X-Git-Url: https://git.creatis.insa-lyon.fr/pubgit/?a=blobdiff_plain;f=src%2FgdcmDocument.cxx;h=8c0ddc3cc5c8195bf9d075e94958c3c7e51c2356;hb=5b89bede4607999aeb8d5b45311e7ee82f9471ef;hp=a252297e52f728abb71c16fadb83b91a4e18c491;hpb=7409414e04772fdb69056cfb8066f392c08014e1;p=gdcm.git diff --git a/src/gdcmDocument.cxx b/src/gdcmDocument.cxx index a252297e..8c0ddc3c 100644 --- a/src/gdcmDocument.cxx +++ b/src/gdcmDocument.cxx @@ -3,8 +3,8 @@ Program: gdcm Module: $RCSfile: gdcmDocument.cxx,v $ Language: C++ - Date: $Date: 2005/01/28 09:31:51 $ - Version: $Revision: 1.215 $ + Date: $Date: 2005/02/02 16:18:48 $ + Version: $Revision: 1.220 $ Copyright (c) CREATIS (Centre de Recherche et d'Applications en Traitement de l'Image). All rights reserved. See Doc/License.txt or @@ -42,7 +42,6 @@ namespace gdcm { - //----------------------------------------------------------------------------- // Refer to Document::CheckSwap() //const unsigned int Document::HEADER_LENGTH_TO_READ = 256; @@ -53,12 +52,13 @@ const unsigned int Document::MAX_SIZE_PRINT_ELEMENT_VALUE = 0x7fffffff; //----------------------------------------------------------------------------- // Constructor / Destructor - +// Constructors and destructors are protected to avoid user to invoke directly /** * \brief constructor * @param filename 'Document' (File or DicomDir) to be opened for parsing */ -Document::Document( std::string const &filename ) : ElementSet(-1) +Document::Document( std::string const &filename ) + :ElementSet(-1) { SetMaxSizeLoadEntry(MAX_SIZE_LOAD_ELEMENT_VALUE); Filename = filename; @@ -75,10 +75,10 @@ Document::Document( std::string const &filename ) : ElementSet(-1) gdcmVerboseMacro( "Starting parsing of file: " << Filename.c_str()); // Fp->seekg( 0, std::ios::beg); - Fp->seekg(0, std::ios::end); + Fp->seekg(0, std::ios::end); long lgt = Fp->tellg(); - Fp->seekg( 0, std::ios::beg); + Fp->seekg(0, std::ios::beg); CheckSwap(); long beg = Fp->tellg(); @@ -125,7 +125,7 @@ Document::Document( std::string const &filename ) : ElementSet(-1) CloseFile(); - // -------------------------------------------------------------- + // ---------------------------- // Specific code to allow gdcm to read ACR-LibIDO formated images // Note: ACR-LibIDO is an extension of the ACR standard that was // used at CREATIS. For the time being (say a couple years) @@ -147,14 +147,15 @@ Document::Document( std::string const &filename ) : ElementSet(-1) SetValEntry(columns, 0x0028, 0x0010); SetValEntry(rows , 0x0028, 0x0011); } - // ----------------- End of ACR-LibIDO kludge ------------------ + // --- End of ACR-LibIDO kludge --- } /** * \brief This default constructor doesn't parse the file. You should * then invoke \ref Document::SetFileName and then the parsing. */ -Document::Document() : ElementSet(-1) +Document::Document() + :ElementSet(-1) { Fp = 0; @@ -174,31 +175,6 @@ Document::~Document () RefShaDict = NULL; } -//----------------------------------------------------------------------------- -// Print - -/** - * \brief Prints The Dict Entries of THE public Dicom Dictionary - * @param os ostream to print to - * @return - */ -void Document::PrintPubDict(std::ostream &os) -{ - RefPubDict->SetPrintLevel(PrintLevel); - RefPubDict->Print(os); -} - -/** - * \brief Prints The Dict Entries of THE shadow Dicom Dictionary - * @param os ostream to print to - * @return - */ -void Document::PrintShaDict(std::ostream &os) -{ - RefShaDict->SetPrintLevel(PrintLevel); - RefShaDict->Print(os); -} - //----------------------------------------------------------------------------- // Public /** @@ -262,6 +238,46 @@ bool Document::IsReadable() return true; } +/** + * \brief Predicate for dicom version 3 file. + * @return True when the file is a dicom version 3. + */ +bool Document::IsDicomV3() +{ + // Checking if Transfer Syntax exists is enough + // Anyway, it's to late check if the 'Preamble' was found ... + // And ... would it be a rich idea to check ? + // (some 'no Preamble' DICOM images exist !) + return GetDocEntry(0x0002, 0x0010) != NULL; +} + +/** + * \brief Predicate for Papyrus file + * Dedicated to whomsoever it may concern + * @return True when the file is a Papyrus file. + */ +bool Document::IsPapyrus() +{ + // check for Papyrus private Sequence + DocEntry *e = GetDocEntry(0x0041, 0x1050); + if ( !e ) + return false; + // check if it's actually a Sequence + if ( !dynamic_cast(e) ) + return false; + return true; +} + +/** + * \brief returns the File Type + * (ACR, ACR_LIBIDO, ExplicitVR, ImplicitVR, Unknown) + * @return the FileType code + */ +FileType Document::GetFileType() +{ + return Filetype; +} + /** * \brief Accessor to the Transfer Syntax (when present) of the * current document (it internally handles reading the @@ -300,45 +316,77 @@ std::string Document::GetTransferSyntax() } /** - * \brief Predicate for dicom version 3 file. - * @return True when the file is a dicom version 3. + * \brief Accesses the info from 0002,0010 : Transfer Syntax and TS + * @return The full Transfer Syntax Name (as opposed to Transfer Syntax UID) */ -bool Document::IsDicomV3() +std::string Document::GetTransferSyntaxName() { - // Checking if Transfer Syntax exists is enough - // Anyway, it's to late check if the 'Preamble' was found ... - // And ... would it be a rich idea to check ? - // (some 'no Preamble' DICOM images exist !) - return GetDocEntry(0x0002, 0x0010) != NULL; -} + // use the TS (TS : Transfer Syntax) + std::string transferSyntax = GetEntryValue(0x0002,0x0010); + + if ( (transferSyntax.find(GDCM_NOTLOADED) < transferSyntax.length()) ) + { + gdcmErrorMacro( "Transfer Syntax not loaded. " << std::endl + << "Better you increase MAX_SIZE_LOAD_ELEMENT_VALUE" ); + return "Uncompressed ACR-NEMA"; + } + if ( transferSyntax == GDCM_UNFOUND ) + { + gdcmVerboseMacro( "Unfound Transfer Syntax (0002,0010)"); + return "Uncompressed ACR-NEMA"; + } + + // we do it only when we need it + const TSKey &tsName = Global::GetTS()->GetValue( transferSyntax ); + // Global::GetTS() is a global static you shall never try to delete it! + return tsName; +} +// +// --------------- Swap Code ------------------ /** - * \brief Predicate for Papyrus file - * Dedicated to whomsoever it may concern - * @return True when the file is a Papyrus file. + * \brief Swaps the bytes so they agree with the processor order + * @return The properly swaped 16 bits integer. */ -bool Document::IsPapyrus() +uint16_t Document::SwapShort(uint16_t a) { - // check for Papyrus private Sequence - DocEntry *e = GetDocEntry(0x0041, 0x1050); - if ( !e ) - return false; - // check if it's actually a Sequence - if ( !dynamic_cast(e) ) - return false; - return true; + if ( SwapCode == 4321 || SwapCode == 2143 ) + { + a = ((( a << 8 ) & 0x0ff00 ) | (( a >> 8 ) & 0x00ff ) ); + } + return a; } /** - * \brief returns the File Type - * (ACR, ACR_LIBIDO, ExplicitVR, ImplicitVR, Unknown) - * @return the FileType code + * \brief Swaps back the bytes of 4-byte long integer accordingly to + * processor order. + * @return The properly swaped 32 bits integer. */ -FileType Document::GetFileType() +uint32_t Document::SwapLong(uint32_t a) { - return Filetype; -} + switch (SwapCode) + { + case 1234 : + break; + case 4321 : + a=( ((a<<24) & 0xff000000) | ((a<<8) & 0x00ff0000) | + ((a>>8) & 0x0000ff00) | ((a>>24) & 0x000000ff) ); + break; + case 3412 : + a=( ((a<<16) & 0xffff0000) | ((a>>16) & 0x0000ffff) ); + break; + case 2143 : + a=( ((a<< 8) & 0xff00ff00) | ((a>>8) & 0x00ff00ff) ); + break; + default : + gdcmErrorMacro( "Unset swap code:" << SwapCode ); + a = 0; + } + return a; +} +// +// -----------------File I/O --------------- /** * \brief Tries to open the file \ref Document::Filename and * checks the preamble when existing. @@ -448,23 +496,23 @@ void Document::WriteContent(std::ofstream *fp, FileType filetype) fp->write("DICM", 4); } -/* - * \todo rewrite later, if really usefull - * - 'Group Length' element is optional in DICOM - * - but un-updated odd groups lengthes can causes pb - * (xmedcon breaker) - * - * if ( (filetype == ImplicitVR) || (filetype == ExplicitVR) ) - * UpdateGroupLength(false,filetype); - * if ( filetype == ACR) - * UpdateGroupLength(true,ACR); - */ - + /* + * \todo rewrite later, if really usefull + * - 'Group Length' element is optional in DICOM + * - but un-updated odd groups lengthes can causes pb + * (xmedcon breaker) + * + * if ( (filetype == ImplicitVR) || (filetype == ExplicitVR) ) + * UpdateGroupLength(false,filetype); + * if ( filetype == ACR) + * UpdateGroupLength(true,ACR); + */ + ElementSet::WriteContent(fp, filetype); // This one is recursive } -//----------------------------------------------------------------------------- -// Protected +// ----------------------------------------- +// Content