private void CheckInsertAndRetrieval(byte[] bytes, string uni)
{
CMapToUnicode c = new CMapToUnicode();
c.AddChar(new PdfString(bytes), new PdfString(uni, "UTF-16BE"));
string lookupResult = c.Lookup(bytes, 0, bytes.Length);
Assert.AreEqual(uni, lookupResult);
}
public void TestCMapWithDefDictionaryKey()
{
byte[] touni = TestResourceUtils.GetResourceAsByteArray(TEST_RESOURCES_PATH, "cmap_with_def_dictionary_key.txt");
CidLocationFromByte lb = new CidLocationFromByte(touni);
CMapToUnicode cmapRet = new CMapToUnicode();
CMapParserEx.ParseCid("", cmapRet, lb);
}
virtual public CMapToUnicode ExportToUnicode() {
CMapToUnicode uni = new CMapToUnicode();
int[] keys = map.GetKeys();
foreach (int key in keys) {
uni.AddChar(map[key], Utilities.ConvertFromUtf32(key));
}
return uni;
}
public static CMapToUnicode GetMapFromOrdering(String ordering)
{
if (ordering.Equals("CNS1"))
{
if (identityCNS == null)
{
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniCNS-UTF16-H");
if (uni == null)
{
return(null);
}
identityCNS = uni.ExportToUnicode();
}
return(identityCNS);
}
else if (ordering.Equals("Japan1"))
{
if (identityJapan == null)
{
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniJIS-UTF16-H");
if (uni == null)
{
return(null);
}
identityJapan = uni.ExportToUnicode();
}
return(identityJapan);
}
else if (ordering.Equals("Korea1"))
{
if (identityKorea == null)
{
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniKS-UTF16-H");
if (uni == null)
{
return(null);
}
identityKorea = uni.ExportToUnicode();
}
return(identityKorea);
}
else if (ordering.Equals("GB1"))
{
if (identityGB == null)
{
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniGB-UTF16-H");
if (uni == null)
{
return(null);
}
identityGB = uni.ExportToUnicode();
}
return(identityGB);
}
return(null);
}
public static CMapToUnicode GetIdentity()
{
CMapToUnicode uni = new CMapToUnicode();
for (int i = 0; i < 65537; i++)
{
uni.AddChar(i, Utilities.ConvertFromUtf32(i));
}
return(uni);
}
virtual public CMapToUnicode ExportToUnicode()
{
CMapToUnicode uni = new CMapToUnicode();
int[] keys = map.GetKeys();
foreach (int key in keys)
{
uni.AddChar(map[key], Utilities.ConvertFromUtf32(key));
}
return(uni);
}
public static CMapToUnicode GetMapFromOrdering(String ordering){
if (ordering.Equals("CNS1")) {
if (identityCNS == null) {
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniCNS-UTF16-H");
if (uni == null)
return null;
identityCNS = uni.ExportToUnicode();
}
return identityCNS;
}
else if (ordering.Equals("Japan1")) {
if (identityJapan == null) {
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniJIS-UTF16-H");
if (uni == null)
return null;
identityJapan = uni.ExportToUnicode();
}
return identityJapan;
}
else if (ordering.Equals("Korea1")) {
if (identityKorea == null) {
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniKS-UTF16-H");
if (uni == null)
return null;
identityKorea = uni.ExportToUnicode();
}
return identityKorea;
}
else if (ordering.Equals("GB1")) {
if (identityGB == null) {
CMapUniCid uni = CMapCache.GetCachedCMapUniCid("UniGB-UTF16-H");
if (uni == null)
return null;
identityGB = uni.ExportToUnicode();
}
return identityGB;
}
else if (ordering.Equals("Identity")) {
if (identityH == null) {
identityH = CMapToUnicode.GetIdentity();
}
return identityH;
}
return null;
}
/**
* Parses the ToUnicode entry, if present, and constructs a CMap for it
* @since 2.1.7
*/
private void ProcessToUnicode(){
