int
pdf_get_rect(char* filename, int page_num, int pdf_box, realrect* box)
/* return the box converted to TeX points */
{
GooString* name = new GooString(filename);
PDFDoc* doc = new PDFDoc(name);
if (!doc) {
delete name;
return -1;
}
/* if the doc got created, it now owns name, so we mustn't delete it! */
if (!doc->isOk()) {
delete doc;
return -1;
}
int pages = doc->getNumPages();
if (page_num > pages)
page_num = pages;
if (page_num < 0)
page_num = pages + 1 + page_num;
if (page_num < 1)
page_num = 1;
Page* page = doc->getCatalog()->getPage(page_num);
PDFRectangle* r;
switch (pdf_box) {
default:
case pdfbox_crop:
r = page->getCropBox();
break;
case pdfbox_media:
r = page->getMediaBox();
break;
case pdfbox_bleed:
r = page->getBleedBox();
break;
case pdfbox_trim:
r = page->getTrimBox();
break;
case pdfbox_art:
r = page->getArtBox();
break;
}
box->x = 72.27 / 72 * my_fmin(r->x1, r->x2);
box->y = 72.27 / 72 * my_fmin(r->y1, r->y2);
box->wd = 72.27 / 72 * fabs(r->x2 - r->x1);
box->ht = 72.27 / 72 * fabs(r->y2 - r->y1);
delete doc;
return 0;
}
KoFilter::ConversionStatus PdfImport::convert(const QByteArray& from, const QByteArray& to)
{
debugPdf << "to:" << to << " from:" << from;
if (from != "application/pdf" || to != "image/svg+xml") {
return KoFilter::NotImplemented;
}
// read config file
globalParams = new GlobalParams();
if (! globalParams)
return KoFilter::NotImplemented;
GooString * fname = new GooString(QFile::encodeName(m_chain->inputFile()).data());
PDFDoc * pdfDoc = new PDFDoc(fname, 0, 0, 0);
if (! pdfDoc) {
delete globalParams;
return KoFilter::StupidError;
}
if (! pdfDoc->isOk()) {
delete globalParams;
delete pdfDoc;
return KoFilter::StupidError;
}
double hDPI = 72.0;
double vDPI = 72.0;
int firstPage = 1;
int lastPage = pdfDoc->getNumPages();
debugPdf << "converting pages" << firstPage << "-" << lastPage;
SvgOutputDev * dev = new SvgOutputDev(m_chain->outputFile());
if (dev->isOk()) {
int rotate = 0;
GBool useMediaBox = gTrue;
GBool crop = gFalse;
GBool printing = gFalse;
pdfDoc->displayPages(dev, firstPage, lastPage, hDPI, vDPI, rotate, useMediaBox, crop, printing);
dev->dumpContent();
}
debugPdf << "wrote file to" << m_chain->outputFile();
delete dev;
delete pdfDoc;
delete globalParams;
globalParams = 0;
// check for memory leaks
Object::memCheck(stderr);
return KoFilter::OK;
}
bool load(const char* fileName)
{
delete _pdfDoc;
delete _outputDev;
delete _nullOutputDev;
for (int i=0; i<_bmpCache.size(); i++) delete _bmpCache[i];
_bmpCache.resize(0);
for (int i=0; i<_textCache.size(); i++) delete _textCache[i];
_textCache.resize(0);
_pdfDoc= new PDFDoc(new GooString(fileName), NULL, NULL, NULL);
if (!_pdfDoc->isOk()) {
printf("error loading pdf");
return false;
}
GBool bitmapTopDown = gTrue;
SplashColor white;
white[0]=0xff;
white[1]=0xff;
white[2]=0xff;
// _outputDev = new SplashOutputDev_mod(splashModeRGB8, 4, gFalse, white, bitmapTopDown);
_outputDev = new SplashOutputDev(splashModeRGB8, 4, gFalse, white, bitmapTopDown);
if(!_outputDev)
{
printf("error loading pdf");
return false;
}
#ifdef USE_NULLOUTPUTDEV
_nullOutputDev=new NullOutputDev();
#endif
#ifdef POPPLER_VERSION
_outputDev->startDoc(_pdfDoc);
#else
_outputDev->startDoc(_pdfDoc->getXRef());
#endif
_bmpCache.resize(_pdfDoc->getNumPages());
