pdfreader.java

处理PDF

            int length = ((PdfNumber)sections.get(idx + 1)).intValue();
            ensureXrefSize((start + length) * 2);
            while (length-- > 0) {
                int type = 1;
                if (wc[0] > 0) {
                    type = 0;
                    for (int k = 0; k < wc[0]; ++k)
                        type = (type << 8) + (b[bptr++] & 0xff);
                }
                int field2 = 0;
                for (int k = 0; k < wc[1]; ++k)
                    field2 = (field2 << 8) + (b[bptr++] & 0xff);
                int field3 = 0;
                for (int k = 0; k < wc[2]; ++k)
                    field3 = (field3 << 8) + (b[bptr++] & 0xff);
                int base = start * 2;
                if (xref[base] == 0 && xref[base + 1] == 0) {
                    switch (type) {
                        case 0:
                            xref[base] = -1;
                            break;
                        case 1:
                            xref[base] = field2;
                            break;
                        case 2:
                            xref[base] = field3;
                            xref[base + 1] = field2;
                            if (partial) {
                                objStmToOffset.put(field2, 0);
                            }
                            else {
                                Integer on = new Integer(field2);
                                IntHashtable seq = (IntHashtable)objStmMark.get(on);
                                if (seq == null) {
                                    seq = new IntHashtable();
                                    seq.put(field3, 1);
                                    objStmMark.put(on, seq);
                                }
                                else
                                    seq.put(field3, 1);
                            }
                            break;
                    }
                }
                ++start;
            }
        }
        thisStream *= 2;
        if (thisStream < xref.length)
            xref[thisStream] = -1;

        if (prev == -1)
            return true;
        return readXRefStream(prev);
    }

    protected void rebuildXref() throws IOException {
        hybridXref = false;
        newXrefType = false;
        tokens.seek(0);
        int xr[][] = new int[1024][];
        int top = 0;
        trailer = null;
        byte line[] = new byte[64];
        for (;;) {
            int pos = tokens.getFilePointer();
            if (!tokens.readLineSegment(line))
                break;
            if (line[0] == 't') {
                if (!PdfEncodings.convertToString(line, null).startsWith("trailer"))
                    continue;
                tokens.seek(pos);
                tokens.nextToken();
                pos = tokens.getFilePointer();
                try {
                    PdfDictionary dic = (PdfDictionary)readPRObject();
                    if (dic.get(PdfName.ROOT) != null)
                        trailer = dic;
                    else
                        tokens.seek(pos);
                }
                catch (Exception e) {
                    tokens.seek(pos);
                }
            }
            else if (line[0] >= '0' && line[0] <= '9') {
                int obj[] = PRTokeniser.checkObjectStart(line);
                if (obj == null)
                    continue;
                int num = obj[0];
                int gen = obj[1];
                if (num >= xr.length) {
                    int newLength = num * 2;
                    int xr2[][] = new int[newLength][];
                    System.arraycopy(xr, 0, xr2, 0, top);
                    xr = xr2;
                }
                if (num >= top)
                    top = num + 1;
                if (xr[num] == null || gen >= xr[num][1]) {
                    obj[0] = pos;
                    xr[num] = obj;
                }
            }
        }
        if (trailer == null)
            throw new IOException("trailer not found.");
        xref = new int[top * 2];
        for (int k = 0; k < top; ++k) {
            int obj[] = xr[k];
            if (obj != null)
                xref[k * 2] = obj[0];
        }
    }

    protected PdfDictionary readDictionary() throws IOException {
        PdfDictionary dic = new PdfDictionary();
        while (true) {
            tokens.nextValidToken();
            if (tokens.getTokenType() == PRTokeniser.TK_END_DIC)
                break;
            if (tokens.getTokenType() != PRTokeniser.TK_NAME)
                tokens.throwError("Dictionary key is not a name.");
            PdfName name = new PdfName(tokens.getStringValue(), false);
            PdfObject obj = readPRObject();
            int type = obj.type();
            if (-type == PRTokeniser.TK_END_DIC)
                tokens.throwError("Unexpected '>>'");
            if (-type == PRTokeniser.TK_END_ARRAY)
                tokens.throwError("Unexpected ']'");
            dic.put(name, obj);
        }
        return dic;
    }

    protected PdfArray readArray() throws IOException {
        PdfArray array = new PdfArray();
        while (true) {
            PdfObject obj = readPRObject();
            int type = obj.type();
            if (-type == PRTokeniser.TK_END_ARRAY)
                break;
            if (-type == PRTokeniser.TK_END_DIC)
                tokens.throwError("Unexpected '>>'");
            array.add(obj);
        }
        return array;
    }

    protected PdfObject readPRObject() throws IOException {
        tokens.nextValidToken();
        int type = tokens.getTokenType();
        switch (type) {
            case PRTokeniser.TK_START_DIC: {
                PdfDictionary dic = readDictionary();
                int pos = tokens.getFilePointer();
                // be careful in the trailer. May not be a "next" token.
                if (tokens.nextToken() && tokens.getStringValue().equals("stream")) {
                    int ch = tokens.read();
                    if (ch != '\n')
                        ch = tokens.read();
                    if (ch != '\n')
                        tokens.backOnePosition(ch);
                    PRStream stream = new PRStream(this, tokens.getFilePointer());
                    stream.putAll(dic);
                    stream.setObjNum(objNum, objGen);
                    return stream;
                }
                else {
                    tokens.seek(pos);
                    return dic;
                }
            }
            case PRTokeniser.TK_START_ARRAY:
                return readArray();
            case PRTokeniser.TK_NUMBER:
                return new PdfNumber(tokens.getStringValue());
            case PRTokeniser.TK_STRING:
                PdfString str = new PdfString(tokens.getStringValue(), null).setHexWriting(tokens.isHexString());
                str.setObjNum(objNum, objGen);
                if (strings != null)
                    strings.add(str);
                return str;
            case PRTokeniser.TK_NAME:
                return new PdfName(tokens.getStringValue(), false);
            case PRTokeniser.TK_REF:
                int num = tokens.getReference();
                PRIndirectReference ref = new PRIndirectReference(this, num, tokens.getGeneration());
                return ref;
            default:
                String sv = tokens.getStringValue();
                if ("null".equals(sv))
                    return PdfNull.PDFNULL;
                else if ("true".equals(sv))
                    return PdfBoolean.PDFTRUE;
                else if ("false".equals(sv))
                    return PdfBoolean.PDFFALSE;
                return new PdfLiteral(-type, tokens.getStringValue());
        }
    }