PdfObject toUni = PdfReader.GetPdfObjectRelease(fontDic.Get(PdfName.TOUNICODE));
if (toUni is PRStream) {
try {
byte[] touni = PdfReader.GetStreamBytes((PRStream)toUni);
CidLocationFromByte lb = new CidLocationFromByte(touni);
toUnicodeCmap = new CMapToUnicode();
CMapParserEx.ParseCid("", toUnicodeCmap, lb);
uni2cid = toUnicodeCmap.CreateReverseMapping();
} catch {
toUnicodeCmap = null;
uni2cid = null;
// technically, we should log this or provide some sort of feedback... but sometimes the cmap will be junk, but it's still possible to get text, so we don't want to throw an exception
//throw new IllegalStateException("Unable to process ToUnicode map - " + e.GetMessage(), e);
}
}
else if(isType0) {
// fake a ToUnicode for CJK Identity-H fonts
try {
PdfName encodingName = fontDic.GetAsName(PdfName.ENCODING);
if(encodingName == null)
return;
String enc = PdfName.DecodeName(encodingName.ToString());
if(!enc.Equals("Identity-H"))
return;
PdfArray df = (PdfArray)PdfReader.GetPdfObjectRelease(fontDic.Get(PdfName.DESCENDANTFONTS));
PdfDictionary cidft = (PdfDictionary)PdfReader.GetPdfObjectRelease(df[0]);
PdfDictionary cidinfo = cidft.GetAsDict(PdfName.CIDSYSTEMINFO);
if(cidinfo == null)
return;
PdfString ordering = cidinfo.GetAsString(PdfName.ORDERING);
if(ordering == null)
return;
CMapToUnicode touni = IdentityToUnicode.GetMapFromOrdering(ordering.ToUnicodeString());
if(touni == null)
return;
toUnicodeCmap = touni;
uni2cid = toUnicodeCmap.CreateReverseMapping();
} catch(IOException e) {
toUnicodeCmap = null;
uni2cid = null;
}
}
}
private CMapToUnicode ProcessToUnicode() {
CMapToUnicode cmapRet = null;
PdfObject toUni = PdfReader.GetPdfObjectRelease(this.font.Get(PdfName.TOUNICODE));
if (toUni is PRStream) {
try {
byte[] touni = PdfReader.GetStreamBytes((PRStream)toUni);
CidLocationFromByte lb = new CidLocationFromByte(touni);
cmapRet = new CMapToUnicode();
CMapParserEx.ParseCid("", cmapRet, lb);
} catch {
cmapRet = null;
}
}
return cmapRet;
}
private void DoType1TT() {
CMapToUnicode toUnicode = null;
PdfObject enc = PdfReader.GetPdfObject(font.Get(PdfName.ENCODING));
if (enc == null) {
FillEncoding(null);
toUnicode = ProcessToUnicode();
if (toUnicode != null) {
IDictionary<int, int> rm = toUnicode.CreateReverseMapping();
foreach (KeyValuePair<int,int> kv in rm) {
uni2byte[kv.Key] = kv.Value;
byte2uni[kv.Value] = kv.Key;
}
}
}
else {
if (enc.IsName())
FillEncoding((PdfName)enc);
else if (enc.IsDictionary()) {
PdfDictionary encDic = (PdfDictionary)enc;
enc = PdfReader.GetPdfObject(encDic.Get(PdfName.BASEENCODING));
if (enc == null)
FillEncoding(null);
else
FillEncoding((PdfName)enc);
PdfArray diffs = encDic.GetAsArray(PdfName.DIFFERENCES);
if (diffs != null) {
diffmap = new IntHashtable();
int currentNumber = 0;
for (int k = 0; k < diffs.Size; ++k) {
PdfObject obj = diffs[k];
if (obj.IsNumber())
currentNumber = ((PdfNumber)obj).IntValue;
else {
int[] c = GlyphList.NameToUnicode(PdfName.DecodeName(((PdfName)obj).ToString()));
if (c != null && c.Length > 0) {
uni2byte[c[0]] = currentNumber;
byte2uni[currentNumber] = c[0];
diffmap[c[0]] = currentNumber;
}
else {
if (toUnicode == null) {
toUnicode = ProcessToUnicode();
if (toUnicode == null) {
toUnicode = new CMapToUnicode();
}
}
string unicode = toUnicode.Lookup(new byte[]{(byte) currentNumber}, 0, 1);
if ((unicode != null) && (unicode.Length == 1)) {
this.uni2byte[unicode[0]] = currentNumber;
this.byte2uni[currentNumber] = unicode[0];
this.diffmap[unicode[0]] = currentNumber;
}
}
++currentNumber;