for (int i=0; i<_bmpCache.size(); i++) _bmpCache[i]=NULL;
_textCache.resize(_pdfDoc->getNumPages());
for (int i=0; i<_textCache.size(); i++) _textCache[i]=NULL;
return true;
}
unsigned char
epdf_document_unlock (Epdf_Document *document, const char *password)
{
if (!document)
return 0;
if (document->locked) {
/* racier then it needs to be */
GooString *pwd = new GooString(password);
PDFDoc *pdfdoc = new PDFDoc(document->pdfdoc->getFileName(), pwd);
delete pwd;
if (pdfdoc->isOk()) {
document->pdfdoc = pdfdoc;
document->locked = false;
}
}
return document->locked;
}
gboolean
import_pdf(const gchar *filename, DiagramData *dia, DiaContext *ctx, void* user_data)
{
PDFDoc *doc;
GooString *fileName = new GooString(filename);
// no passwords yet
GooString *ownerPW = NULL;
GooString *userPW = NULL;
gboolean ret = FALSE;
// without this we will get strange crashes (at least with /O2 build)
globalParams = new GlobalParams();
doc = PDFDocFactory().createPDFDoc(*fileName, ownerPW, userPW);
if (!doc->isOk()) {
dia_context_add_message (ctx, _("PDF document not OK.\n%s"),
dia_context_get_filename (ctx));
} else {
DiaOutputDev *diaOut = new DiaOutputDev(dia, doc->getNumPages());
for (int pg = 1; pg <= doc->getNumPages(); ++pg) {
Page *page = doc->getPage (pg);
if (!page || !page->isOk())
continue;
doc->displayPage(diaOut, pg,
72.0, 72.0, /* DPI, scaling elsewhere */
0, /* rotate */
gTrue, /* useMediaBox */
gTrue, /* Crop */
gFalse /* printing */
);
}
delete diaOut;
ret = TRUE;
}
delete doc;
delete globalParams;
delete fileName;
return ret;
}
int
pdf_count_pages(char* filename)
{
int pages = 0;
GooString* name = new GooString(filename);
PDFDoc* doc = new PDFDoc(name);
if (!doc) {
delete name;
return 0;
}
/* if the doc got created, it now owns name, so we mustn't delete it! */
if (doc->isOk())
pages = doc->getNumPages();
delete doc;
return pages;
}
bool XojPopplerDocument::load(const char * filename, const char * password, GError ** error) {
XOJ_CHECK_TYPE(XojPopplerDocument);
PDFDoc * newDoc;
GooString * filename_g;
GooString * password_g;
if (!globalParams) {
globalParams = new GlobalParams();
}
if (!filename) {
return false;
}
password_g = NULL;
if (password != NULL) {
if (g_utf8_validate(password, -1, NULL)) {
gchar *password_latin;
password_latin = g_convert(password, -1, "ISO-8859-1", "UTF-8", NULL, NULL, NULL);
password_g = new GooString(password_latin);
g_free(password_latin);
} else {
password_g = new GooString(password);
}
}
#ifdef G_OS_WIN32
wchar_t *filenameW;
int wlen;
wlen = MultiByteToWideChar(CP_UTF8, 0, filename, -1, NULL, 0);
filenameW = new WCHAR[wlen];
if (!filenameW)
return NULL;
wlen = MultiByteToWideChar(CP_UTF8, 0, filename, -1, filenameW, wlen);
newDoc = new PDFDoc(filenameW, wlen, password_g, password_g);
delete filenameW;
#else
filename_g = new GooString(filename);
newDoc = new PDFDoc(filename_g, password_g, password_g);
#endif
delete password_g;
if (!newDoc->isOk()) {
int fopen_errno;
switch (newDoc->getErrorCode()) {
case errOpenFile:
// If there was an error opening the file, count it as a G_FILE_ERROR
// and set the GError parameters accordingly. (this assumes that the
// only way to get an errOpenFile error is if newDoc was created using
// a filename and thus fopen was called, which right now is true.