entries /** * \brief Loads (from disk) the element content * when a string is not suitable @@ -524,32 +572,6 @@ void Document::LoadEntryBinArea(BinEntry *elem) CloseFile(); } -/** - * \brief Sets a 'non string' value to a given Dicom Element - * @param area area containing the 'non string' value - * @param group Group number of the searched Dicom Element - * @param elem Element number of the searched Dicom Element - * @return - */ -/*bool Document::SetEntryBinArea(uint8_t *area, - uint16_t group, uint16_t elem) -{ - DocEntry *currentEntry = GetDocEntry(group, elem); - if ( !currentEntry ) - { - return false; - } - - if ( BinEntry *binEntry = dynamic_cast(currentEntry) ) - { - binEntry->SetBinArea( area ); - return true; - } - - return false; -}*/ - - /** * \brief Loads the element while preserving the current * underlying file position indicator as opposed to @@ -568,101 +590,218 @@ void Document::LoadDocEntrySafe(DocEntry *entry) } /** - * \brief Swaps back the bytes of 4-byte long integer accordingly to - * processor order. - * @return The properly swaped 32 bits integer. + * \brief Compares two documents, according to \ref DicomDir rules + * \warning Does NOT work with ACR-NEMA files + * \todo Find a trick to solve the pb (use RET fields ?) + * @param document + * @return true if 'smaller' */ -uint32_t Document::SwapLong(uint32_t a) +bool Document::operator<(Document &document) { - switch (SwapCode) + // Patient Name + std::string s1 = GetEntryValue(0x0010,0x0010); + std::string s2 = document.GetEntryValue(0x0010,0x0010); + if(s1 < s2) { - case 1234 : - break; - case 4321 : - a=( ((a<<24) & 0xff000000) | ((a<<8) & 0x00ff0000) | - ((a>>8) & 0x0000ff00) | ((a>>24) & 0x000000ff) ); - break; - case 3412 : - a=( ((a<<16) & 0xffff0000) | ((a>>16) & 0x0000ffff) ); - break; - case 2143 : - a=( ((a<< 8) & 0xff00ff00) | ((a>>8) & 0x00ff00ff) ); - break; - default : - gdcmErrorMacro( "Unset swap code:" << SwapCode ); - a = 0; + return true; } - return a; -} - -/** - * \brief Unswaps back the bytes of 4-byte long integer accordingly to - * processor order. - * @return The properly unswaped 32 bits integer. - */ -uint32_t Document::UnswapLong(uint32_t a) -{ - return SwapLong(a); -} - -/** - * \brief Swaps the bytes so they agree with the processor order - * @return The properly swaped 16 bits integer. - */ -uint16_t Document::SwapShort(uint16_t a) -{ - if ( SwapCode == 4321 || SwapCode == 2143 ) + else if( s1 > s2 ) { - a = ((( a << 8 ) & 0x0ff00 ) | (( a >> 8 ) & 0x00ff ) ); + return false; } - return a; -} - -/** - * \brief Unswaps the bytes so they agree with the processor order - * @return The properly unswaped 16 bits integer. - */ -uint16_t Document::UnswapShort(uint16_t a) -{ - return SwapShort(a); -} - -//----------------------------------------------------------------------------- -// Private - -/** - * \brief Parses a DocEntrySet (Zero-level DocEntries or SQ Item DocEntries) - * @return length of the parsed set. - */ -void Document::ParseDES(DocEntrySet *set, long offset, - long l_max, bool delim_mode) -{ - DocEntry *newDocEntry = 0; - ValEntry *newValEntry; - BinEntry *newBinEntry; - SeqEntry *newSeqEntry; - VRKey vr; - bool used = false; - - while (true) + else { - if ( !delim_mode && ((long)(Fp->tellg())-offset) >= l_max) + // Patient ID + s1 = GetEntryValue(0x0010,0x0020); + s2 = document.GetEntryValue(0x0010,0x0020); + if ( s1 < s2 ) { - break; + return true; } - - used = true; - newDocEntry = ReadNextDocEntry( ); - - if ( !newDocEntry ) + else if ( s1 > s2 ) { - break; + return false; } - - vr = newDocEntry->GetVR(); - newValEntry = dynamic_cast(newDocEntry); - newBinEntry = dynamic_cast(newDocEntry); - newSeqEntry = dynamic_cast(newDocEntry); + else + { + // Study Instance UID + s1 = GetEntryValue(0x0020,0x000d); + s2 = document.GetEntryValue(0x0020,0x000d); + if ( s1 < s2 ) + { + return true; + } + else if( s1 > s2 ) + { + return false; + } + else + { + // Serie Instance UID + s1 = GetEntryValue(0x0020,0x000e); + s2 = document.GetEntryValue(0x0020,0x000e); + if ( s1 < s2 ) + { + return true; + } + else if( s1 > s2 ) + { + return false; + } + } + } + } + return false; +} + +//----------------------------------------------------------------------------- +// Protected +/** + * \brief Reads a supposed to be 16 Bits integer + * (swaps it depending on processor endianity) + * @return read