public IReadOnlyList <IFilter> GetFilters(PdfDictionary streamDictionary)
{
if (streamDictionary == null)
{
throw new ArgumentNullException(nameof(streamDictionary));
}
var filterObject = streamDictionary.GetItemOrDefault(CosName.FILTER);
if (filterObject == null)
{
return(new IFilter[0]);
}
switch (filterObject)
{
case COSArray filters:
// TODO: presumably this may be invalid...
return(filters.Select(x => GetFilterStrict((CosName)x)).ToList());
case CosName name:
return(new[] { GetFilterStrict(name) });
default:
throw new InvalidOperationException("The filter for a stream may be either a string or an array, instead this Pdf has: "
+ filterObject.GetType());
}
}
public PdfRawStream Parse(IRandomAccessRead reader, PdfDictionary streamDictionary, bool isLenientParsing, IPdfObjectParser parser)
{
PdfRawStream result;
// read 'stream'; this was already tested in parseObjectsDynamically()
ReadHelper.ReadExpectedString(reader, "stream");
skipWhiteSpaces(reader);
// This needs to be streamDictionary.getItem because when we are parsing, the underlying object might still be null.
ICosNumber streamLength = GetLength(reader, streamDictionary.GetItemOrDefault(CosName.LENGTH), streamDictionary.GetName(CosName.TYPE), isLenientParsing, parser);
ValidateStreamLength(reader, isLenientParsing, streamLength);
// get output stream to copy data to
using (var stream = new MemoryStream())
using (var writer = new BinaryWriter(stream))
{
if (streamLength != null && validateStreamLength(reader, streamLength.AsLong(), reader.Length()))
{
ReadValidStream(reader, writer, streamLength);
}
else
{
ReadUntilEndStream(reader, writer);
}
result = new PdfRawStream(stream.ToArray(), streamDictionary);
}
String endStream = ReadHelper.ReadString(reader);
if (endStream.Equals("endobj") && isLenientParsing)
{
log.Warn($"stream ends with \'endobj\' instead of \'endstream\' at offset {reader.GetPosition()}");
// avoid follow-up warning about missing endobj
reader.Rewind("endobj".Length);
}
else if (endStream.Length > 9 && isLenientParsing && endStream.Substring(0, 9).Equals("endstream"))
{
log.Warn("stream ends with '" + endStream + "' instead of 'endstream' at offset " + reader.GetPosition());
// unread the "extra" bytes
reader.Rewind(OtherEncodings.StringAsLatin1Bytes(endStream.Substring(9)).Length);
}
else if (!endStream.Equals("endstream"))
{
throw new InvalidOperationException("Error reading stream, expected='endstream' actual='"
+ endStream + "' at offset " + reader.GetPosition());
}
return(result);
}
public Page Create(int number, PdfDictionary dictionary, PageTreeMembers pageTreeMembers, IRandomAccessRead reader,
bool isLenientParsing)
{
if (dictionary == null)
{
throw new ArgumentNullException(nameof(dictionary));
}
var type = dictionary.GetName(CosName.TYPE);
if (type != null && !type.Equals(CosName.PAGE) && !isLenientParsing)
{
throw new InvalidOperationException($"Page {number} had its type was specified as {type} rather than 'Page'.");
}
MediaBox mediaBox = GetMediaBox(number, dictionary, pageTreeMembers, isLenientParsing);
CropBox cropBox = GetCropBox(dictionary, pageTreeMembers, mediaBox);
UserSpaceUnit userSpaceUnit = GetUserSpaceUnits(dictionary);
LoadResources(dictionary, reader, isLenientParsing);
PageContent content = default(PageContent);
var contentObject = dictionary.GetItemOrDefault(CosName.CONTENTS) as CosObject;
if (contentObject != null)
{
var contentStream = pdfObjectParser.Parse(contentObject.ToIndirectReference(), reader, false) as PdfRawStream;
if (contentStream == null)
{
throw new InvalidOperationException("Failed to parse the content for the page: " + number);
}
var contents = contentStream.Decode(filterProvider);
var txt = OtherEncodings.BytesAsLatin1String(contents);
var operations = pageContentParser.Parse(new ByteArrayInputBytes(contents));
var context = new ContentStreamProcessor(cropBox.Bounds, resourceStore, userSpaceUnit);
content = context.Process(operations);
}
var page = new Page(number, mediaBox, cropBox, content);
return(page);
}
public void LoadResources(PdfDictionary dictionary, IRandomAccessRead reader, bool isLenientParsing)
{
var resources = dictionary.GetItemOrDefault(CosName.RESOURCES);
if (resources is PdfDictionary resource)
{
resourceStore.LoadResourceDictionary(resource, reader, isLenientParsing);
return;
}
if (resources is CosObject resourceObject)
{
var resourceDictionary =
pdfObjectParser.Parse(resourceObject.ToIndirectReference(), reader, isLenientParsing);
if (resourceDictionary is PdfDictionary resolvedDictionary)
{
resourceStore.LoadResourceDictionary(resolvedDictionary, reader, isLenientParsing);
}
}
}
public CosBase Get(CosName name)
{
return(catalogDictionary.GetItemOrDefault(name));
}