fopen_errno = newDoc->getFopenErrno();
g_set_error(error, G_FILE_ERROR, g_file_error_from_errno(fopen_errno), "%s", g_strerror(fopen_errno));
break;
case errBadCatalog:
g_set_error(error, 0, 0, "Failed to read the document catalog");
break;
case errDamaged:
g_set_error(error, 0, 0, "PDF document is damaged");
break;
case errEncrypted:
g_set_error(error, 0, 0, "Document is encrypted");
break;
default:
g_set_error(error, 0, 0, "Failed to load document");
}
delete newDoc;
return false;
}
if (this->data) {
this->data->unreference();
}
this->data = new _IntPopplerDocument(newDoc);
return true;
}
bool PdfPlug::convert(QString fn)
{
bool firstPg = true;
int currentLayer = m_Doc->activeLayer();
int baseLayer = m_Doc->activeLayer();
importedColors.clear();
if(progressDialog)
{
progressDialog->setOverallProgress(2);
progressDialog->setLabel("GI", tr("Generating Items"));
qApp->processEvents();
}
QFile f(fn);
oldDocItemCount = m_Doc->Items->count();
if (progressDialog)
{
progressDialog->setBusyIndicator("GI");
qApp->processEvents();
}
globalParams = new GlobalParams();
if (globalParams)
{
GooString *fname = new GooString(QFile::encodeName(fn).data());
globalParams->setErrQuiet(gTrue);
GBool hasOcg = gFalse;
QList<OptionalContentGroup*> ocgGroups;
// globalParams->setPrintCommands(gTrue);
PDFDoc *pdfDoc = new PDFDoc(fname, 0, 0, 0);
if (pdfDoc)
{
if (pdfDoc->isOk())
{
double hDPI = 72.0;
double vDPI = 72.0;
int firstPage = 1;
int lastPage = pdfDoc->getNumPages();
SlaOutputDev *dev = new SlaOutputDev(m_Doc, &Elements, &importedColors, importerFlags);
if (dev->isOk())
{
OCGs* ocg = pdfDoc->getOptContentConfig();
if (ocg)
{
hasOcg = ocg->hasOCGs();
if (hasOcg)
{
QStringList ocgNames;
Array *order = ocg->getOrderArray();
if (order)
{
for (int i = 0; i < order->getLength (); ++i)
{
Object orderItem;
order->get(i, &orderItem);
if (orderItem.isDict())
{
Object ref;
order->getNF(i, &ref);
if (ref.isRef())
{
OptionalContentGroup *oc = ocg->findOcgByRef(ref.getRef());
QString ocgName = UnicodeParsedString(oc->getName());
if (!ocgNames.contains(ocgName))
{
ocgGroups.prepend(oc);
ocgNames.append(ocgName);
}
}
ref.free();
}
else
{
GooList *ocgs;
int i;
ocgs = ocg->getOCGs ();
for (i = 0; i < ocgs->getLength (); ++i)
{
OptionalContentGroup *oc = (OptionalContentGroup *)ocgs->get(i);
QString ocgName = UnicodeParsedString(oc->getName());
if (!ocgNames.contains(ocgName))
{
ocgGroups.prepend(oc);
ocgNames.append(ocgName);
}
}
}
}
}
else
{
GooList *ocgs;
int i;
ocgs = ocg->getOCGs ();
for (i = 0; i < ocgs->getLength (); ++i)
{
OptionalContentGroup *oc = (OptionalContentGroup *)ocgs->get(i);
QString ocgName = UnicodeParsedString(oc->getName());
if (!ocgNames.contains(ocgName))
{
ocgGroups.prepend(oc);
ocgNames.append(ocgName);
}
}
}
}
}
GBool useMediaBox = gTrue;
GBool crop = gFalse;
GBool printing = gFalse;
dev->startDoc(pdfDoc, pdfDoc->getXRef(), pdfDoc->getCatalog());
int rotate = pdfDoc->getPageRotate(firstPage);
if (importerFlags & LoadSavePlugin::lfCreateDoc)
{
// POPPLER_VERSION appeared in 0.19.0 first
#ifdef POPPLER_VERSION
if (hasOcg)
{
QString actL = m_Doc->activeLayerName();
for (int a = 0; a < ocgGroups.count(); a++)
{
OptionalContentGroup *oc = ocgGroups[a];
if (actL != UnicodeParsedString(oc->getName()))
