private void ParseCIDChar(CMap codes, IList <object> operands)
{
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
MoveNext();
var inputCode = ParseInputCode();
int mappedCID = ConvertUtils.ByteArrayToInt(inputCode);
MoveNext();
var mappedCode = ParseUnicode();
codes.AddCIDMapping(mappedCode, mappedCID);
}
}
private void ParseBFChar(CMap codes, IList <object> operands)
{
//NOTE: The first element on each line is the input code of the template font;
//the second element is the code or name of the character.
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
MoveNext();
var inputCode = ParseInputCode();
MoveNext();
// FIXME: Unicode character sequences (such as ligatures) have not been supported yet [BUG:72].
try
{
codes.AddCharMapping(inputCode, ParseUnicode());
}
catch (OverflowException)
{ Debug.WriteLine($"WARN: Unable to process Unicode sequence from {codes.CMapName} CMap: {Token}"); }
}
}
private void ParseCIDRange(CMap codes, IList <object> operands)
{
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
// 1. Beginning input code.
MoveNext();
byte[] beginInputCode = ParseInputCode();
int beginInput = ConvertUtils.ByteArrayToInt(beginInputCode);
// 2. Ending input code.
MoveNext();
byte[] endInputCode = ParseInputCode();
int entInput = ConvertUtils.ByteArrayToInt(endInputCode);
MoveNext();
int mappedCode = ParseUnicode();
// 3. Character codes.
if (beginInputCode.Length <= 2 && endInputCode.Length <= 2)
{
// some CMaps are using CID ranges to map single values
if (beginInput == entInput)
{
codes.AddCIDMapping(mappedCode, beginInput);
}
else
{
codes.AddCIDRange((char)beginInput, (char)entInput, mappedCode);
}
}
else
{
// TODO Is this even possible?
int endOfMappings = mappedCode + entInput - beginInput;
while (mappedCode <= endOfMappings)
{
int mappedCID = ConvertUtils.ByteArrayToInt(beginInputCode);
codes.AddCIDMapping(mappedCode++, mappedCID);
OperationUtils.Increment(beginInputCode);
}
}
}
}
/**
* <summary>Parses the character-code-to-unicode mapping [PDF:1.6:5.9.1].</summary>
*/
public CMap Parse()
{
Stream.Seek(0);
var codes = new CMap();
{
IList <object> operands = new List <object>();
while (MoveNext())
{
switch (TokenType)
{
case TokenTypeEnum.Keyword:
{
string @operator = (string)Token;
if (@operator.Equals(BeginCodeSpaceRangeOperator, StringComparison.Ordinal))
{
ParseCodeSpaceRange(codes, operands);
}
else if (@operator.Equals(BeginBaseFontCharOperator, StringComparison.Ordinal))
{
ParseBFChar(codes, operands);
}
else if (@operator.Equals(BeginCIDCharOperator, StringComparison.Ordinal))
{
ParseCIDChar(codes, operands);
}
else if (@operator.Equals(BeginBaseFontRangeOperator, StringComparison.Ordinal))
{
ParseBFRange(codes, operands);
}
else if (@operator.Equals(BeginCIDRangeOperator, StringComparison.Ordinal))
{
ParseCIDRange(codes, operands);
}
else if (@operator.Equals(UseCMapOperator, StringComparison.Ordinal))
{
var useCMap = CMap.Get((string)operands[0]);
codes.UseCmap(useCMap);
}
else if (@operator.Equals(DefOperator, StringComparison.Ordinal) && operands.Count != 0)
{
if (CMapName.Equals((string)operands[0], StringComparison.Ordinal))
{
codes.CMapName = (string)operands[1];
}
else if (CMapType.Equals((string)operands[0], StringComparison.Ordinal))
{
codes.CMapType = (int)operands[1];
}
else if (Registry.Equals((string)operands[0], StringComparison.Ordinal))
{
codes.Registry = (string)operands[1];
}
else if (Ordering.Equals((string)operands[0], StringComparison.Ordinal))
{
codes.Ordering = (string)operands[1];
}
else if (WMode.Equals((string)operands[0], StringComparison.Ordinal))
{
codes.WMode = (int)operands[1];
}
}
operands.Clear();
break;
}
case TokenTypeEnum.ArrayBegin:
case TokenTypeEnum.DictionaryBegin:
{
// Skip.
while (MoveNext())
{
if (TokenType == TokenTypeEnum.ArrayEnd ||
TokenType == TokenTypeEnum.DictionaryEnd)
{
break;
}
}
break;
}
case TokenTypeEnum.Comment:
// Skip.
break;
default:
{
operands.Add(Token);
break;
}
}
}
}
return(codes);
}
private void ParseBFRange(CMap codes, IList <object> operands)
{
//NOTE: The first and second elements in each line are the beginning and
//ending valid input codes for the template font; the third element is
//the beginning character code for the range.
