Mime Type Problem when Uploading DOC, DOCX, ODT Files in Symfony

If you create a DOC file with OpenOffice, its mimetype will be application/octet-stream, that can be anything binary data. Of course, you can't allow this mimetype at uploading.

If you try to upload DOCX and ODT files, symfony will recognise them as a ZIP archive. To corrrect this and the OpenOffice DOC mimetype problem, add the 'mime_type_guessers' option to sfValidatorFile:

$this->setValidator('filename', new sfValidatorFile(array(
      'max_size' => ...,
      'path' => ...,
      'mime_type_guessers' => array('guessFromFileinfo'),
      'required' => ...,
      ...

Convert ODT to TEXT with PHP

Shorter way (requires odt2txt installed and shell_exec enabled):

echo shell_exec("odt2txt --encoding=utf8 test.odt");

Longer way (requires PHP 5.2+ and ZIP extension enabled):

function odt2text($filename) {
    return readZippedXML($filename, "content.xml");
}

function readZippedXML($archiveFile, $dataFile) {
    // Create new ZIP archive
    $zip = new ZipArchive;

    // Open received archive file
    if (true === $zip->open($archiveFile)) {
        // If done, search for the data file in the archive
        if (($index = $zip->locateName($dataFile)) !== false) {
            // If found, read it to the string
            $data = $zip->getFromIndex($index);
            // Close archive file
            $zip->close();
            // Load XML from a string
            // Skip errors and warnings
            $xml = DOMDocument::loadXML($data, LIBXML_NOENT | LIBXML_XINCLUDE | LIBXML_NOERROR | LIBXML_NOWARNING);
            // Return data without XML formatting tags
            return strip_tags($xml->saveXML());
        }
        $zip->close();
    }

    // In case of failure return empty string
    return "";
}

echo odt2text("test.odt");