currentLayer = m_Doc->addLayer(UnicodeParsedString(oc->getName()), false);
else
currentLayer = m_Doc->layerIDFromName(UnicodeParsedString(oc->getName()));
// POPPLER_VERSION appeared in 0.19.0 first
#ifdef POPPLER_VERSION
if ((oc->getViewState() == OptionalContentGroup::ocUsageOn) || (oc->getViewState() == OptionalContentGroup::ocUsageUnset))
m_Doc->setLayerVisible(currentLayer, true);
else
m_Doc->setLayerVisible(currentLayer, false);
if ((oc->getPrintState() == OptionalContentGroup::ocUsageOn) || (oc->getPrintState() == OptionalContentGroup::ocUsageUnset))
m_Doc->setLayerPrintable(currentLayer, true);
else
m_Doc->setLayerPrintable(currentLayer, false);
#else
if (oc->getState() == OptionalContentGroup::On)
{
m_Doc->setLayerVisible(currentLayer, true);
m_Doc->setLayerPrintable(currentLayer, true);
}
else
{
m_Doc->setLayerVisible(currentLayer, false);
m_Doc->setLayerPrintable(currentLayer, false);
}
#endif
oc->setState(OptionalContentGroup::Off);
}
dev->layersSetByOCG = true;
}
#endif
Object info;
pdfDoc->getDocInfo(&info);
if (info.isDict())
{
Object obj;
GooString *s1;
Dict *infoDict = info.getDict();
if (infoDict->lookup((char*)"Title", &obj )->isString())
{
s1 = obj.getString();
m_Doc->documentInfo().setTitle(UnicodeParsedString(obj.getString()));
obj.free();
}
if (infoDict->lookup((char*)"Author", &obj )->isString())
{
s1 = obj.getString();
m_Doc->documentInfo().setAuthor(UnicodeParsedString(obj.getString()));
obj.free();
}
if (infoDict->lookup((char*)"Subject", &obj )->isString())
{
s1 = obj.getString();
m_Doc->documentInfo().setSubject(UnicodeParsedString(obj.getString()));
obj.free();
}
if (infoDict->lookup((char*)"Keywords", &obj )->isString())
{
s1 = obj.getString();
m_Doc->documentInfo().setKeywords(UnicodeParsedString(obj.getString()));
obj.free();
}
}
info.free();
for (int pp = 0; pp < lastPage; pp++)
{
m_Doc->setActiveLayer(baseLayer);
if (firstPg)
firstPg = false;
else
m_Doc->addPage(pp);
m_Doc->currentPage()->setInitialHeight(pdfDoc->getPageMediaHeight(pp + 1));
m_Doc->currentPage()->setInitialWidth(pdfDoc->getPageMediaWidth(pp + 1));
m_Doc->currentPage()->setHeight(pdfDoc->getPageMediaHeight(pp + 1));
m_Doc->currentPage()->setWidth(pdfDoc->getPageMediaWidth(pp + 1));
m_Doc->currentPage()->MPageNam = CommonStrings::trMasterPageNormal;
m_Doc->currentPage()->m_pageSize = "Custom";
m_Doc->setPageSize("Custom");
m_Doc->reformPages(true);
if (hasOcg)
{
for (int a = 0; a < ocgGroups.count(); a++)
{
OptionalContentGroup *oc = ocgGroups[a];
// m_Doc->setActiveLayer(UnicodeParsedString(oc->getName()));
// currentLayer = m_Doc->activeLayer();
oc->setState(OptionalContentGroup::On);
// pdfDoc->displayPage(dev, pp + 1, hDPI, vDPI, rotate, useMediaBox, crop, printing);
// oc->setState(OptionalContentGroup::Off);
}
pdfDoc->displayPage(dev, pp + 1, hDPI, vDPI, rotate, useMediaBox, crop, printing);
}
else
pdfDoc->displayPage(dev, pp + 1, hDPI, vDPI, rotate, useMediaBox, crop, printing);
}
}
else
{
if (hasOcg)
{
for (int a = 0; a < ocgGroups.count(); a++)
{
ocgGroups[a]->setState(OptionalContentGroup::On);
}
}
pdfDoc->displayPage(dev, firstPage, hDPI, vDPI, rotate, useMediaBox, crop, printing);
}
}
delete dev;
}
}
delete pdfDoc;
}
delete globalParams;
globalParams = 0;
// qDebug() << "converting finished";