value + */ +uint16_t Document::ReadInt16() + throw( FormatError ) +{ + uint16_t g; + Fp->read ((char*)&g, (size_t)2); + if ( Fp->fail() ) + { + throw FormatError( "Document::ReadInt16()", " file error." ); + } + if( Fp->eof() ) + { + throw FormatError( "Document::ReadInt16()", "EOF." ); + } + g = SwapShort(g); + return g; +} + +/** + * \brief Reads a supposed to be 32 Bits integer + * (swaps it depending on processor endianity) + * @return read value + */ +uint32_t Document::ReadInt32() + throw( FormatError ) +{ + uint32_t g; + Fp->read ((char*)&g, (size_t)4); + if ( Fp->fail() ) + { + throw FormatError( "Document::ReadInt32()", " file error." ); + } + if( Fp->eof() ) + { + throw FormatError( "Document::ReadInt32()", "EOF." ); + } + g = SwapLong(g); + return g; +} + +/** + * \brief skips bytes inside the source file + * \warning NOT end user intended method ! + * @return + */ +void Document::SkipBytes(uint32_t nBytes) +{ + //FIXME don't dump the returned value + Fp->seekg((long)nBytes, std::ios::cur); +} + +/** + * \brief Re-computes the length of a ACR-NEMA/Dicom group from a DcmHeader + * @param filetype Type of the File to be written + */ +int Document::ComputeGroup0002Length( FileType filetype ) +{ + uint16_t gr; + std::string vr; + + int groupLength = 0; + bool found0002 = false; + + // for each zero-level Tag in the DCM Header + DocEntry *entry = GetFirstEntry(); + while( entry ) + { + gr = entry->GetGroup(); + + if( gr == 0x0002 ) + { + found0002 = true; + + if( entry->GetElement() != 0x0000 ) + { + vr = entry->GetVR(); + + if( filetype == ExplicitVR ) + { + if ( (vr == "OB") || (vr == "OW") || (vr == "SQ") ) + { + // explicit VR AND OB, OW, SQ : 4 more bytes + groupLength += 4; + } + } + groupLength += 2 + 2 + 4 + entry->GetLength(); + } + } + else if (found0002 ) + break; + + entry = GetNextEntry(); + } + return groupLength; +} + +//----------------------------------------------------------------------------- +// Private +/** + * \brief Loads all the needed Dictionaries + * \warning NOT end user intended method ! + */ +void Document::Initialize() +{ + RefPubDict = Global::GetDicts()->GetDefaultPubDict(); + RefShaDict = NULL; + Filetype = Unknown; +} + +/** + * \brief Parses a DocEntrySet (Zero-level DocEntries or SQ Item DocEntries) + * @return length of the parsed set. + */ +void Document::ParseDES(DocEntrySet *set, long offset, + long l_max, bool delim_mode) +{ + DocEntry *newDocEntry = 0; + ValEntry *newValEntry; + BinEntry *newBinEntry; + SeqEntry *newSeqEntry; + VRKey vr; + bool used = false; + + while (true) + { + if ( !delim_mode && ((long)(Fp->tellg())-offset) >= l_max) + { + break; + } + + used = true; + newDocEntry = ReadNextDocEntry( ); + + if ( !newDocEntry ) + { + break; + } + + vr = newDocEntry->GetVR(); + newValEntry = dynamic_cast(newDocEntry); + newBinEntry = dynamic_cast(newDocEntry); + newSeqEntry = dynamic_cast(newDocEntry); if ( newValEntry || newBinEntry ) { @@ -885,9 +1024,9 @@ void Document::ParseSQ( SeqEntry *seqEntry, } /** - * \brief Loads the element content if its length doesn't exceed - * the value specified with Document::SetMaxSizeLoadEntry() - * @param entry Header Entry (Dicom Element) to be dealt with + * \brief Loads the element content if its length doesn't exceed + * the value specified with Document::SetMaxSizeLoadEntry() + * @param entry Header Entry (Dicom Element) to be dealt with */ void Document::LoadDocEntry(DocEntry *entry) { @@ -1056,7 +1195,6 @@ void Document::LoadDocEntry(DocEntry *entry) } } - /** * \brief Find the value Length of the passed Header Entry * @param entry Header Entry whose length of the value shall be loaded. @@ -1204,6 +1342,63 @@ void Document::FindDocEntryLength( DocEntry *entry ) } } +/** + * \brief Find the Length till the next sequence delimiter + * \warning NOT end user intended method ! + * @return + */ +uint32_t Document::FindDocEntryLengthOBOrOW() + throw( FormatUnexpected ) +{ + // See PS 3.5-2001, section A.4 p. 49 on encapsulation of encoded pixel data. + long positionOnEntry = Fp->tellg(); + bool foundSequenceDelimiter = false; + uint32_t totalLength = 0; + + while ( !foundSequenceDelimiter ) + { + uint16_t group; + uint16_t elem; + try + { + group = ReadInt16(); + elem = ReadInt16(); + } + catch ( FormatError ) + { + throw FormatError("Unexpected end of file encountered during ", + "Document::FindDocEntryLengthOBOrOW()"); + } + // We