#include <unicodetext.h>

Public Types
typedef std::bidirectional_iterator_tag	iterator_category

typedef char32	value_type

typedef ptrdiff_t	difference_type

typedef void	pointer

Public Member Functions
	const_iterator ()

	const_iterator (const const_iterator &other)

const_iterator &	operator= (const const_iterator &other)

char32	operator* () const

const_iterator &	operator++ ()

const_iterator	operator++ (int)

const_iterator &	operator-- ()

const_iterator	operator-- (int)

int	get_utf8 (char *buf) const

string	get_utf8_string () const

int	utf8_length () const

const char *	utf8_data () const

string	DebugString () const

Public Attributes
const typedef char32	reference

Friends
class	UnicodeText

class	UnicodeTextUtils

class	UTF8StateTableProperty

bool	operator== (const CI &lhs, const CI &rhs)

bool	operator!= (const CI &lhs, const CI &rhs)

bool	operator< (const CI &lhs, const CI &rhs)

bool	operator> (const CI &lhs, const CI &rhs)

bool	operator<= (const CI &lhs, const CI &rhs)

bool	operator>= (const CI &lhs, const CI &rhs)

difference_type	distance (const CI &first, const CI &last)

Detailed Description

Definition at line 176 of file unicodetext.h.

Member Typedef Documentation

◆ difference_type

typedef ptrdiff_t UnicodeText::const_iterator::difference_type

Definition at line 181 of file unicodetext.h.

◆ iterator_category

typedef std::bidirectional_iterator_tag UnicodeText::const_iterator::iterator_category

Definition at line 179 of file unicodetext.h.

◆ pointer

typedef void UnicodeText::const_iterator::pointer

Definition at line 182 of file unicodetext.h.

◆ value_type

typedef char32 UnicodeText::const_iterator::value_type

Definition at line 180 of file unicodetext.h.

Constructor & Destructor Documentation

◆ const_iterator() [1/2]

UnicodeText::const_iterator::const_iterator ( )

Definition at line 395 of file unicodetext.cc.

395 : it_(nullptr) {}

◆ const_iterator() [2/2]

UnicodeText::const_iterator::const_iterator ( const const_iterator & other )

Definition at line 397 of file unicodetext.cc.

398 : it_(other.it_) {

399 }

Member Function Documentation

◆ DebugString()

string UnicodeText::const_iterator::DebugString ( ) const

Definition at line 495 of file unicodetext.cc.

                                                     {
   return tensorflow::strings::Printf("{iter %p}", it_);
 }

◆ get_utf8()

int UnicodeText::const_iterator::get_utf8 ( char * buf ) const

Definition at line 460 of file unicodetext.cc.

                                                                {
   utf8_output[0] = it_[0]; if ((it_[0] & 0xff) < 0x80) return 1;
   utf8_output[1] = it_[1]; if ((it_[0] & 0xff) < 0xE0) return 2;
   utf8_output[2] = it_[2]; if ((it_[0] & 0xff) < 0xF0) return 3;
   utf8_output[3] = it_[3];
   return 4;
 }

◆ get_utf8_string()

string UnicodeText::const_iterator::get_utf8_string ( ) const

Definition at line 468 of file unicodetext.cc.

                                                         {
   return string(utf8_data(), utf8_length());
 }

◆ operator*()

char32 UnicodeText::const_iterator::operator* ( ) const

Definition at line 421 of file unicodetext.cc.

                                                   {
   // (We could call chartorune here, but that does some
   // error-checking, and we're guaranteed that our data is valid
   // UTF-8. Also, we expect this routine to be called very often. So
   // for speed, we do the calculation ourselves.)
  
   // Convert from UTF-8
   unsigned char byte1 = it_[0];
   if (byte1 < 0x80)
     return byte1;
  
   unsigned char byte2 = it_[1];
   if (byte1 < 0xE0)
     return ((byte1 & 0x1F) << 6)
           | (byte2 & 0x3F);
  
   unsigned char byte3 = it_[2];
   if (byte1 < 0xF0)
     return ((byte1 & 0x0F) << 12)
          | ((byte2 & 0x3F) << 6)
          |  (byte3 & 0x3F);
  
   unsigned char byte4 = it_[3];
   return ((byte1 & 0x07) << 18)
        | ((byte2 & 0x3F) << 12)
        | ((byte3 & 0x3F) << 6)
        |  (byte4 & 0x3F);
 }

◆ operator++() [1/2]

UnicodeText::const_iterator & UnicodeText::const_iterator::operator++ ( )

Definition at line 450 of file unicodetext.cc.