// qDebug() << "Imported" << Elements.count() << "Elements";
if (Elements.count() == 0)
{
if (importedColors.count() != 0)
{
for (int cd = 0; cd < importedColors.count(); cd++)
{
m_Doc->PageColors.remove(importedColors[cd]);
}
}
}
if (progressDialog)
progressDialog->close();
return true;
}
int main(int argc, char *argv[])
{
// parse args
bool ok = parseArgs(argDesc, &argc, argv);
if (!ok || argc < 2 || argc > 3 || printHelp) {
fprintf(stderr, "pdftoipe version %s\n", PDFTOIPE_VERSION);
printUsage("pdftoipe", "<PDF-file> [<XML-file>]", argDesc);
return 1;
}
GooString *fileName = new GooString(argv[1]);
globalParams = new GlobalParams();
if (quiet)
globalParams->setErrQuiet(quiet);
GooString *ownerPW, *userPW;
if (ownerPassword[0]) {
ownerPW = new GooString(ownerPassword);
} else {
ownerPW = 0;
}
if (userPassword[0]) {
userPW = new GooString(userPassword);
} else {
userPW = 0;
}
// open PDF file
PDFDoc *doc = new PDFDoc(fileName, ownerPW, userPW);
delete userPW;
delete ownerPW;
if (!doc->isOk())
return 1;
// construct XML file name
std::string xmlFileName;
if (argc == 3) {
xmlFileName = argv[2];
} else {
const char *p = fileName->c_str() + fileName->getLength() - 4;
if (!strcmp(p, ".pdf") || !strcmp(p, ".PDF")) {
xmlFileName = std::string(fileName->c_str(),
fileName->getLength() - 4);
} else {
xmlFileName = fileName->c_str();
}
xmlFileName += ".ipe";
}
// get page range
if (firstPage < 1)
firstPage = 1;
if (lastPage < 1 || lastPage > doc->getNumPages())
lastPage = doc->getNumPages();
// write XML file
XmlOutputDev *xmlOut =
new XmlOutputDev(xmlFileName, doc->getXRef(),
doc->getCatalog(), firstPage, lastPage);
// tell output device about text handling
xmlOut->setTextHandling(math, notext, literal, mergeLevel, unicodeLevel);
int exitCode = 2;
if (xmlOut->isOk()) {
doc->displayPages(xmlOut, firstPage, lastPage,
// double hDPI, double vDPI, int rotate,
// bool useMediaBox, bool crop, bool printing,
72.0, 72.0, 0, false, false, false);
exitCode = 0;
}
if (xmlOut->hasUnicode()) {
fprintf(stderr, "The document contains Unicode (non-ASCII) text.\n");
if (unicodeLevel <= 1)
fprintf(stderr, "Unknown Unicode characters were replaced by [U+XXX].\n");
else
fprintf(stderr, "UTF-8 was set as document encoding in the preamble.\n");
}
// clean up
delete xmlOut;
delete doc;
delete globalParams;
return exitCode;
}
int main(int argc, char *argv[]) {
PDFDoc *doc;
GString *fileName;
GString *ownerPW, *userPW;
SplashColor paperColor;
SplashOutputDev *splashOut;
GBool ok;
int exitCode;
int pg;
exitCode = 99;
// parse args
ok = parseArgs(argDesc, &argc, argv);
if (!ok || argc != 2 || printVersion || printHelp) {
fprintf(stderr, "pdf2jpeg version %s\n", xpdfVersion);
fprintf(stderr, "%s\n", xpdfCopyright);
if (!printVersion) {
printUsage("pdf2jpeg", "<PDF-file> -o <jpegfile>", argDesc);
}
goto err0;
}
fileName = new GString(argv[1]);
// read config file
globalParams = new GlobalParams(cfgFileName);
globalParams->setupBaseFonts(NULL);
// open PDF file
if (ownerPassword[0]) {
ownerPW = new GString(ownerPassword);
} else {
ownerPW = NULL;
}
if (userPassword[0]) {
userPW = new GString(userPassword);
} else {
userPW = NULL;
}
doc = new PDFDoc(fileName, ownerPW, userPW);
if (userPW) {
delete userPW;
}
if (ownerPW) {
delete ownerPW;
}
if (!doc->isOk()) {
exitCode = 1;
