PDFColorSpace PDFAnalyzer::getCSType(PdfObject* cs)
{
try {
// colorspace is either a name or an array
if (cs && cs->IsName())
{
PdfName csName = cs->GetName();
if (csName == "DeviceGray")
return CS_DeviceGray;
else if (csName == "DeviceRGB")
return CS_DeviceRGB;
else if (csName == "DeviceCMYK")
return CS_DeviceCMYK;
}
else if (cs && cs->IsArray())
{
PdfArray csArr = cs->GetArray();
PdfObject csTypePdfName = csArr[0];
if (csTypePdfName.IsName())
{
PdfName csTypeName = csTypePdfName.GetName();
if (csTypeName == "ICCBased")
return CS_ICCBased;
else if (csTypeName == "CalGray")
return CS_CalGray;
else if (csTypeName == "CalRGB")
return CS_CalRGB;
else if (csTypeName == "Lab")
return CS_Lab;
else if (csTypeName == "Indexed")
{
PdfObject base = cs->GetArray()[1];
PdfObject* pBase = &base;
if (base.IsReference())
{
pBase = cs->GetOwner()->GetObject(base.GetReference());
}
pBase->SetOwner(cs->GetOwner());
return getCSType(pBase);
}
else if (csTypeName == "Separation")
return CS_Separation;
else if (csTypeName == "DeviceN")
return CS_DeviceN;
else if (csTypeName == "Pattern")
return CS_Pattern;
}
}
}
catch (PdfError & e)
{
qDebug() << "Error in identifying the color type";
e.PrintErrorMsg();
return CS_Unknown;
}
return CS_Unknown;
}
PDFFont PDFAnalyzer::getFontInfo(PdfObject* fontObj)
{
PDFFont currFont;
PdfObject* subtype = fontObj->GetIndirectKey("Subtype");
if (subtype && subtype->IsName())
{
PdfObject* fontDesc = fontObj->GetIndirectKey("FontDescriptor");
if (subtype->GetName() == "Type1")
currFont.fontType = F_Type1;
else if (subtype->GetName() == "MMType1")
currFont.fontType = F_MMType1;
else if (subtype->GetName() == "TrueType")
currFont.fontType = F_TrueType;
else if (subtype->GetName() == "Type3")
{
currFont.fontType = F_Type3;
currFont.isEmbedded = true;
fontDesc = NULL;
}
else if (subtype->GetName() == "Type0")
{
PdfObject* descendantFonts = fontObj->GetIndirectKey("DescendantFonts");
if (descendantFonts && descendantFonts->IsArray())
{
PdfObject descendantFont = descendantFonts->GetArray()[0];
descendantFont.SetOwner(descendantFonts->GetOwner());
PdfObject* subtypeDescFont = descendantFont.GetIndirectKey("Subtype");
fontDesc = descendantFont.MustGetIndirectKey("FontDescriptor");
if (subtypeDescFont && subtypeDescFont->IsName())
{
if (subtypeDescFont->GetName() == "CIDFontType0")
currFont.fontType = F_CIDFontType0;
else if (subtypeDescFont->GetName() == "CIDFontType2")
currFont.fontType = F_CIDFontType2;
}
}
}
if (fontDesc)
{
PdfObject* fontFile = fontDesc->GetIndirectKey("FontFile");
PdfObject* fontFile2 = fontDesc->GetIndirectKey("FontFile2");
PdfObject* fontFile3 = fontDesc->GetIndirectKey("FontFile3");
if (fontFile && fontFile->HasStream())
currFont.isEmbedded = true;
if (fontFile2 && fontFile2->HasStream())
currFont.isEmbedded = true;
if (fontFile3 && fontFile3->HasStream())
{
currFont.isEmbedded = true;
PdfObject* ff3Subtype = fontFile3->GetIndirectKey("Subtype");
if (ff3Subtype && ff3Subtype->IsName() && ff3Subtype->GetName() == "OpenType")
currFont.isOpenType = true;
}
}
}
return currFont;
}
TVecFilters PdfFilterFactory::CreateFilterList( const PdfObject* pObject )
{
TVecFilters filters;
const PdfObject* pObj = NULL;
if( pObject->IsDictionary() && pObject->GetDictionary().HasKey( "Filter" ) )
pObj = pObject->GetDictionary().GetKey( "Filter" );
else if( pObject->IsArray() )
pObj = pObject;
else if( pObject->IsName() )
pObj = pObject;
if (!pObj)
// Object had no /Filter key . Return a null filter list.