}
}
}
}
}
PdfArray newWidths = font.GetAsArray(PdfName.WIDTHS);
PdfNumber first = font.GetAsNumber(PdfName.FIRSTCHAR);
PdfNumber last = font.GetAsNumber(PdfName.LASTCHAR);
if (BuiltinFonts14.ContainsKey(fontName)) {
BaseFont bf = BaseFont.CreateFont(fontName, WINANSI, false);
int[] e = uni2byte.ToOrderedKeys();
for (int k = 0; k < e.Length; ++k) {
int n = uni2byte[e[k]];
widths[n] = bf.GetRawWidth(n, GlyphList.UnicodeToName(e[k]));
}
if (diffmap != null) { //widths for differences must override existing ones
e = diffmap.ToOrderedKeys();
for (int k = 0; k < e.Length; ++k) {
int n = diffmap[e[k]];
widths[n] = bf.GetRawWidth(n, GlyphList.UnicodeToName(e[k]));
}
diffmap = null;
}
Ascender = bf.GetFontDescriptor(ASCENT, 1000);
CapHeight = bf.GetFontDescriptor(CAPHEIGHT, 1000);
Descender = bf.GetFontDescriptor(DESCENT, 1000);
ItalicAngle = bf.GetFontDescriptor(ITALICANGLE, 1000);
fontWeight = bf.GetFontDescriptor(FONT_WEIGHT, 1000);
llx = bf.GetFontDescriptor(BBOXLLX, 1000);
lly = bf.GetFontDescriptor(BBOXLLY, 1000);
urx = bf.GetFontDescriptor(BBOXURX, 1000);
ury = bf.GetFontDescriptor(BBOXURY, 1000);
}
if (first != null && last != null && newWidths != null) {
int f = first.IntValue;
int nSize = f + newWidths.Size;
if (widths.Length < nSize) {
int[] tmp = new int[nSize];
System.Array.Copy(widths, 0, tmp, 0, f);
widths = tmp;
}
for (int k = 0; k < newWidths.Size; ++k) {
widths[f + k] = newWidths.GetAsNumber(k).IntValue;
}
}
FillFontDesc(font.GetAsDict(PdfName.FONTDESCRIPTOR));
}
private void FillDiffMap(PdfDictionary encDic, CMapToUnicode toUnicode) {
PdfArray diffs = encDic.GetAsArray(PdfName.DIFFERENCES);
if (diffs != null) {
diffmap = new IntHashtable();
int currentNumber = 0;
for (int k = 0; k < diffs.Size; ++k) {
PdfObject obj = diffs[k];
if (obj.IsNumber())
currentNumber = ((PdfNumber)obj).IntValue;
else {
int[] c = GlyphList.NameToUnicode(PdfName.DecodeName(((PdfName)obj).ToString()));
if (c != null && c.Length > 0) {
uni2byte[c[0]] = currentNumber;
byte2uni[currentNumber] = c[0];
diffmap[c[0]] = currentNumber;
}
else {
if (toUnicode == null) {
toUnicode = ProcessToUnicode();
if (toUnicode == null) {
toUnicode = new CMapToUnicode();
}
}
string unicode = toUnicode.Lookup(new byte[]{(byte) currentNumber}, 0, 1);
if ((unicode != null) && (unicode.Length == 1)) {
this.uni2byte[unicode[0]] = currentNumber;
this.byte2uni[currentNumber] = unicode[0];
this.diffmap[unicode[0]] = currentNumber;
}
}
++currentNumber;
}
}
}
}
/**
* Parses the ToUnicode entry, if present, and constructs a CMap for it
* @since 2.1.7
*/
private void ProcessToUnicode(){
PdfObject toUni = PdfReader.GetPdfObjectRelease(fontDic.Get(PdfName.TOUNICODE));
if (toUni is PRStream){
try {
byte[] touni = PdfReader.GetStreamBytes((PRStream)toUni);
CidLocationFromByte lb = new CidLocationFromByte(touni);
toUnicodeCmap = new CMapToUnicode();
CMapParserEx.ParseCid("", toUnicodeCmap, lb);
uni2cid = toUnicodeCmap.CreateReverseMapping();
} catch {
toUnicodeCmap = null;
uni2cid = null;
// technically, we should log this or provide some sort of feedback... but sometimes the cmap will be junk, but it's still possible to get text, so we don't want to throw an exception
//throw new IllegalStateException("Unable to process ToUnicode map - " + e.GetMessage(), e);
}
}
}
public static CMapToUnicode GetIdentity() {
CMapToUnicode uni = new CMapToUnicode();
for (int i = 0; i < 65537; i++) {
uni.AddChar(i, Utilities.ConvertFromUtf32(i));
}
return uni;
}