goto err1;
}
paperColor[0] = paperColor[1] = paperColor[2] = 0xff;
splashOut = new SplashOutputDev(splashModeRGB8, 1, gFalse, paperColor);
splashOut->startDoc(doc->getXRef());
if(page>=1 && page<=doc->getNumPages()) {
double r = resolution;
if(width) {
int old_width = doc->getPageCropWidth(page);
r = 72.0*width/old_width;
}
doc->displayPage(splashOut, page, r, r, 0, gFalse, gTrue, gFalse);
SplashBitmap*bitmap = splashOut->getBitmap();
if(bitmap) {
Guchar*rgb = bitmap->getDataPtr();
int width = bitmap->getWidth();
int height = bitmap->getHeight();
jpeg_save(rgb, width, height, quality, output);
}
}
delete splashOut;
exitCode = 0;
// clean up
err1:
delete doc;
delete globalParams;
err0:
// check for memory leaks
Object::memCheck(stderr);
gMemReport(stderr);
return exitCode;
}
int extract_images_from_pdf(char* filename,
char* target,
char* owner_password,
char* user_password,
char* range,
char* format,
int jpg_quality,
GBool dump_jpg,
GBool tiff_jpg)
{
if (user_cancelled)
return gpret_user_cancelled;
// load config
xpdf_rc xrc;
// open file
xpdf_doc xdoc(filename, owner_password, user_password);
PDFDoc* doc = xdoc.get_doc();
if (!doc->isOk())
return doc->getErrorCode() == errEncrypted ?
gpret_pdf_encrypted :
gpret_cant_open_pdf;
// check for copy permission
// if (!doc->okToCopy())
// return gpret_dont_allow_copy;
// get page range
page_range range_list(range);
if (*range == '\0')
{
range_list.add_item(range_item(1, doc->getNumPages()));
}
if (user_cancelled)
return gpret_user_cancelled;
// write image files
fi_loader fi;
int progress = 0;
image_extractor img_out(target, dump_jpg, format, jpg_quality, tiff_jpg);
for (int i = 0; i < range_list.item_count(); i++)
{
range_item& item = range_list.get_item(i);
for (int pg = item.first;
pg <= min(item.last, doc->getNumPages());
pg++)
{
if (user_cancelled)
return gpret_user_cancelled;
doc->displayPage(&img_out, pg, 72, 72, 0, gFalse, gTrue, gFalse);
printf("progress: %d\n",
++progress * 100 / range_list.page_count());
}
}
printf("image count: %d\n", img_out.get_image_number());
return gpret_success;
}
int main(int argc, char *argv[])
{
int opt;
PDFRectangle rect;
while ((opt = getopt(argc, argv, "dop:m:")) != -1) {
switch (opt) {
case 'd':
opt_debug++;
break;
case 'o':
opt_oreilly = 1;
break;
case 'p':
opt_page = atoi(optarg);
break;
case 'm':
sscanf(optarg, "%lf %lf %lf %lf",
&rect.x1, &rect.y1, &rect.x2, &rect.y2);
opt_margin = ▭
break;
}
}
if (optind + 2 > argc) {
fprintf(stderr, "Usage: %s [infile] [outfile]\n", argv[0]);
fprintf(stderr, "\n");
return EXIT_FAILURE;
}
globalParams = new GlobalParams();
GooString *in_file = new GooString(argv[optind]);
GooString *out_file = new GooString(argv[optind+1]);
if(opt_debug){
printf("Input File: %s\n", in_file->getCString());
printf("Output File: %s\n", out_file->getCString());
}
//GooString *owner_pw = new GooString("");
PDFDoc *doc = new PDFDoc(in_file, NULL, NULL);
if (!doc->isOk()) {
perror("pdf open error");
return EXIT_FAILURE;
}
int ret;
ret = pdfcrop(doc);
if (ret) {
printf("crop failed: %d\n", ret);
return EXIT_FAILURE;
}
//ret = doc->saveAs(out_file, writeStandard);
ret = doc->saveAs(out_file, writeForceRewrite);
if (ret) {
printf("save failed: %d\n", ret);
return EXIT_FAILURE;
}
delete doc;
// double free?