return filters;
if( pObj->IsName() )
filters.push_back( PdfFilterFactory::FilterNameToType( pObj->GetName() ) );
else if( pObj->IsArray() )
{
TCIVariantList it = pObj->GetArray().begin();
while( it != pObj->GetArray().end() )
{
if ( (*it).IsName() )
{
filters.push_back( PdfFilterFactory::FilterNameToType( (*it).GetName() ) );
}
else if ( (*it).IsReference() )
{
PdfObject* pFilter = pObject->GetOwner()->GetObject( (*it).GetReference() );
if( pFilter == NULL )
{
PODOFO_RAISE_ERROR_INFO( ePdfError_InvalidDataType, "Filter array contained unexpected reference" );
}
filters.push_back( PdfFilterFactory::FilterNameToType( pFilter->GetName() ) );
}
else
{
PODOFO_RAISE_ERROR_INFO( ePdfError_InvalidDataType, "Filter array contained unexpected non-name type" );
}
++it;
}
}
return filters;
}
void PDFProcessor::start ()
{
int nNum = 0;
try
{
PdfObject* pObj = NULL;
// open document
qDebug() << "Opening file: " << filename.toStdString().c_str();
PdfMemDocument document( filename.toStdString().c_str() );
// m_pszOutputDirectory = const_cast<char*>(pszOutput);
TCIVecObjects it = document.GetObjects().begin();
while( it != document.GetObjects().end() )
{
if( (*it)->IsDictionary() )
{
PdfObject* pObjType = (*it)->GetDictionary().GetKey( PdfName::KeyType );
PdfObject* pObjSubType = (*it)->GetDictionary().GetKey( PdfName::KeySubtype );
if( ( pObjType && pObjType->IsName() && ( pObjType->GetName().GetName() == "XObject" ) ) ||
( pObjSubType && pObjSubType->IsName() && ( pObjSubType->GetName().GetName() == "Image" ) ) )
{
pObj = (*it)->GetDictionary().GetKey( PdfName::KeyFilter );
if( pObj && pObj->IsArray() && pObj->GetArray().GetSize() == 1 &&
pObj->GetArray()[0].IsName() && (pObj->GetArray()[0].GetName().GetName() == "DCTDecode") )
pObj = &pObj->GetArray()[0];
std::string filterName = pObj->GetName().GetName();
bool processed = 0;
if( pObj && pObj->IsName() && ( filterName == "DCTDecode" ) )
{
// The only filter is JPEG -> create a JPEG file
qDebug() << "JPG found.\n";
processed = true;
nNum++;
}
if( pObj && pObj->IsName() && ( filterName == "JPXDecode" ) )
{
// The only filter is JPEG -> create a JPEG file
qDebug() << "JPG found.\n";
processed = true;
nNum++;
}
if( pObj && pObj->IsName() && ( filterName == "FlateDecode" ) )
{
// The only filter is JPEG -> create a JPEG file
qDebug() << "JPG found.\n";
processed = true;
nNum++;
}
// else we found something strange, we do not care about it for now.
if (processed == false)
{
qDebug() << "Unknown image type found:" << QString::fromStdString(filterName) << "\n";
nNum++;
}
document.FreeObjectMemory( *it );
}
}
++it;
}
}
catch( PdfError & e )
{
qDebug() << "Error: An error ocurred during processing the pdf file:" << e.GetError();
e.PrintErrorMsg();
return;// e.GetError();
}
// TODO: statistics of no of images etc
// nNum = extractor.GetNumImagesExtracted();
qDebug() << "Extracted " << nNum << " images successfully from the PDF file.\n";
}
PdfFontMetricsObject::PdfFontMetricsObject( PdfObject* pFont, PdfObject* pDescriptor, const PdfEncoding* const pEncoding )
: PdfFontMetrics( ePdfFontType_Unknown, "", NULL ),
m_pEncoding( pEncoding ), m_dDefWidth(0.0)
{
if( !pDescriptor )
{
PODOFO_RAISE_ERROR( ePdfError_InvalidHandle );
}
const PdfName & rSubType = pFont->GetDictionary().GetKey( PdfName::KeySubtype )->GetName();
// OC 15.08.2010 BugFix: /FirstChar /LastChar /Widths are in the Font dictionary and not in the FontDescriptor
if ( rSubType == PdfName("Type1") || rSubType == PdfName("TrueType") ) {
m_sName = pDescriptor->GetIndirectKey( "FontName" )->GetName();
m_bbox = pDescriptor->GetIndirectKey( "FontBBox" )->GetArray();
m_nFirst = static_cast<int>(pFont->GetDictionary().GetKeyAsLong( "FirstChar", 0L ));
m_nLast = static_cast<int>(pFont->GetDictionary().GetKeyAsLong( "LastChar", 0L ));
// OC 15.08.2010 BugFix: GetIndirectKey() instead of GetDictionary().GetKey() and "Widths" instead of "Width"
PdfObject* widths = pFont->GetIndirectKey( "Widths" );
if( widths != NULL )
{
m_width = widths->GetArray();
m_missingWidth = NULL;
}
else
{
widths = pDescriptor->GetDictionary().GetKey( "MissingWidth" );
if( widths == NULL )
{
PODOFO_RAISE_ERROR_INFO( ePdfError_NoObject, "Font object defines neither Widths, nor MissingWidth values!" );
m_missingWidth = widths;
}
}
} else if ( rSubType == PdfName("CIDFontType0") || rSubType == PdfName("CIDFontType2") ) {
PdfObject *pObj = pDescriptor->GetIndirectKey( "FontName" );