                                                                  {
   it_ += UniLib::OneCharLen(it_);
   return *this;
 }

◆ operator++() [2/2]

const_iterator UnicodeText::const_iterator::operator++ ( int )

inline

Definition at line 195 of file unicodetext.h.

                                    {  // (iter++)
       const_iterator result(*this);
       ++*this;
       return result;
     }

◆ operator--() [1/2]

UnicodeText::const_iterator & UnicodeText::const_iterator::operator-- ( )

Definition at line 455 of file unicodetext.cc.

                                                                  {
   while (UniLib::IsTrailByte(*--it_));
   return *this;
 }

◆ operator--() [2/2]

const_iterator UnicodeText::const_iterator::operator-- ( int )

inline

Definition at line 202 of file unicodetext.h.

                                    {  // (iter--)
       const_iterator result(*this);
       --*this;
       return result;
     }

◆ operator=()

UnicodeText::const_iterator & UnicodeText::const_iterator::operator= ( const const_iterator & other )

Definition at line 402 of file unicodetext.cc.

                                                                 {
   if (&other != this)
     it_ = other.it_;
   return *this;
 }

◆ utf8_data()

const char* UnicodeText::const_iterator::utf8_data ( ) const

inline

Definition at line 233 of file unicodetext.h.

233 { return it_; }

◆ utf8_length()

int UnicodeText::const_iterator::utf8_length ( ) const

Definition at line 472 of file unicodetext.cc.

                                                  {
   if ((it_[0] & 0xff) < 0x80) {
     return 1;
   } else if ((it_[0] & 0xff) < 0xE0) {
     return 2;
   } else if ((it_[0] & 0xff) < 0xF0) {
     return 3;
   } else {
     return 4;
   }
 }

Friends And Related Function Documentation

◆ distance

difference_type distance	(	const CI &	first,
		const CI &	last
	)

friend

Definition at line 44 of file unicodetext.cc.

                                                 {
   return CodepointDistance(first.it_, last.it_);
 }

◆ operator!=

bool operator!=	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 211 of file unicodetext.h.

211 {

212 return !(lhs == rhs); }

◆ operator<

bool operator<	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 416 of file unicodetext.cc.

                                                      {
   return lhs.it_ < rhs.it_;
 }

◆ operator<=

bool operator<=	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 216 of file unicodetext.h.

216 {

217 return !(rhs < lhs); }

◆ operator==

bool operator==	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 209 of file unicodetext.h.

209 {

210 return lhs.it_ == rhs.it_; }

◆ operator>

bool operator>	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 214 of file unicodetext.h.

214 {

215 return rhs < lhs; }

◆ operator>=

bool operator>=	(	const CI &	lhs,
		const CI &	rhs
	)

friend

Definition at line 218 of file unicodetext.h.

218 {

219 return !(lhs < rhs); }

◆ UnicodeText

friend class UnicodeText

friend

Definition at line 238 of file unicodetext.h.

◆ UnicodeTextUtils

friend class UnicodeTextUtils

friend

Definition at line 239 of file unicodetext.h.

◆ UTF8StateTableProperty

friend class UTF8StateTableProperty

friend

Definition at line 240 of file unicodetext.h.

Member Data Documentation

◆ reference

const typedef char32 UnicodeText::const_iterator::reference

Definition at line 183 of file unicodetext.h.

The documentation for this class was generated from the following files:

/tesseract/unittest/util/utf8/unicodetext.h
/tesseract/unittest/util/utf8/unicodetext.cc

Public Types

Public Member Functions

Public Attributes

Friends

Detailed Description

Member Typedef Documentation

◆ difference_type

◆ iterator_category

◆ pointer

◆ value_type

Constructor & Destructor Documentation

◆ const_iterator() [1/2]

◆ const_iterator() [2/2]

Member Function Documentation

◆ DebugString()

◆ get_utf8()

◆ get_utf8_string()

◆ operator*()

◆ operator++() [1/2]

◆ operator++() [2/2]

◆ operator--() [1/2]

◆ operator--() [2/2]

◆ operator=()

◆ utf8_data()

◆ utf8_length()

Friends And Related Function Documentation

◆ distance

◆ operator!=

◆ operator<

◆ operator<=

◆ operator==

◆ operator>

◆ operator>=

◆ UnicodeText

◆ UnicodeTextUtils

◆ UTF8StateTableProperty

Member Data Documentation

◆ reference