ToUnicodeWriter.java

package com.tom_roush.pdfbox.pdmodel.font;

import com.tom_roush.pdfbox.util.Charsets;

import java.io.BufferedWriter;
import java.io.IOException;
import java.io.OutputStream;
import java.io.OutputStreamWriter;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import java.util.TreeMap;

/**
 * Writes ToUnicode Mapping Files.
 *
 * @author John Hewson
 */
final class ToUnicodeWriter
{
    private final Map<Integer, String> cidToUnicode = new TreeMap<Integer, String>();
    private int wMode;

    /**
     * Creates a new ToUnicode CMap writer.
     */
    ToUnicodeWriter()
    {
        this.wMode = 0;
    }

    /**
     * Sets the WMode (writing mode) of this CMap.
     *
     * @param wMode 1 for vertical, 0 for horizontal (default)
     */
    public void setWMode(int wMode)
    {
        this.wMode = wMode;
    }

    /**
     * Adds the given CID to Unicode mapping.
     *
     * @param cid CID
     * @param text Unicode text, up to 512 bytes.
     */
    public void add(int cid, String text)
    {
        if (cid < 0 || cid > 0xFFFF)
        {
            throw new IllegalArgumentException("CID is not valid");
        }

        if (text == null || text.isEmpty())
        {
            throw new IllegalArgumentException("Text is null or empty");
        }

        cidToUnicode.put(cid, text);
    }

    /**
     * Writes the CMap as ASCII to the given output stream.
     *
     * @param out ASCII output stream
     * @throws IOException if the stream could not be written
     */
    public void writeTo(OutputStream out) throws IOException
    {
        BufferedWriter writer = new BufferedWriter(new OutputStreamWriter(out, Charsets.US_ASCII));

        writeLine(writer, "/CIDInit /ProcSet findresource begin");
        writeLine(writer, "12 dict begin\n");

        writeLine(writer, "begincmap");
        writeLine(writer, "/CIDSystemInfo");
        writeLine(writer, "<< /Registry ()");
        writeLine(writer, "/Ordering ()");
        writeLine(writer, "/Supplement ");
        writeLine(writer, ">> def\n");

        writeLine(writer, "/CMapName /Adobe-Identity-UCS" + " def");
        writeLine(writer, "/CMapType 2 def\n"); // 2 = ToUnicode

        if (wMode != 0)
        {
            writeLine(writer, "/WMode /" + wMode + " def");
        }

        // ToUnicode always uses 16-bit CIDs
        writeLine(writer, "1 begincodespacerange");
        writeLine(writer, "<0000> <FFFF>");
        writeLine(writer, "endcodespacerange\n");

        // CID -> Unicode mappings, we use ranges to generate a smaller CMap
        List<Integer> srcFrom = new ArrayList<Integer>();
        List<Integer> srcTo = new ArrayList<Integer>();
        List<String> dstString = new ArrayList<String>();

        int srcPrev = -1;
        String dstPrev = null;
        
        int srcCode1 = -1;

        for (Map.Entry<Integer, String> entry : cidToUnicode.entrySet())
        {
        	int cid = entry.getKey();
        	String text = entry.getValue();

        	if (cid == srcPrev + 1 && // CID must be last CID + 1
        			dstPrev.codePointCount(0, dstPrev.length()) == 1 && // no UTF-16 surrogates
        			text.codePointAt(0) == dstPrev.codePointAt(0) + 1 && // dstString must be prev + 1
        			dstPrev.codePointAt(0) + 1 <= 255 - (cid - srcCode1)) // increment last byte only
        	{
        		// extend range
        		srcTo.set(srcTo.size() - 1, cid);
        	}
        	else
        	{
        		// begin range
        		srcCode1 = cid;
        		srcFrom.add(cid);
        		srcTo.add(cid);
                dstString.add(text);
            }
            srcPrev = cid;
            dstPrev = text;
        }

        // limit of 100 entries per operator
        int batchCount = (int)Math.ceil(srcFrom.size() / 100.0);
        for (int batch = 0; batch < batchCount; batch++)
        {
            int count = batch == batchCount - 1 ? srcFrom.size() % 100 : 100;
            writer.write(count + " beginbfrange\n");
            for (int j = 0; j < count; j++)
            {
                int index = batch * 100 + j;
                writer.write('<');
                writer.write(toHex(srcFrom.get(index)));
                writer.write("> ");

                writer.write('<');
                writer.write(toHex(srcTo.get(index)));
                writer.write("> ");

                writer.write("<");
                writer.write(stringToHex(dstString.get(index)));
                writer.write(">\n");
            }
            writeLine(writer, "endbfrange\n");
        }

        // footer
        writeLine(writer, "endcmap");
        writeLine(writer, "CMapName currentdict /CMap defineresource pop");
        writeLine(writer, "end");
        writeLine(writer, "end");

        writer.flush();
    }

    private void writeLine(BufferedWriter writer, String text) throws IOException
    {
        writer.write(text);
        writer.write('\n');
    }

    private String toHex(int num)
    {
        return String.format("%04X", num);
    }

    private String stringToHex(String text)
    {
        // use of non-BMP code points requires PDF 1.5 or later, otherwise we're limited to UCS-2
        StringBuilder sb = new StringBuilder();
        for (byte b : text.getBytes(Charsets.UTF_16BE))
        {
            sb.append(String.format("%02X", b));
        }
        return sb.toString();
    }
}