if (pObj) {
m_sName = pObj->GetName();
}
pObj = pDescriptor->GetIndirectKey( "FontBBox" );
if (pObj) {
m_bbox = pObj->GetArray();
}
m_nFirst = 0;
m_nLast = 0;
m_dDefWidth = static_cast<double>(pFont->GetDictionary().GetKeyAsLong( "DW", 1000L ));
PdfVariant default_width(m_dDefWidth);
PdfObject * pw = pFont->GetIndirectKey( "W" );
for (int i = m_nFirst; i <= m_nLast; ++i) {
m_width.push_back(default_width);
}
if (pw) {
PdfArray w = pw->GetArray();
int pos = 0;
while (pos < static_cast<int>(w.GetSize())) {
int start = static_cast<int>(w[pos++].GetNumber());
PODOFO_ASSERT (start >= 0);
if (w[pos].IsArray()) {
PdfArray widths = w[pos++].GetArray();
int length = start + static_cast<int>(widths.GetSize());
PODOFO_ASSERT (length >= start);
if (length > static_cast<int>(m_width.GetSize())) {
m_width.resize(length, default_width);
}
for (int i = 0; i < static_cast<int>(widths.GetSize()); ++i) {
m_width[start + i] = widths[i];
}
} else {
int end = static_cast<int>(w[pos++].GetNumber());
int length = start + end;
PODOFO_ASSERT (length >= start);
if (length > static_cast<int>(m_width.GetSize())) {
m_width.resize(length, default_width);
}
pdf_int64 width = w[pos++].GetNumber();
for (int i = start; i <= end; ++i)
m_width[i] = PdfVariant(width);
}
}
}
m_nLast = m_width.GetSize() - 1;
} else {
PODOFO_RAISE_ERROR_INFO( ePdfError_UnsupportedFontFormat, rSubType.GetEscapedName().c_str() );
}
m_nWeight = static_cast<unsigned int>(pDescriptor->GetDictionary().GetKeyAsLong( "FontWeight", 400L ));
m_nItalicAngle = static_cast<int>(pDescriptor->GetDictionary().GetKeyAsLong( "ItalicAngle", 0L ));
m_dPdfAscent = pDescriptor->GetDictionary().GetKeyAsReal( "Ascent", 0.0 );
m_dAscent = m_dPdfAscent / 1000.0;
m_dPdfDescent = pDescriptor->GetDictionary().GetKeyAsReal( "Descent", 0.0 );
m_dDescent = m_dPdfDescent / 1000.0;
m_dLineSpacing = m_dAscent + m_dDescent;
// Try to fine some sensible values
m_dUnderlineThickness = 1.0;
m_dUnderlinePosition = 0.0;
m_dStrikeOutThickness = m_dUnderlinePosition;
m_dStrikeOutPosition = m_dAscent / 2.0;
m_bSymbol = false; // TODO
}
PdfFont* PdfFontFactory::CreateFont( FT_Library*, PdfObject* pObject )
{
PdfFontMetrics* pMetrics = NULL;
PdfFont* pFont = NULL;
PdfObject* pDescriptor = NULL;
PdfObject* pEncoding = NULL;
if( pObject->GetDictionary().GetKey( PdfName::KeyType )->GetName() != PdfName("Font") )
{
PODOFO_RAISE_ERROR( ePdfError_InvalidDataType );
}
const PdfName & rSubType = pObject->GetDictionary().GetKey( PdfName::KeySubtype )->GetName();
if( rSubType == PdfName("Type0") )
{
// The PDF reference states that DescendantFonts must be an array,
// some applications (e.g. MS Word) put the array into an indirect object though.
const PdfArray & descendant =
pObject->GetIndirectKey( "DescendantFonts" )->GetArray();
PdfObject* pFontObject = pObject->GetOwner()->GetObject( descendant[0].GetReference() );
pDescriptor = pFontObject->GetIndirectKey( "FontDescriptor" );
pEncoding = pObject->GetIndirectKey( "Encoding" );
if ( pEncoding && pDescriptor ) // OC 18.08.2010: Avoid sigsegv
{
const PdfEncoding* const pPdfEncoding =
PdfEncodingObjectFactory::CreateEncoding( pEncoding );
// OC 15.08.2010 BugFix: Parameter pFontObject added: TODO: untested
pMetrics = new PdfFontMetricsObject( pFontObject, pDescriptor, pPdfEncoding );
pFont = new PdfFontCID( pMetrics, pPdfEncoding, pObject, false );
}
}
else if( rSubType == PdfName("Type1") )
{
// TODO: Old documents do not have a FontDescriptor for
// the 14 standard fonts. This suggestions is
// deprecated now, but give us problems with old documents.
pDescriptor = pObject->GetIndirectKey( "FontDescriptor" );
pEncoding = pObject->GetIndirectKey( "Encoding" );
// OC 13.08.2010: Handle missing FontDescriptor for the 14 standard fonts:
if( !pDescriptor )
{
// Check if its a PdfFontType1Base14
PdfObject* pBaseFont = NULL;
pBaseFont = pObject->GetIndirectKey( "BaseFont" );
const char* pszBaseFontName = pBaseFont->GetName().GetName().c_str();
PdfFontMetricsBase14* pMetrics = PODOFO_Base14FontDef_FindBuiltinData(pszBaseFontName);
if ( pMetrics != NULL )
{
// pEncoding may be undefined, found a valid pdf with
// 20 0 obj
// <<
// /Type /Font
// /BaseFont /ZapfDingbats
// /Subtype /Type1
// >>
// endobj
// If pEncoding is null then
// use StandardEncoding for Courier, Times, Helvetica font families
// and special encodings for Symbol and ZapfDingbats
const PdfEncoding* pPdfEncoding = NULL;
