Question

Can anyone suggest an easy to implement way to extract ToUnicode tables from PDF? I can extract fonts using pdfextract from mupdf, now I’m looking for a way to extract ToUnicode tables for those fonts.

Answer 1

You can modify pdfextract to extract the ToUnicode CMaps (not tables, CMaps).

You might look at the code in savefont and add something like :

obj = fz_dict_gets(dict, "ToUnicode");
if (obj)
{
    stream = obj;
}

If there is a ToUnicode (there need not be) then you could dump the stream in a similar way to the way the font stream is written to file.

obj = fz_dict_gets(dict, "ToUnicode");
if (obj)
{
    stream = obj;
        buf = fz_new_buffer(0);

        error = pdf_load_stream(&buf, xref, fz_to_num(stream), fz_to_gen(stream));
        if (error)
        die(error);
            /* Do something with the data */
    }

buf->data (of size buf->len) would then contain the CMap, which you could write to file, or whatever.