
| 67 | int m_length; |
|---|
| 68 | }; |
|---|
| 69 | |
|---|
| 70 | /** |
|---|
| 71 | * Checks how similar two strings are. |
|---|
| 72 | * The return value is the score, and a higher score is more similar |
|---|
| 73 | * than one with a low score. |
|---|
| 74 | * Linguist considers a score over 190 to be a good match. |
|---|
| 75 | * \sa StringSimilarityMatcher |
|---|
| 76 | */ |
|---|
| 77 | static inline int getSimilarityScore(const QString &str1, const QString &str2) |
|---|
| 78 | { |
|---|
| 79 | return StringSimilarityMatcher(str1).getSimilarityScore(strCandidate: str2); |
|---|
| 80 | } |
|---|
| 81 | |
|---|
| 82 | CandidateList similarTextHeuristicCandidates( const Translator *tor, |
|---|
| 83 | const QString &text, |
|---|
| 84 | int maxCandidates ); |
|---|
| 85 | |
|---|
| 86 | QT_END_NAMESPACE |
|---|
| 87 | |
|---|
| 88 | #endif |
|---|
| 89 | |
|---|