if ( pEncoding!= NULL )
pPdfEncoding = PdfEncodingObjectFactory::CreateEncoding( pEncoding );
else if ( !pMetrics->IsSymbol() )
pPdfEncoding = PdfEncodingFactory::GlobalStandardEncodingInstance();
else if ( strcmp(pszBaseFontName, "Symbol") == 0 )
pPdfEncoding = PdfEncodingFactory::GlobalSymbolEncodingInstance();
else if ( strcmp(pszBaseFontName, "ZapfDingbats") == 0 )
pPdfEncoding = PdfEncodingFactory::GlobalZapfDingbatsEncodingInstance();
return new PdfFontType1Base14(pMetrics, pPdfEncoding, pObject);
}
}
const PdfEncoding* pPdfEncoding = NULL;
if ( pEncoding != NULL )
pPdfEncoding = PdfEncodingObjectFactory::CreateEncoding( pEncoding );
else if ( pDescriptor )
{
// OC 18.08.2010 TODO: Encoding has to be taken from the font's built-in encoding
// Its extremely complicated to interpret the type1 font programs
// so i try to determine if its a symbolic font by reading the FontDescriptor Flags
// Flags & 4 --> Symbolic, Flags & 32 --> Nonsymbolic
pdf_int32 lFlags = static_cast<pdf_int32>(pDescriptor->GetDictionary().GetKeyAsLong( "Flags", 0L ));
if ( lFlags & 32 ) // Nonsymbolic, otherwise pEncoding remains NULL
pPdfEncoding = PdfEncodingFactory::GlobalStandardEncodingInstance();
}
if ( pPdfEncoding && pDescriptor ) // OC 18.08.2010: Avoid sigsegv
{
// OC 15.08.2010 BugFix: Parameter pObject added:
pMetrics = new PdfFontMetricsObject( pObject, pDescriptor, pPdfEncoding );
pFont = new PdfFontType1( pMetrics, pPdfEncoding, pObject );
}
}
else if( rSubType == PdfName("TrueType") )
{
pDescriptor = pObject->GetIndirectKey( "FontDescriptor" );
pEncoding = pObject->GetIndirectKey( "Encoding" );
if ( pEncoding && pDescriptor ) // OC 18.08.2010: Avoid sigsegv
{
const PdfEncoding* const pPdfEncoding =
PdfEncodingObjectFactory::CreateEncoding( pEncoding );
// OC 15.08.2010 BugFix: Parameter pObject added:
pMetrics = new PdfFontMetricsObject( pObject, pDescriptor, pPdfEncoding );
pFont = new PdfFontTrueType( pMetrics, pPdfEncoding, pObject );
}
}
return pFont;
}
void PdfMemStream::FlateCompress()
{
PdfObject* pObj;
PdfVariant vFilter( PdfName("FlateDecode" ) );
PdfVariant vFilterList;
PdfArray tFilters;
PdfArray::const_iterator tciFilters;
if( !m_lLength )
return; // ePdfError_ErrOk
// TODO: Handle DecodeParms
if( m_pParent->GetDictionary().HasKey( "Filter" ) )
{
pObj = m_pParent->GetIndirectKey( "Filter" );
if( pObj->IsName() )
{
if( pObj->GetName() != "DCTDecode" && pObj->GetName() != "FlateDecode" )
{
tFilters.push_back( vFilter );
tFilters.push_back( *pObj );
}
}
else if( pObj->IsArray() )
{
tciFilters = pObj->GetArray().begin();
while( tciFilters != pObj->GetArray().end() )
{
if( (*tciFilters).IsName() )
{
// do not compress DCTDecoded are already FlateDecoded streams again
if( (*tciFilters).GetName() == "DCTDecode" || (*tciFilters).GetName() == "FlateDecode" )
{
return;
}
}
++tciFilters;
}
tFilters.push_back( vFilter );
tciFilters = pObj->GetArray().begin();
while( tciFilters != pObj->GetArray().end() )
{
tFilters.push_back( (*tciFilters) );
++tciFilters;
}
}
else
return;
vFilterList = PdfVariant( tFilters );
m_pParent->GetDictionary().AddKey( "Filter", vFilterList );
FlateCompressStreamData(); // throws an exception on error
}
else
{
m_pParent->GetDictionary().AddKey( "Filter", PdfName( "FlateDecode" ) );
FlateCompressStreamData();
}
}
void PDFAnalyzer::inspectExtGStateObj(PdfObject* extGStateObj, QList<PDFColorSpace> & usedColorSpaces, bool & hasTransparency, QList<PDFFont> & usedFonts, PDFGraphicState & currGS)
{
PdfObject* bmObj = extGStateObj->GetIndirectKey("BM");
if (bmObj && bmObj->IsName())
{
currGS.blendModes.clear();
currGS.blendModes.append(bmObj->GetName().GetEscapedName().c_str());
if (!(bmObj->GetName() == "Normal" || bmObj->GetName() == "Compatible"))
hasTransparency = true;
}
else if (bmObj && bmObj->IsArray())
{
PdfArray arr = bmObj->GetArray();
currGS.blendModes.clear();
for(int i=0; i<arr.GetSize(); ++i)
currGS.blendModes.append(arr[i].GetName().GetEscapedName().c_str());
if (arr[0].IsName() && !(arr[0].GetName() == "Normal" || arr[0].GetName() == "Compatible"))
hasTransparency = true;
}
PdfObject* caObj = extGStateObj->GetIndirectKey("ca");
if (caObj && (caObj->IsReal() || caObj->IsNumber()))
{
currGS.fillAlphaConstant = caObj->GetReal();
if (caObj->GetReal() < 1)
hasTransparency = true;
}
PdfObject* cAObj = extGStateObj->GetIndirectKey("CA");
if (cAObj && (cAObj->IsReal() || cAObj->IsNumber()))
{