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
// 1. Beginning input code.
MoveNext();
byte[] beginInputCode = ParseInputCode();
int beginInput = ConvertUtils.ByteArrayToInt(beginInputCode);
// 2. Ending input code.
MoveNext();
byte[] endInputCode = ParseInputCode();
int entInput = ConvertUtils.ByteArrayToInt(endInputCode);
MoveNext();
switch (TokenType)
{
case TokenTypeEnum.ArrayBegin:
{
byte[] inputCode = beginInputCode;
while (MoveNext() &&
TokenType != TokenTypeEnum.ArrayEnd)
{
// FIXME: Unicode character sequences (such as ligatures) have not been supported yet [BUG:72].
try
{
codes.AddCharMapping(inputCode, ParseUnicode());
}
catch (OverflowException)
{ Debug.WriteLine($"WARN: Unable to process Unicode sequence from {codes.CMapName} CMap: {Token}"); }
OperationUtils.Increment(inputCode);
}
break;
}
default:
{
var tokenBytes = ParseInputCode();
if (tokenBytes.Length > 0)
{
// some pdfs use the malformed bfrange <0000> <FFFF> <0000>. Add support by adding a identity
// mapping for the whole range instead of cutting it after 255 entries
// TODO find a more efficient method to represent all values for a identity mapping
if (tokenBytes.Length == 2 && beginInput == 0 && entInput == 0xffff &&
tokenBytes[0] == 0 && tokenBytes[1] == 0)
{
for (int i = 0; i < 256; i++)
{
beginInputCode[1] = (byte)i;
tokenBytes[1] = (byte)i;
AddMappingFrombfrange(codes, beginInputCode, 0xff, tokenBytes);
}
}
else
{
// PDFBOX-4661: avoid overflow of the last byte, all following values are undefined
int values = Math.Min(entInput - beginInput,
255 - (tokenBytes[tokenBytes.Length - 1] & 0xFF)) + 1;
AddMappingFrombfrange(codes, beginInputCode, values, tokenBytes);
}
}
break;
}
}
}
}
/**
* <summary>Parses the character-code-to-unicode mapping [PDF:1.6:5.9.1].</summary>
*/
public IDictionary <ByteArray, int> Parse()
{
Stream.Seek(0);
IDictionary <ByteArray, int> codes = new Dictionary <ByteArray, int>();
{
IList <object> operands = new List <object>();
string cmapName = null;
int cmapType;
while (MoveNext())
{
switch (TokenType)
{
case TokenTypeEnum.Keyword:
{
string @operator = (string)Token;
if (@operator.Equals(BeginBaseFontCharOperator, StringComparison.Ordinal) ||
@operator.Equals(BeginCIDCharOperator, StringComparison.Ordinal))
{
/*
* NOTE: The first element on each line is the input code of the template font;
* the second element is the code or name of the character.
*/
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
MoveNext();
ByteArray inputCode = new ByteArray(ParseInputCode());
MoveNext();
// FIXME: Unicode character sequences (such as ligatures) have not been supported yet [BUG:72].
try
{
codes[inputCode] = ParseUnicode();
}
catch (OverflowException)
{ Debug.WriteLine(String.Format("WARN: Unable to process Unicode sequence from {0} CMap: {1}", cmapName, Token)); }
}
}
else if (@operator.Equals(BeginBaseFontRangeOperator, StringComparison.Ordinal) ||
@operator.Equals(BeginCIDRangeOperator, StringComparison.Ordinal))
{
/*
* NOTE: The first and second elements in each line are the beginning and
* ending valid input codes for the template font; the third element is
* the beginning character code for the range.
*/
for (int itemIndex = 0, itemCount = (int)operands[0]; itemIndex < itemCount; itemIndex++)
{
// 1. Beginning input code.
MoveNext();
byte[] beginInputCode = ParseInputCode();
// 2. Ending input code.
MoveNext();
byte[] endInputCode = ParseInputCode();
// 3. Character codes.