//delete in_file;
//delete out_file;
// some times blocked.
//delete globalParams;
return EXIT_SUCCESS;
}
indri::parse::UnparsedDocument* indri::parse::PDFDocumentExtractor::nextDocument() {
if( !_documentPath.length() )
return 0;
PDFDoc* doc = 0;
TextOutputDev* textOut = 0;
GString* gfilename = new GString(_documentPath.c_str());
doc = new PDFDoc( gfilename );
// if the doc is not ok, or ok to copy, it
// will be a document of length 0.
if( doc->isOk() && doc->okToCopy() ) {
void* stream = &_documentTextBuffer;
textOut = new TextOutputDev( buffer_write, stream, gFalse, gFalse);
if ( textOut->isOk() ) {
int firstPage = 1;
int lastPage = doc->getNumPages();
double hDPI=72.0;
double vDPI=72.0;
int rotate=0;
GBool useMediaBox=gFalse;
GBool crop=gTrue;
GBool printing=gFalse;
if(doc->readMetadata()!=NULL)
{
GString rawMetaData = doc->readMetadata();
GString preparedMetaData="";
//zoek <rdf:RDF en eindig bij </rdf:RDF>!!
for(int x=0; x<rawMetaData.getLength(); x++) {
if(rawMetaData.getChar(x)!='?' && rawMetaData.getChar(x)!=':') {
//skip characters which the XMLReader doesn't understand
preparedMetaData.append(rawMetaData.getChar(x));
}
}
std::string metaData(preparedMetaData.getCString());
int startbegin = metaData.find("<rdf");
int stopend = metaData.find(">", metaData.rfind("</rdf") );
metaData = metaData.substr(startbegin, (stopend-startbegin)+1 );
indri::xml::XMLReader reader;
try {
std::auto_ptr<indri::xml::XMLNode> result( reader.read( metaData.c_str() ) );
appendPdfMetaData( result.get() );
} catch( lemur::api::Exception& e ) {
LEMUR_RETHROW( e, "Had trouble reading PDF metadata" );
}
if( _author.length()>0 || _title.length()>0 )
{
std::string createdPdfHeader;
createdPdfHeader="<head>\n";
if(_title.length()>0) {
createdPdfHeader+="<title>";
createdPdfHeader+=_title;
createdPdfHeader+="</title>\n";
}
if(_author.length()>0) {
createdPdfHeader+="<author>";
createdPdfHeader+=_author;
createdPdfHeader+="</author>\n";
}
createdPdfHeader+="</head>\n";
char *metastream = _documentTextBuffer.write( createdPdfHeader.length()+1 );
strcpy(metastream, createdPdfHeader.c_str());
}
}
doc->displayPages(textOut, firstPage, lastPage, hDPI, vDPI, rotate, useMediaBox, crop, printing);
}
}
delete textOut;
delete doc;
_unparsedDocument.textLength = _documentTextBuffer.position();
_unparsedDocument.contentLength = _unparsedDocument.textLength ? _documentTextBuffer.position() - 1 : 0 ; // no null 0 if text is empty.
char* docnoPoint = _documentTextBuffer.write( _documentPath.length()+1 );
strcpy( docnoPoint, _documentPath.c_str() );
_unparsedDocument.text = _documentTextBuffer.front();
_unparsedDocument.content = _documentTextBuffer.front();
_unparsedDocument.metadata.clear();
indri::parse::MetadataPair pair;
pair.key = "path";
pair.value = docnoPoint;
pair.valueLength = _documentPath.length()+1;
_unparsedDocument.metadata.push_back( pair );
_docnostring.assign(_documentPath.c_str() );
cleanDocno();
pair.value = _docnostring.c_str();
pair.valueLength = _docnostring.length()+1;
pair.key = "docno";
_unparsedDocument.metadata.push_back( pair );
_documentPath = "";
return &_unparsedDocument;
}