if (cAObj->GetReal() < 1)
hasTransparency = true;
}
PdfObject* sMaskObj = extGStateObj->GetIndirectKey("SMask");
if (sMaskObj && !(sMaskObj->IsName() && sMaskObj->GetName() == "None"))
hasTransparency = true;
PdfObject* fontObj = extGStateObj->GetIndirectKey("Font");
if (fontObj && fontObj->IsArray())
{
PdfArray arr = fontObj->GetArray();
if (arr[0].IsReference())
{
PdfReference ref = arr[0].GetReference();
PdfObject* fontObject = m_doc->GetObjects().GetObject(ref);
if (fontObject)
{
PDFFont font = getFontInfo(fontObject);
usedFonts.append(font);
currGS.font.first = font;
currGS.font.second = arr[1].GetReal();
}
}
}
PdfObject* lwObj = extGStateObj->GetIndirectKey("LW");
if (lwObj)
currGS.lineWidth = lwObj->GetReal();
PdfObject* lcObj = extGStateObj->GetIndirectKey("LC");
if (lcObj)
currGS.lineCap = lcObj->GetNumber();
PdfObject* ljObj = extGStateObj->GetIndirectKey("LJ");
if (ljObj)
currGS.lineJoin = ljObj->GetNumber();
PdfObject* mlObj = extGStateObj->GetIndirectKey("ML");
if (mlObj)
currGS.miterLimit = mlObj->GetReal();
PdfObject* dObj = extGStateObj->GetIndirectKey("D");
if (dObj)
{
PdfArray dashArr = dObj->GetArray()[0];
currGS.dashPattern.first.clear();
for (int i=0; i<dashArr.GetSize(); ++i)
currGS.dashPattern.first.append(dashArr[i].GetNumber());
currGS.dashPattern.second = dObj->GetArray()[1].GetNumber();
}
}
bool PDFAnalyzer::inspectCanvas(PdfCanvas* canvas, QList<PDFColorSpace> & usedColorSpaces, bool & hasTransparency, QList<PDFFont> & usedFonts, QList<PDFImage> & imgs)
{
// this method can be used to get used color spaces, detect transparency, and get used fonts in either PdfPage or PdfXObject
PdfObject* colorSpaceRes;
PdfObject* xObjects;
PdfObject* transGroup;
PdfObject* extGState;
PdfObject* fontRes;
QMap<PdfName, PDFColorSpace> processedNamedCS;
QMap<PdfName, PDFFont> processedNamedFont;
QList<PdfName> processedNamedXObj;
QList<PdfName> processedNamedGS;
try {
// get hold of a PdfObject pointer of this canvas
// needed for the finding resources code below to work
PdfPage* page = dynamic_cast<PdfPage*>(canvas);
PdfObject* canvasObject = page?(page->GetObject()):((dynamic_cast<PdfXObject*>(canvas))->GetObject());
// find a resource with ColorSpace entry
PdfObject* resources = canvas->GetResources();
for (PdfObject* par = canvasObject; par && !resources; par = par->GetIndirectKey("Parent"))
{
resources = par->GetIndirectKey("Resources");
}
colorSpaceRes = resources?resources->GetIndirectKey("ColorSpace"):NULL;
xObjects = resources?resources->GetIndirectKey("XObject"):NULL;
extGState = resources?resources->GetIndirectKey("ExtGState"):NULL;
fontRes = resources?resources->GetIndirectKey("Font"):NULL;
// getting the transparency group of this content stream (if available)
transGroup = canvasObject?canvasObject->GetIndirectKey("Group"):NULL;
if (transGroup)
{
PdfObject* subtype = transGroup->GetIndirectKey("S");
if (subtype && subtype->GetName() == "Transparency")
{
// having transparency group means there's transparency in the PDF
hasTransparency = true;
// reporting the color space used in transparency group (Section 7.5.5, PDF 1.6 Spec)
PdfObject* cs = transGroup->GetIndirectKey("CS");
if (cs)
{
PDFColorSpace retval = getCSType(cs);
if (retval != CS_Unknown && !usedColorSpaces.contains(retval))
usedColorSpaces.append(retval);
}
}
}
}
catch (PdfError & e)
{
qDebug() << "Error in analyzing stream's resources.";
e.PrintErrorMsg();
return false;
}
try {
// start parsing the content stream
PdfContentsTokenizer tokenizer(canvas);
EPdfContentsType t;
const char * kwText;
PdfVariant var;
bool readToken;
int tokenNumber = 0;
QList<PdfVariant> args;
bool inlineImgDict = false;
QStack<PDFGraphicState> gsStack;
PDFGraphicState currGS;
while ((readToken = tokenizer.ReadNext(t, kwText, var)))
{
++tokenNumber;
if (t == ePdfContentsType_Variant)
{
args.append(var);
}
else if (t == ePdfContentsType_Keyword)
{
QString kw(kwText);
switch(kwNameMap.value(kw, KW_Undefined))
{
case KW_q:
gsStack.push(currGS);
break;
case KW_Q:
currGS = gsStack.pop();
break;
case KW_cm:
{
if (args.size() == 6)
{
double mt[6];
for (int i=0; i<6; ++i)
{
mt[i] = args[i].GetReal();
}
QMatrix transMatrix(mt[0], mt[1], mt[2], mt[3], mt[4], mt[5]);
currGS.ctm = transMatrix*currGS.ctm;
}
}
break;
case KW_w:
currGS.lineWidth = args[0].GetReal();
break;
case KW_J:
currGS.lineCap = args[0].GetNumber();
break;
case KW_j:
currGS.lineJoin = args[0].GetNumber();
break;
