Some Japanese characters not shown/extracted properly about pdftotext HOT 6 OPEN

Hello, Far-east languages are a big issue to me ! the sample PDF file you sent internally contains lots of different font maps and also syntactic constructs I was not aware of. Hopefully, you provided me with a short example, which will be easier for me to debug my class. I have put this issue on my todo list ; this will take a little time� I will come back to you when the issue is fixed. With kind regards, Christian.

Hello, Thanks again for sending me this sample ! this issue has been fixed with version 1.4.1. Please feel free to contact me if you have any other issues. With kind regards, Christian.

Dear Sir, I have had a look at your PDF file. Unfortunately, my PdfToText class won�t be of great help for you ! This file uses CID (character ID) fonts. CID fonts were implemented by Adobe long time before the Unicode standard emerged. They describe how to draw a character, but they do not give any information about which character it is. Even Adobe tools ignore what the underlying character is ! To convince yourself, just open the PDF file with Acrobat Reader, select text then paste it to a text editor such as Notepad++ : you will get the same result as my class does, because we have no correspondence between a CID and its Unicode equivalent. I have started an experimental implementation of CID fonts ; for the moment, it handles more or less correctly some languages of eastern Europe. Regarding your PDF file, I suppose it is written in Chinese. Anyway, it uses an Adobe CID font known as �GB 1-5� ; there is a document from Adobe which describes all the characters in this font, giving their corresponding Character ID. But there are two bad news : the first one is that this document from Adobe does not give the Unicode equivalent of each character ID ; the second one is that it defines a little bit more than 30000 characters ! So, I am really sorry to say that I do currently not have a solution to your problem ; maybe in such cases you could use something like OCR ? Please feel free to contact me if you have any question, Christian.