MoveNext();
switch (TokenType)
{
case TokenTypeEnum.ArrayBegin:
{
byte[] inputCode = beginInputCode;
while (MoveNext() &&
TokenType != TokenTypeEnum.ArrayEnd)
{
// FIXME: Unicode character sequences (such as ligatures) have not been supported yet [BUG:72].
try
{
codes[new ByteArray(inputCode)] = ParseUnicode();
}
catch (OverflowException)
{ Debug.WriteLine(String.Format("WARN: Unable to process Unicode sequence from {0} CMap: {1}", cmapName, Token)); }
OperationUtils.Increment(inputCode);
}
break;
}
default:
{
byte[] inputCode = beginInputCode;
int charCode = ParseUnicode();
int endCharCode = charCode + (ConvertUtils.ByteArrayToInt(endInputCode) - ConvertUtils.ByteArrayToInt(beginInputCode));
while (true)
{
codes[new ByteArray(inputCode)] = charCode;
if (charCode == endCharCode)
{
break;
}
OperationUtils.Increment(inputCode);
charCode++;
}
break;
}
}
}
}
else if (@operator.Equals(UseCMapOperator, StringComparison.Ordinal))
{
codes = CMap.Get((string)operands[0]);
}
else if (@operator.Equals(DefOperator, StringComparison.Ordinal) && operands.Count != 0)
{
if (CMapName.Equals((string)operands[0], StringComparison.Ordinal))
{
cmapName = (string)operands[1];
}
if (CMapType.Equals((string)operands[0], StringComparison.Ordinal))
{
cmapType = (int)operands[1];
}
}
operands.Clear();
break;
}
case TokenTypeEnum.ArrayBegin:
case TokenTypeEnum.DictionaryBegin:
{
// Skip.
while (MoveNext())
{
if (TokenType == TokenTypeEnum.ArrayEnd ||
TokenType == TokenTypeEnum.DictionaryEnd)
{
break;
}
}
break;
}
case TokenTypeEnum.Comment:
// Skip.
break;
default:
{
operands.Add(Token);
break;
}
}
}
}
return(codes);
}
protected void LoadEncoding()
{
PdfDataObject encodingObject = BaseDataObject.Resolve(PdfName.Encoding);
// CMap [PDF:1.6:5.6.4].
IDictionary <ByteArray, int> cmap = CMap.Get(encodingObject);
// 1. Unicode.
if (codes == null)
{
codes = new BiDictionary <ByteArray, int>();
if (encodingObject is PdfName &&
!(encodingObject.Equals(PdfName.IdentityH) ||
encodingObject.Equals(PdfName.IdentityV)))
{
/*
* NOTE: According to [PDF:1.6:5.9.1], the fallback method to retrieve
* the character-code-to-Unicode mapping implies getting the UCS2 CMap
* (Unicode value to CID) corresponding to the font's one (character code to CID);
* CIDs are the bridge from character codes to Unicode values.
*/
BiDictionary <ByteArray, int> ucs2CMap;
{
PdfDictionary cidSystemInfo = (PdfDictionary)CIDFontDictionary.Resolve(PdfName.CIDSystemInfo);
String registry = (String)((PdfTextString)cidSystemInfo[PdfName.Registry]).Value;
String ordering = (String)((PdfTextString)cidSystemInfo[PdfName.Ordering]).Value;
String ucs2CMapName = registry + "-" + ordering + "-" + "UCS2";
ucs2CMap = new BiDictionary <ByteArray, int>(CMap.Get(ucs2CMapName));
}
if (ucs2CMap.Count > 0)
{
foreach (KeyValuePair <ByteArray, int> cmapEntry in cmap)
{
codes[cmapEntry.Key] = ConvertUtils.ByteArrayToInt(ucs2CMap.GetKey(cmapEntry.Value).Data);
}
}
}
if (codes.Count == 0)
{
/*
* NOTE: In case no clue is available to determine the Unicode resolution map,
* the font is considered symbolic and an identity map is synthesized instead.
*/
symbolic = true;
foreach (KeyValuePair <ByteArray, int> cmapEntry in cmap)
{
codes[cmapEntry.Key] = ConvertUtils.ByteArrayToInt(cmapEntry.Key.Data);
}
}
}
// 2. Glyph indexes.
/*
* TODO: gids map for glyph indexes as glyphIndexes is used to map cids!!!
*/
// Character-code-to-CID mapping [PDF:1.6:5.6.4,5].
glyphIndexes = new Dictionary <int, int>();
foreach (KeyValuePair <ByteArray, int> cmapEntry in cmap)
{
if (!codes.ContainsKey(cmapEntry.Key))
{
continue;
}
glyphIndexes[codes[cmapEntry.Key]] = cmapEntry.Value;
}
}