case KW_M:
currGS.lineJoin = args[0].GetReal();
break;
case KW_d:
{
currGS.dashPattern.first.clear();
PdfArray dashArr = args[0].GetArray();
for (int i=0; i<dashArr.size(); ++i)
currGS.dashPattern.first.append(dashArr[i].GetNumber());
currGS.dashPattern.second = args[0].GetNumber();
}
break;
case KW_g:
if (!usedColorSpaces.contains(CS_DeviceGray))
usedColorSpaces.append(CS_DeviceGray);
currGS.fillCS = CS_DeviceGray;
currGS.fillColor.clear();
currGS.fillColor.append(args[0].GetReal());
break;
case KW_G:
if (!usedColorSpaces.contains(CS_DeviceGray))
usedColorSpaces.append(CS_DeviceGray);
currGS.strokeCS = CS_DeviceGray;
currGS.strokeColor.clear();
currGS.strokeColor.append(args[0].GetReal());
break;
case KW_rg:
if (!usedColorSpaces.contains(CS_DeviceRGB))
usedColorSpaces.append(CS_DeviceRGB);
currGS.fillCS = CS_DeviceRGB;
currGS.fillColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.fillColor.append(args[i].GetReal());
break;
case KW_RG:
if (!usedColorSpaces.contains(CS_DeviceRGB))
usedColorSpaces.append(CS_DeviceRGB);
currGS.strokeCS = CS_DeviceRGB;
currGS.strokeColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.strokeColor.append(args[i].GetReal());
break;
case KW_k:
if (!usedColorSpaces.contains(CS_DeviceCMYK))
usedColorSpaces.append(CS_DeviceCMYK);
currGS.fillCS = CS_DeviceCMYK;
currGS.fillColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.fillColor.append(args[i].GetReal());
break;
case KW_K:
if (!usedColorSpaces.contains(CS_DeviceCMYK))
usedColorSpaces.append(CS_DeviceCMYK);
currGS.strokeCS = CS_DeviceCMYK;
currGS.strokeColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.strokeColor.append(args[i].GetReal());
break;
case KW_cs:
{
if (args.size() == 1 && args[0].IsName())
{
if (args[0].GetName() == "DeviceGray")
{
currGS.fillCS = CS_DeviceGray;
currGS.fillColor.clear();
currGS.fillColor.append(0);
if (!usedColorSpaces.contains(CS_DeviceGray))
usedColorSpaces.append(CS_DeviceGray);
}
else if (args[0].GetName() == "DeviceRGB")
{
currGS.fillCS = CS_DeviceRGB;
currGS.fillColor.clear();
for (int i=0; i<3; ++i)
currGS.fillColor.append(0);
if (!usedColorSpaces.contains(CS_DeviceRGB))
usedColorSpaces.append(CS_DeviceRGB);
}
else if (args[0].GetName() == "DeviceCMYK")
{
currGS.fillCS = CS_DeviceCMYK;
currGS.fillColor.clear();
for (int i=0; i<3; ++i)
currGS.fillColor.append(0);
currGS.fillColor.append(1);
if (!usedColorSpaces.contains(CS_DeviceCMYK))
usedColorSpaces.append(CS_DeviceCMYK);
}
else if (args[0].GetName() == "Pattern")
{
currGS.fillCS = CS_Pattern;
if (!usedColorSpaces.contains(CS_Pattern))
usedColorSpaces.append(CS_Pattern);
}
else
{
if (processedNamedCS.contains(args[0].GetName()))
{
currGS.fillCS = processedNamedCS.value(args[0].GetName());
}
else
{
if (colorSpaceRes && colorSpaceRes->GetIndirectKey(args[0].GetName()))
{
PdfObject* csEntry = colorSpaceRes->GetIndirectKey(args[0].GetName());
PDFColorSpace retval = getCSType(csEntry);
if (retval != CS_Unknown && !usedColorSpaces.contains(retval))
usedColorSpaces.append(retval);
currGS.fillCS = retval;
processedNamedCS.insert(args[0].GetName(), retval);
}
else
{
qDebug() << "Supplied colorspace is undefined!";
return false;
}
}
}
}
else
{
qDebug() << "Wrong syntax in specifying color space!";
return false;
}
}
break;
case KW_CS:
{
if (args.size() == 1 && args[0].IsName())
{
if (args[0].GetName() == "DeviceGray")
{
currGS.strokeCS = CS_DeviceGray;
currGS.strokeColor.clear();
currGS.strokeColor.append(0);
if (!usedColorSpaces.contains(CS_DeviceGray))
usedColorSpaces.append(CS_DeviceGray);
}
else if (args[0].GetName() == "DeviceRGB")
{
currGS.fillCS = CS_DeviceRGB;
currGS.strokeColor.clear();
for (int i=0; i<3; ++i)
currGS.strokeColor.append(0);
if (!usedColorSpaces.contains(CS_DeviceRGB))
usedColorSpaces.append(CS_DeviceRGB);
}
else if (args[0].GetName() == "DeviceCMYK")
{
currGS.fillCS = CS_DeviceCMYK;
currGS.strokeColor.clear();
for (int i=0; i<3; ++i)
currGS.strokeColor.append(0);
currGS.strokeColor.append(1);
if (!usedColorSpaces.contains(CS_DeviceCMYK))
usedColorSpaces.append(CS_DeviceCMYK);
}
else if (args[0].GetName() == "Pattern")
{
currGS.fillCS = CS_Pattern;
if (!usedColorSpaces.contains(CS_Pattern))
usedColorSpaces.append(CS_Pattern);
}
else
{
if (processedNamedCS.contains(args[0].GetName()))
{
currGS.strokeCS = processedNamedCS.value(args[0].GetName());
}
else
{
if (colorSpaceRes && colorSpaceRes->GetIndirectKey(args[0].GetName()))
{
