PHP and antiword are not correctly parsing cyrillic text

Question

I am trying to parse an MS Office 2003 document using antiword on my Linux server. But it will not parse cyrillic text correctly.

It returns something like this:

??? ???? ???????????

Does anybody know of a way to correctly parse a MS Office 2003 document that contains cyrillic text?

Answer 1

I resolved this issue with cyrillic text

good documentation you may see here

working code follows:

$content = shell_exec('/usr/bin/antiword -m cp1251.txt '.$filename);
var_dump($content);

Pay attention to param -m ( character mapping file)

you forgot to set correct mapping file

---

Piece of doc concerns mapping file:

Q9: Which mapping file (-m option) is correct in my situation?
A9: The correct mapping file depends on the character set you need for output
    in a specific language.
    For Western European languages (like English, French, German) this is
    8859-1.txt. (OS/2: cp1252.txt) (DOS: cp850.txt)
    For Eastern European languages (like Polish, Czech, Slovak, Croatian) this
    is 8859-2.txt. (OS/2: cp1250.txt) (DOS: cp852.txt)
    For Esperanto use 8859-3.txt.
    For Russian use 8859-5.txt or koi8-r.txt. (OS/2: cp1251.txt)
     (DOS: cp866.txt)
    For Ukrainian use koi8-u.txt.
    For Arabic use 8859-6.txt. (DOS: cp864.txt)
    For Hebrew use 8859-8.txt. (DOS: cp862.txt)
    For Thai use 8859-11.txt.
    If your system supports it, you might also try UTF-8.txt.

    NOTE: UTF-8 also enables Antiword to show text in languages like Chinese,
          Japanese and Korean.

Answer 2

Antiword has an encoding parameter, maybe you give that a try:

 shell_exec('antiword -X UTF-8 test.doc')

Or use koi8-r , and then convert in php via iconv()

---

Alternatively try LibreOffice in cmdline mode

 shell_exec('soffice --headless --convert-to txt test.doc')