have to decount the group and element we just read + totalLength += 4; + if ( group != 0xfffe || ( ( elem != 0xe0dd ) && ( elem != 0xe000 ) ) ) + { + long filePosition = Fp->tellg(); + gdcmVerboseMacro( "Neither an Item tag nor a Sequence delimiter tag on :" + << std::hex << group << " , " << elem + << ") -before- position x(" << filePosition << ")" ); + + Fp->seekg(positionOnEntry, std::ios::beg); + throw FormatUnexpected( "Neither an Item tag nor a Sequence delimiter tag."); + } + if ( elem == 0xe0dd ) + { + foundSequenceDelimiter = true; + } + uint32_t itemLength = ReadInt32(); + // We add 4 bytes since we just read the ItemLength with ReadInt32 + totalLength += itemLength + 4; + SkipBytes(itemLength); + + if ( foundSequenceDelimiter ) + { + break; + } + } + Fp->seekg( positionOnEntry, std::ios::beg); + return totalLength; +} + /** * \brief Find the Value Representation of the current Dicom Element. * @return Value Representation of the current Entry @@ -1244,9 +1439,6 @@ std::string Document::FindDocEntryVR() */ bool Document::CheckDocEntryVR(VRKey vr) { - // CLEANME searching the dicom_vr at each occurence is expensive. - // PostPone this test in an optional integrity check at the end - // of parsing or only in debug mode. if ( !Global::GetVR()->IsValidVR(vr) ) return false; @@ -1324,7 +1516,6 @@ std::string Document::GetDocEntryValue(DocEntry *entry) #endif //GDCM_NO_ANSI_STRING_STREAM return s.str(); } - return ((ValEntry *)entry)->GetValue(); } @@ -1395,7 +1586,7 @@ void Document::SkipDocEntry(DocEntry *entry) } /** - * \brief Skips to the begining of the next Header Entry + * \brief Skips to the beginning of the next Header Entry * \warning NOT end user intended method ! * @param currentDocEntry entry to skip */ @@ -1474,8 +1665,7 @@ void Document::FixDocEntryFoundLength(DocEntry *entry, { foundLength = 0; } - } - + } entry->SetLength(foundLength); } @@ -1487,10 +1677,10 @@ void Document::FixDocEntryFoundLength(DocEntry *entry, */ bool Document::IsDocEntryAnInteger(DocEntry *entry) { - uint16_t elem = entry->GetElement(); - uint16_t group = entry->GetGroup(); + uint16_t elem = entry->GetElement(); + uint16_t group = entry->GetGroup(); const std::string &vr = entry->GetVR(); - uint32_t length = entry->GetLength(); + uint32_t length = entry->GetLength(); // When we have some semantics on the element we just read, and if we // a priori know we are dealing with an integer, then we shall be @@ -1526,142 +1716,14 @@ bool Document::IsDocEntryAnInteger(DocEntry *entry) } /** - * \brief Find the Length till the next sequence delimiter - * \warning NOT end user intended method ! - * @return + * \brief Discover what the swap code is (among little endian, big endian, + * bad little endian, bad big endian). + * sw is set + * @return false when we are absolutely sure + * it's neither ACR-NEMA nor DICOM + * true when we hope ours assuptions are OK */ - -uint32_t Document::FindDocEntryLengthOBOrOW() - throw( FormatUnexpected ) -{ - // See PS 3.5-2001, section A.4 p. 49 on encapsulation of encoded pixel data. - long positionOnEntry = Fp->tellg(); - bool foundSequenceDelimiter = false; - uint32_t totalLength = 0; - - while ( !foundSequenceDelimiter ) - { - uint16_t group; - uint16_t elem; - try - { - group = ReadInt16(); - elem = ReadInt16(); - } - catch ( FormatError ) - { - throw FormatError("Unexpected end of file encountered during ", - "Document::FindDocEntryLengthOBOrOW()"); - } - - // We have to decount the group and element we just read - totalLength += 4; - - if ( group != 0xfffe || ( ( elem != 0xe0dd ) && ( elem != 0xe000 ) ) ) - { - long filePosition = Fp->tellg(); - gdcmVerboseMacro( "Neither an Item tag nor a Sequence delimiter tag on :" - << std::hex << group << " , " << elem - << ") -before- position x(" << filePosition << ")" ); - - Fp->seekg(positionOnEntry, std::ios::beg); - throw FormatUnexpected( "Neither an Item tag nor a Sequence delimiter tag."); - } - - if ( elem == 0xe0dd ) - { - foundSequenceDelimiter = true; - } - - uint32_t itemLength = ReadInt32(); - // We add 4 bytes since we just read the ItemLength with ReadInt32 - totalLength += itemLength + 4; - SkipBytes(itemLength); - - if ( foundSequenceDelimiter ) - { - break; - } - } - Fp->seekg( positionOnEntry, std::ios::beg); - return totalLength; -} - -/** - * \brief Reads a supposed to be 16 Bits integer - * (swaps it depending on processor endianity) - * @return read value - */ -uint16_t Document::ReadInt16() - throw( FormatError ) -{ - uint16_t g; - Fp->read ((char*)&g, (size_t)2); - if ( Fp->fail() ) - { - throw FormatError( "Document::ReadInt16()", " file error." ); - } - if( Fp->eof() ) - { - throw FormatError( "Document::ReadInt16()", "EOF." ); - } - g = SwapShort(g); - return g; -} - -/** - * \brief Reads a supposed to be 32 Bits integer - * (swaps it depending on processor endianity) - * @return read value - */ -uint32_t Document::ReadInt32() - throw( FormatError ) -{ - uint32_t g; - Fp->read ((char*)&g, (size_t)4); - if ( Fp->fail() ) - { - throw FormatError( "Document::ReadInt32()", " file error." ); - } - if( Fp->eof() ) - { - throw FormatError( "Document::ReadInt32()", "EOF." ); - } - g = SwapLong(g); - return g; -} - -/** - * \brief skips bytes inside the source file - * \warning NOT end user intended method ! - * @return - */ -void Document::SkipBytes(uint32_t nBytes) -{ - //FIXME don't dump the returned value - Fp->seekg((long)nBytes, std::ios::cur); -} - -/** - * \brief Loads all the needed Dictionaries - * \warning NOT end user intended method ! - */ -void Document::Initialize() -{ - RefPubDict = Global::GetDicts()->GetDefaultPubDict(); - RefShaDict = NULL; - Filetype = Unknown; -} - -/** - * \brief Discover what the swap code is (among little endian, big endian, - * bad little endian, bad big endian). - * sw is set - * @return false when we are absolutely sure - * it's neither ACR-NEMA nor DICOM - * true when we hope ours assuptions are OK - */ -bool Document::CheckSwap() +bool Document::CheckSwap() { // The only guaranted way of finding the swap code is to find a // group tag since we know it's length has to be of four bytes i.e. @@ -1879,7 +1941,6 @@ void Document::SetMaxSizeLoadEntry(long newSize) MaxSizeLoadEntry = newSize; } - /** * \brief Header Elements too long will not be printed * \todo See comments of \ref Document::MAX_SIZE_PRINT_ELEMENT_VALUE @@ -1900,104 +1961,6 @@ void Document::SetMaxSizePrintEntry(long newSize) } - -/** - * \brief Handle broken private tag from Philips NTSCAN - * where the endianess is being switch to BigEndian for no - * apparent reason - * @return no return - */ -void Document::HandleBrokenEndian(uint16_t &group, uint16_t &elem) -{ - // Endian reversion. Some files contain groups of tags with reversed endianess. - static int reversedEndian = 0; - // try to fix endian switching in the middle of headers - if ((group == 0xfeff) && (elem == 0x00e0)) - { - // start endian swap mark for group found - reversedEndian++; - SwitchByteSwapCode(); - // fix the tag - group = 0xfffe; - elem = 0xe000; - } - else if (group == 0xfffe && elem == 0xe00d && reversedEndian) - { - // end of reversed endian group - reversedEndian--; - SwitchByteSwapCode(); - } -} - -/** - * \brief Accesses the info from 0002,0010 : Transfer Syntax and TS - * @return The full Transfer Syntax Name (as opposed to Transfer Syntax UID) - */ -std::string Document::GetTransferSyntaxName() -{ - // use the TS (TS : Transfer Syntax) - std::string transferSyntax = GetEntryValue(0x0002,0x0010); - - if ( (transferSyntax.find(GDCM_NOTLOADED) < transferSyntax.length()) ) - { - gdcmErrorMacro( "Transfer Syntax not loaded. " << std::endl - << "Better you increase MAX_SIZE_LOAD_ELEMENT_VALUE" ); - return "Uncompressed ACR-NEMA"; - } - if ( transferSyntax == GDCM_UNFOUND ) - { - gdcmVerboseMacro( "Unfound Transfer Syntax (0002,0010)"); - return "Uncompressed ACR-NEMA"; - } - - // we do it only when we need it - const TSKey &tsName = Global::GetTS()->GetValue( transferSyntax ); - - // Global::GetTS() is a global static you shall never try to delete it! - return tsName; -} - -/** - * \brief Group 0002 is always coded Little Endian - * whatever Transfer Syntax is - * @return no return - */ -void Document::HandleOutOfGroup0002(uint16_t &group, uint16_t &elem) -{ - // Endian reversion. Some files contain groups of tags with reversed endianess. - if ( !Group0002Parsed && group != 0x0002) - { - Group0002Parsed = true; - // we just came out of group 0002 - // if Transfer syntax is Big Endian we have to change CheckSwap - - std::string ts = GetTransferSyntax(); - if ( !Global::GetTS()->IsTransferSyntax(ts) ) - { - gdcmVerboseMacro("True DICOM File, with NO Tansfer Syntax: " << ts ); - return; - } - - // Group 0002 