PdfObject* csEntry = colorSpaceRes->GetIndirectKey(args[0].GetName());
PDFColorSpace retval = getCSType(csEntry);
if (retval != CS_Unknown && !usedColorSpaces.contains(retval))
usedColorSpaces.append(retval);
currGS.strokeCS = retval;
processedNamedCS.insert(args[0].GetName(), retval);
}
else
{
qDebug() << "Supplied colorspace is undefined!";
return false;
}
}
}
}
else
{
qDebug() << "Wrong syntax in specifying color space!";
return false;
}
}
break;
case KW_sc:
currGS.fillColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.fillColor.append(args[i].GetReal());
break;
case KW_SC:
currGS.strokeColor.clear();
for (int i=0; i<args.size(); ++i)
currGS.strokeColor.append(args[i].GetReal());
break;
case KW_scn:
currGS.fillColor.clear();
for (int i=0; i<args.size(); ++i)
{
if (args[i].IsReal() || args[i].IsNumber())
currGS.fillColor.append(args[i].GetReal());
}
break;
case KW_SCN:
currGS.strokeColor.clear();
for (int i=0; i<args.size(); ++i)
{
if (args[i].IsReal() || args[i].IsNumber())
currGS.strokeColor.append(args[i].GetReal());
}
break;
case KW_Do: // image or form XObject
{
if (!processedNamedXObj.contains(args[0].GetName()))
{
if (args.size() == 1 && args[0].IsName() && xObjects)
{
PdfObject* xObject = xObjects->GetIndirectKey(args[0].GetName());
PdfObject* subtypeObject = xObject?xObject->GetIndirectKey("Subtype"):NULL;
if (subtypeObject && subtypeObject->IsName())
{
if (subtypeObject->GetName() == "Image")
{
PdfObject* imgColorSpace = xObject->GetIndirectKey("ColorSpace");
if (imgColorSpace)
{
PDFColorSpace retval = getCSType(imgColorSpace);
if (retval != CS_Unknown && !usedColorSpaces.contains(retval))
usedColorSpaces.append(retval);
}
PdfObject* sMaskObj = xObject->GetIndirectKey("SMask");
if (sMaskObj)
hasTransparency = true;
PDFImage img;
img.imgName = args[0].GetName().GetEscapedName().c_str();
double width = xObject->GetIndirectKey("Width")->GetReal();
double height = xObject->GetIndirectKey("Height")->GetReal();
img.dpiX = qRound(width/(currGS.ctm.m11()/72));
img.dpiY = qRound(height/(currGS.ctm.m22()/72));
imgs.append(img);
}
else if (subtypeObject->GetName() == "Form")
{
PdfXObject xObj(xObject);
inspectCanvas(&xObj, usedColorSpaces, hasTransparency, usedFonts, imgs); // recursive call
}
}
else
{
qDebug() << "Supplied external object is undefined!";
return false;
}
processedNamedXObj.append(args[0].GetName());
}
else
{
qDebug() << "Wrong syntax for Do operator or there's no XObject defined!";
return false;
}
}
}
break;
case KW_BI:
inlineImgDict = true;
break;
case KW_ID:
if (inlineImgDict)
{
PdfName colorspace("ColorSpace");
PdfName cs("CS");
if (args.contains(colorspace) || args.contains(cs))
{
int csIdx = args.contains(colorspace)?args.indexOf(colorspace):args.indexOf(cs);
if (args[csIdx+1].IsName())
{
PdfName csName = args[csIdx+1].GetName();
if ((csName == "G" || csName == "DeviceGray") && !usedColorSpaces.contains(CS_DeviceGray))
usedColorSpaces.append(CS_DeviceGray);
else if ((csName == "RGB" || csName == "DeviceRGB") && !usedColorSpaces.contains(CS_DeviceRGB))
usedColorSpaces.append(CS_DeviceRGB);
else if ((csName == "CMYK" || csName == "DeviceCMYK") && !usedColorSpaces.contains(CS_DeviceCMYK))
usedColorSpaces.append(CS_DeviceCMYK);
else if (!processedNamedCS.contains(csName))
{
if (colorSpaceRes && colorSpaceRes->GetIndirectKey(csName))
{
PdfObject* csEntry = colorSpaceRes->GetIndirectKey(csName);
if (csEntry)
{
PDFColorSpace retval = getCSType(csEntry);
if (retval != CS_Unknown && !usedColorSpaces.contains(retval))
usedColorSpaces.append(retval);
processedNamedCS.insert(csName, retval);
}
}
else
{
qDebug() << "Supplied colorspace for inline image is undefined!";
return false;
}
}
}
}
PdfName height("Height");
PdfName h("H");
PdfName width("Width");
PdfName w("W");
if ((args.contains(height) || args.contains(h)) && (args.contains(width) || args.contains(w)))
{
int heightIdx = args.contains(height)?args.indexOf(height):args.indexOf(h);
int widthIdx = args.contains(width)?args.indexOf(width):args.indexOf(w);
double height = args[heightIdx+1].GetReal();
double width = args[widthIdx+1].GetReal();
PDFImage img;
img.imgName = "Inline Image";
img.dpiX = qRound(width/(currGS.ctm.m11()/72));
img.dpiY = qRound(height/(currGS.ctm.m22()/72));
imgs.append(img);
}
inlineImgDict = false;
}
break;
case KW_gs:
{
if (!processedNamedGS.contains(args[0].GetName()))
{
if (args.size() == 1 && args[0].IsName() && extGState)
{