    /** Decodes a stream that has the FlateDecode filter.
     * @param in the input data
     * @return the decoded data
     */
    public static byte[] FlateDecode(byte in[]) {
        byte b[] = FlateDecode(in, true);
        if (b == null)
            return FlateDecode(in, false);
        return b;
    }

    /**
     * @param in
     * @param dicPar
     * @return a byte array
     */
    public static byte[] decodePredictor(byte in[], PdfObject dicPar) {
        if (dicPar == null || !dicPar.isDictionary())
            return in;
        PdfDictionary dic = (PdfDictionary)dicPar;
        PdfObject obj = getPdfObject(dic.get(PdfName.PREDICTOR));
        if (obj == null || !obj.isNumber())
            return in;
        int predictor = ((PdfNumber)obj).intValue();
        if (predictor < 10)
            return in;
        int width = 1;
        obj = getPdfObject(dic.get(PdfName.COLUMNS));
        if (obj != null && obj.isNumber())
            width = ((PdfNumber)obj).intValue();
        int colors = 1;
        obj = getPdfObject(dic.get(PdfName.COLORS));
        if (obj != null && obj.isNumber())
            colors = ((PdfNumber)obj).intValue();
        int bpc = 8;
        obj = getPdfObject(dic.get(PdfName.BITSPERCOMPONENT));
        if (obj != null && obj.isNumber())
            bpc = ((PdfNumber)obj).intValue();
        DataInputStream dataStream = new DataInputStream(new ByteArrayInputStream(in));
        ByteArrayOutputStream fout = new ByteArrayOutputStream(in.length);
        int bytesPerPixel = colors * bpc / 8;
        int bytesPerRow = (colors*width*bpc + 7)/8;
        byte[] curr = new byte[bytesPerRow];
        byte[] prior = new byte[bytesPerRow];

        // Decode the (sub)image row-by-row
        while (true) {
            // Read the filter type byte and a row of data
            int filter = 0;
            try {
                filter = dataStream.read();
                if (filter < 0) {
                    return fout.toByteArray();
                }
                dataStream.readFully(curr, 0, bytesPerRow);
            } catch (Exception e) {
                return fout.toByteArray();
            }

            switch (filter) {
                case 0: //PNG_FILTER_NONE
                    break;
                case 1: //PNG_FILTER_SUB
                    for (int i = bytesPerPixel; i < bytesPerRow; i++) {
                        curr[i] += curr[i - bytesPerPixel];
                    }
                    break;
                case 2: //PNG_FILTER_UP
                    for (int i = 0; i < bytesPerRow; i++) {
                        curr[i] += prior[i];
                    }
                    break;
                case 3: //PNG_FILTER_AVERAGE
                    for (int i = 0; i < bytesPerPixel; i++) {
                        curr[i] += prior[i] / 2;
                    }
                    for (int i = bytesPerPixel; i < bytesPerRow; i++) {
                        curr[i] += ((curr[i - bytesPerPixel] & 0xff) + (prior[i] & 0xff))/2;
                    }
                    break;
                case 4: //PNG_FILTER_PAETH
                    for (int i = 0; i < bytesPerPixel; i++) {
                        curr[i] += prior[i];
                    }

                    for (int i = bytesPerPixel; i < bytesPerRow; i++) {
                        int a = curr[i - bytesPerPixel] & 0xff;
                        int b = prior[i] & 0xff;
                        int c = prior[i - bytesPerPixel] & 0xff;

                        int p = a + b - c;
                        int pa = Math.abs(p - a);
                        int pb = Math.abs(p - b);
                        int pc = Math.abs(p - c);

                        int ret;

                        if ((pa <= pb) && (pa <= pc)) {
                            ret = a;
                        } else if (pb <= pc) {
                            ret = b;
                        } else {
                            ret = c;
                        }
                        curr[i] += (byte)(ret);
                    }
                    break;
                default:
                    // Error -- unknown filter type
                    throw new RuntimeException("PNG filter unknown.");
            }
            try {
                fout.write(curr);
            }
            catch (IOException ioe) {
                // Never happens
            }

            // Swap curr and prior
            byte[] tmp = prior;
            prior = curr;
            curr = tmp;
        }
    }

    /** A helper to FlateDecode.
     * @param in the input data
     * @param strict <CODE>true</CODE> to read a correct stream. <CODE>false</CODE>
     * to try to read a corrupted stream
     * @return the decoded data
     */
    public static byte[] FlateDecode(byte in[], boolean strict) {
        ByteArrayInputStream stream = new ByteArrayInputStream(in);
        InflaterInputStream zip = new InflaterInputStream(stream);
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        byte b[] = new byte[strict ? 4092 : 1];
        try {
            int n;
            while ((n = zip.read(b)) >= 0) {
                out.write(b, 0, n);
            }
            zip.close();
            out.close();
            return out.toByteArray();
        }
        catch (Exception e) {
            if (strict)