is always 'Explicit ...' enven when Transfer Syntax says 'Implicit ..." - - if ( Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ImplicitVRLittleEndian ) - { - Filetype = ImplicitVR; - } - - // FIXME Strangely, this works with - //'Implicit VR Transfer Syntax (GE Private) - if ( Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ExplicitVRBigEndian ) - { - gdcmVerboseMacro("Transfer Syntax Name = [" - << GetTransferSyntaxName() << "]" ); - SwitchByteSwapCode(); - group = SwapShort(group); - elem = SwapShort(elem); - } - } -} - /** * \brief Read the next tag but WITHOUT loading it's value * (read the 'Group Number', the 'Element Number', @@ -2019,14 +1982,13 @@ DocEntry *Document::ReadNextDocEntry() { // We reached the EOF (or an error occured) therefore // header parsing has to be considered as finished. - //std::cout << e; return 0; } // Sometimes file contains groups of tags with reversed endianess. HandleBrokenEndian(group, elem); -// In 'true DICOM' files Group 0002 is always little endian + // In 'true DICOM' files Group 0002 is always little endian if ( HasDCMPreamble ) HandleOutOfGroup0002(group, elem); @@ -2082,222 +2044,77 @@ DocEntry *Document::ReadNextDocEntry() return newEntry; } -//GenerateFreeTagKeyInGroup? What was it designed for ?!? -/** - * \brief Generate a free TagKey i.e. a TagKey that is not present - * in the TagHt dictionary. - * @param group The generated tag must belong to this group. - * @return The element of tag with given group which is fee. - */ -//uint32_t Document::GenerateFreeTagKeyInGroup(uint16_t group) -//{ -// for (uint32_t elem = 0; elem < UINT32_MAX; elem++) -// { -// TagKey key = DictEntry::TranslateToKey(group, elem); -// if (TagHT.count(key) == 0) -// { -// return elem; -// } -// } -// return UINT32_MAX; -//} - /** - * \brief Compares two documents, according to \ref DicomDir rules - * \warning Does NOT work with ACR-NEMA files - * \todo Find a trick to solve the pb (use RET fields ?) - * @param document - * @return true if 'smaller' + * \brief Handle broken private tag from Philips NTSCAN + * where the endianess is being switch to BigEndian for no + * apparent reason + * @return no return */ -bool Document::operator<(Document &document) +void Document::HandleBrokenEndian(uint16_t &group, uint16_t &elem) { - // Patient Name - std::string s1 = GetEntryValue(0x0010,0x0010); - std::string s2 = document.GetEntryValue(0x0010,0x0010); - if(s1 < s2) - { - return true; - } - else if( s1 > s2 ) + // Endian reversion. Some files contain groups of tags with reversed endianess. + static int reversedEndian = 0; + // try to fix endian switching in the middle of headers + if ((group == 0xfeff) && (elem == 0x00e0)) { - return false; - } - else + // start endian swap mark for group found + reversedEndian++; + SwitchByteSwapCode(); + // fix the tag + group = 0xfffe; + elem = 0xe000; + } + else if (group == 0xfffe && elem == 0xe00d && reversedEndian) { - // Patient ID - s1 = GetEntryValue(0x0010,0x0020); - s2 = document.GetEntryValue(0x0010,0x0020); - if ( s1 < s2 ) - { - return true; - } - else if ( s1 > s2 ) - { - return false; - } - else - { - // Study Instance UID - s1 = GetEntryValue(0x0020,0x000d); - s2 = document.GetEntryValue(0x0020,0x000d); - if ( s1 < s2 ) - { - return true; - } - else if( s1 > s2 ) - { - return false; - } - else - { - // Serie Instance UID - s1 = GetEntryValue(0x0020,0x000e); - s2 = document.GetEntryValue(0x0020,0x000e); - if ( s1 < s2 ) - { - return true; - } - else if( s1 > s2 ) - { - return false; - } - } - } + // end of reversed endian group + reversedEndian--; + SwitchByteSwapCode(); } - return false; } /** - * \brief Re-computes the length of a ACR-NEMA/Dicom group from a DcmHeader - * @param filetype Type of the File to be written + * \brief Group 0002 is always coded Little Endian + * whatever Transfer Syntax is + * @return no return */ -int Document::ComputeGroup0002Length( FileType filetype ) +void Document::HandleOutOfGroup0002(uint16_t &group, uint16_t &elem) { - uint16_t gr; - std::string vr; - - int groupLength = 0; - bool found0002 = false; - - // for each zero-level Tag in the DCM Header - DocEntry *entry = GetFirstEntry(); - while(entry) + // Endian reversion. Some files contain groups of tags with reversed endianess. + if ( !Group0002Parsed && group != 0x0002) { - gr = entry->GetGroup(); + Group0002Parsed = true; + // we