Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/tesseract/src/training/unicharset/validate_myanmar.h @ 3:2c135c81b16c

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
MERGE: upstream PyMuPDF 1.26.4 with MuPDF 1.26.7
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:44:09 +0200
parents b50eed0cc0ef
children
comparison
equal deleted inserted replaced
0:6015a75abc2d 3:2c135c81b16c
1 #ifndef TESSERACT_TRAINING_VALIDATE_MYANMAR_H_
2 #define TESSERACT_TRAINING_VALIDATE_MYANMAR_H_
3
4 #include "validator.h"
5
6 namespace tesseract {
7
8 // Subclass of Validator that validates and segments Myanmar.
9 class ValidateMyanmar : public Validator {
10 public:
11 ValidateMyanmar(ViramaScript script, bool report_errors) : Validator(script, report_errors) {}
12 ~ValidateMyanmar() override = default;
13
14 protected:
15 // Returns whether codes matches the pattern for a Myanmar Grapheme.
16 // Consumes the next Grapheme in codes_[codes_used_++...] and copies it to
17 // parts_ and output_. Returns true if a valid Grapheme was consumed,
18 // otherwise does not increment codes_used_.
19 bool ConsumeGraphemeIfValid() override;
20 // Returns the CharClass corresponding to the given Unicode ch.
21 Validator::CharClass UnicodeToCharClass(char32 ch) const override;
22
23 private:
24 // Helper consumes/copies a virama and any subscript consonant.
25 // Returns true if the end of input is reached.
26 bool ConsumeSubscriptIfPresent();
27 // Helper consumes/copies a series of optional signs.
28 // Returns true if the end of input is reached.
29 bool ConsumeOptionalSignsIfPresent();
30 // Returns true if the unicode is a Myanmar "letter" including consonants
31 // and independent vowels. Although table 16-3 distinguishes between some
32 // base consonants and vowels, the extensions make no such distinction, so we
33 // put them all into a single bucket.
34 static bool IsMyanmarLetter(char32 ch);
35 // Returns true if ch is a Myanmar digit or other symbol that does not take
36 // part in being a syllable.
37 static bool IsMyanmarOther(char32 ch);
38
39 // Some special unicodes used only for Myanmar processing.
40 static const char32 kMyanmarAsat = 0x103a;
41 static const char32 kMyanmarMedialYa = 0x103b;
42 };
43
44 } // namespace tesseract
45
46 #endif // TESSERACT_TRAINING_VALIDATE_MYANMAR_H_