PdfObject* extGStateObj = extGState->GetIndirectKey(args[0].GetName());
if (extGStateObj)
{
inspectExtGStateObj(extGStateObj, usedColorSpaces, hasTransparency, usedFonts, currGS);
}
else
{
qDebug() << "Named graphic state used with gs operator is undefined in current ExtGState";
return false;
}
processedNamedGS.append(args[0].GetName());
}
else
{
qDebug() << "Wrong syntax in applying extended graphic state (gs operator) or there's no ExtGState defined!";
return false;
}
}
}
break;
case KW_Tf:
{
if (processedNamedFont.contains(args[0].GetName()))
{
currGS.font.first = processedNamedFont.value(args[0].GetName());
currGS.font.second = args[1].GetReal();
}
else
{
if (args.size() == 2 && args[0].IsName() && fontRes)
{
PdfObject* fontObj = fontRes->GetIndirectKey(args[0].GetName());
if (fontObj)
{
PDFFont retval = getFontInfo(fontObj);
usedFonts.append(retval);
processedNamedFont.insert(args[0].GetName(), retval);
currGS.font.first = retval;
currGS.font.second = args[1].GetReal();
}
else
{
qDebug() << "The specified font cannot be found in current Resources!";
return false;
}
}
else
{
qDebug() << "Wrong syntax in use of Tf operator or there's no Font defined in current Resources dictionary!";
return false;
}
}
}
break;
case KW_Undefined:
default:
break;
}
args.clear();
}
}
}
catch (PdfError & e)
{
qDebug() << "Error in parsing content stream";
e.PrintErrorMsg();
return false;
}
return true;
}
void PdfParser::ReadXRefStreamContents( pdf_long lOffset, bool bReadOnlyTrailer )
{
char* pBuffer;
char* pStart;
pdf_long lBufferLen;
long long lSize = 0;
PdfVariant vWArray;
PdfObject* pObj;
long nW[W_ARRAY_SIZE] = { 0, 0, 0 };
int i;
m_device.Device()->Seek( lOffset );
PdfParserObject xrefObject( m_vecObjects, m_device, m_buffer );
// Ignore the encryption in the XREF as the XREF stream must no be encrypted (see PDF Reference 3.4.7)
xrefObject.ParseFile( NULL );
if( !xrefObject.GetDictionary().HasKey( PdfName::KeyType ) )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
pObj = xrefObject.GetDictionary().GetKey( PdfName::KeyType );
if( !pObj->IsName() || ( pObj->GetName() != "XRef" ) )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
if( !m_pTrailer )
m_pTrailer = new PdfParserObject( m_vecObjects, m_device, m_buffer );
MergeTrailer( &xrefObject );
if( bReadOnlyTrailer )
return;
if( !xrefObject.GetDictionary().HasKey( PdfName::KeySize ) || !xrefObject.GetDictionary().HasKey( "W" ) )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
lSize = xrefObject.GetDictionary().GetKeyAsLong( PdfName::KeySize, 0 );
vWArray = *(xrefObject.GetDictionary().GetKey( "W" ));
// The pdf reference states that W is always an array with 3 entries
// all of them have to be integeres
if( !vWArray.IsArray() || vWArray.GetArray().size() != 3 )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
for( i=0;i<W_ARRAY_SIZE;i++ )
{
if( !vWArray.GetArray()[i].IsNumber() )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
nW[i] = static_cast<long>(vWArray.GetArray()[i].GetNumber());
}
std::vector<long long> vecIndeces;
// get the first object number in this crossref stream.
// it is not required to have an index key though.
if( xrefObject.GetDictionary().HasKey( "Index" ) )
{
// reuse vWArray!!
vWArray = *(xrefObject.GetDictionary().GetKey( "Index" ));
if( !vWArray.IsArray() )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
TCIVariantList it = vWArray.GetArray().begin();
while ( it != vWArray.GetArray().end() )
{
vecIndeces.push_back( (*it).GetNumber() );
++it;
}
}
else
{
vecIndeces.push_back( 0 );
vecIndeces.push_back( lSize );
}
if( vecIndeces.size() % 2 )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
if( !xrefObject.HasStreamToParse() )
{
PODOFO_RAISE_ERROR( ePdfError_NoXRef );
}
xrefObject.GetStream()->GetFilteredCopy( &pBuffer, &lBufferLen );
pStart = pBuffer;
int nCurIndex = 0;
while( nCurIndex < static_cast<pdf_long>(vecIndeces.size()) && pBuffer - pStart < lBufferLen )
{
int nFirstObj = static_cast<int>(vecIndeces[nCurIndex]);
long long nCount = vecIndeces[nCurIndex+1];
while( nCount-- && pBuffer - pStart < lBufferLen )
{
ReadXRefStreamEntry( pBuffer, lBufferLen, nW, nFirstObj++ );
pBuffer += (nW[0] + nW[1] + nW[2]);
}
nCurIndex += 2;
}
free( pStart );
if( xrefObject.GetDictionary().HasKey("Prev") )
{
lOffset = static_cast<pdf_long>(xrefObject.GetDictionary().GetKeyAsLong( "Prev", 0 ));
ReadXRefStreamContents( lOffset, bReadOnlyTrailer );
}
}