just came out of group 0002 + // if Transfer syntax is Big Endian we have to change CheckSwap - if (gr == 0x0002) + std::string ts = GetTransferSyntax(); + if ( !Global::GetTS()->IsTransferSyntax(ts) ) { - found0002 = true; - vr = entry->GetVR(); - - if (filetype == ExplicitVR) - { - if ( (vr == "OB") || (vr == "OW") || (vr == "SQ") ) - { - groupLength += 4; // explicit VR AND OB, OW, SQ : 4 more bytes - } - } - groupLength += 2 + 2 + 4 + entry->GetLength(); + gdcmVerboseMacro("True DICOM File, with NO Tansfer Syntax: " << ts ); + return; } - else if (found0002 ) - break; - entry = GetNextEntry(); - } - return groupLength; -} + // Group 0002 is always 'Explicit ...' enven when Transfer Syntax says 'Implicit ..." -/* - * \brief Walk recursively the given \ref DocEntrySet, and feed - * the given hash table (\ref TagDocEntryHT) with all the - * \ref DocEntry (Dicom entries) encountered. - * This method does the job for \ref BuildFlatHashTable. - * @param builtHT Where to collect all the \ref DocEntry encountered - * when recursively walking the given set. - * @param set The structure to be traversed (recursively). - */ -/*void Document::BuildFlatHashTableRecurse( TagDocEntryHT &builtHT, - DocEntrySet *set ) -{ - if (ElementSet *elementSet = dynamic_cast< ElementSet* > ( set ) ) - { - TagDocEntryHT const ¤tHT = elementSet->GetTagHT(); - for( TagDocEntryHT::const_iterator i = currentHT.begin(); - i != currentHT.end(); - ++i) - { - DocEntry *entry = i->second; - if ( SeqEntry *seqEntry = dynamic_cast(entry) ) + if ( Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ImplicitVRLittleEndian ) { - const ListSQItem& items = seqEntry->GetSQItems(); - for( ListSQItem::const_iterator item = items.begin(); - item != items.end(); - ++item) - { - BuildFlatHashTableRecurse( builtHT, *item ); - } - continue; + Filetype = ImplicitVR; } - builtHT[entry->GetKey()] = entry; - } - return; - } - - if (SQItem *SQItemSet = dynamic_cast< SQItem* > ( set ) ) - { - const ListDocEntry& currentList = SQItemSet->GetDocEntries(); - for (ListDocEntry::const_iterator i = currentList.begin(); - i != currentList.end(); - ++i) + + // FIXME Strangely, this works with + //'Implicit VR Transfer Syntax (GE Private) + if ( Global::GetTS()->GetSpecialTransferSyntax(ts) == TS::ExplicitVRBigEndian ) { - DocEntry *entry = *i; - if ( SeqEntry *seqEntry = dynamic_cast(entry) ) - { - const ListSQItem& items = seqEntry->GetSQItems(); - for( ListSQItem::const_iterator item = items.begin(); - item != items.end(); - ++item) - { - BuildFlatHashTableRecurse( builtHT, *item ); - } - continue; - } - builtHT[entry->GetKey()] = entry; + gdcmVerboseMacro("Transfer Syntax Name = [" + << GetTransferSyntaxName() << "]" ); + SwitchByteSwapCode(); + group = SwapShort(group); + elem = SwapShort(elem); } - } -}*/ - -/* - * \brief Build a \ref TagDocEntryHT (i.e. a std::map<>) from the current - * Document. - * - * The structure used by a Document (through \ref ElementSet), - * in order to hold the parsed entries of a Dicom header, is a recursive - * one. This is due to the fact that the sequences (when present) - * can be nested. Additionaly, the sequence items (represented in - * gdcm as \ref SQItem) add an extra complexity to the data - * structure. Hence, a gdcm user whishing to visit all the entries of - * a Dicom header will need to dig in the gdcm internals (which - * implies exposing all the internal data structures to the API). - * In order to avoid this burden to the user, \ref BuildFlatHashTable - * recursively builds a temporary hash table, which holds all the - * Dicom entries in a flat structure (a \ref TagDocEntryHT i.e. a - * std::map<>). - * \warning Of course there is NO integrity constrain between the - * returned \ref TagDocEntryHT and the \ref ElementSet used - * to build it. Hence if the underlying \ref ElementSet is - * altered, then it is the caller responsability to invoke - * \ref BuildFlatHashTable again... - * @return The flat std::map<> we juste build. - */ -/*TagDocEntryHT *Document::BuildFlatHashTable() -{ - TagDocEntryHT *FlatHT = new TagDocEntryHT; - BuildFlatHashTableRecurse( *FlatHT, this ); - return FlatHT; -}*/ +} -} // end namespace gdcm +//----------------------------------------------------------------------------- +// Print //----------------------------------------------------------------------------- +} // end namespace gdcm