/**
* Have a high offset to simulate a missing packet,
* which should cause it to ignore the ExtendedXMP packet.
*
* @covers XMPReader::parseExtended
*/
public function testExtendedXMPMissingPacket()
{
$xmpPath = __DIR__ . '/../../data/xmp/';
$standardXMP = file_get_contents($xmpPath . 'xmpExt.xmp');
$extendedXMP = file_get_contents($xmpPath . 'xmpExt2.xmp');
$md5sum = '28C74E0AC2D796886759006FBE2E57B7';
// of xmpExt2.xmp
$length = pack('N', strlen($extendedXMP));
$offset = pack('N', 2048);
$extendedPacket = $md5sum . $length . $offset . $extendedXMP;
$reader = new XMPReader();
$reader->parse($standardXMP);
$reader->parseExtended($extendedPacket);
$actual = $reader->getResults();
$expected = array('xmp-exif' => array('DigitalZoomRatio' => '0/10', 'Flash' => 9));
$this->assertEquals($expected, $actual);
}
/** function for gif images.
*
* They don't really have native metadata, so just merges together
* XMP and image comment.
*
* @param string $filename full path to file
* @return Array metadata array
*/
public static function GIF($filename)
{
$meta = new self();
$baseArray = GIFMetadataExtractor::getMetadata($filename);
if (count($baseArray['comment']) > 0) {
$meta->addMetadata(array('GIFFileComment' => $baseArray['comment']), 'native');
}
if ($baseArray['xmp'] !== '' && function_exists('xml_parser_create_ns')) {
$xmp = new XMPReader();
$xmp->parse($baseArray['xmp']);
$xmpRes = $xmp->getResults();
foreach ($xmpRes as $type => $xmpSection) {
$meta->addMetadata($xmpSection, $type);
}
}
unset($baseArray['comment']);
unset($baseArray['xmp']);
$baseArray['metadata'] = $meta->getMetadataArray();
$baseArray['metadata']['_MW_GIF_VERSION'] = GIFMetadataExtractor::VERSION;
return $baseArray;
}
/**
* Test for multi-section, hostile XML
* @covers checkParseSafety
*/
public function testCheckParseSafety()
{
// Test for detection
$xmpPath = __DIR__ . '/../../data/xmp/';
$file = fopen($xmpPath . 'doctype-included.xmp', 'rb');
$valid = false;
$reader = new XMPReader();
do {
$chunk = fread($file, 10);
$valid = $reader->parse($chunk, feof($file));
} while (!feof($file));
$this->assertFalse($valid, 'Check that doctype is detected in fragmented XML');
$this->assertEquals(array(), $reader->getResults(), 'Check that doctype is detected in fragmented XML');
fclose($file);
unset($reader);
// Test for false positives
$file = fopen($xmpPath . 'doctype-not-included.xmp', 'rb');
$valid = false;
$reader = new XMPReader();
do {
$chunk = fread($file, 10);
$valid = $reader->parse($chunk, feof($file));
} while (!feof($file));
$this->assertTrue($valid, 'Check for false-positive detecting doctype in fragmented XML');
$this->assertEquals(array('xmp-exif' => array('DigitalZoomRatio' => '0/10', 'Flash' => '9')), $reader->getResults(), 'Check that doctype is detected in fragmented XML');
}
/**
* Postprocess the metadata (convert xmp into useful form, etc)
*
* This is used to generate the metadata table at the bottom
* of the image description page.
*
* @param $data Array metadata
* @return Array post-processed metadata
*/
protected function postProcessDump(array $data)
{
$meta = new BitmapMetadataHandler();
$items = array();
foreach ($data as $key => $val) {
switch ($key) {
case 'Title':
$items['ObjectName'] = $val;
break;
case 'Subject':
$items['ImageDescription'] = $val;
break;
case 'Keywords':
// Sometimes we have empty keywords. This seems
// to be a product of how pdfinfo deals with keywords
// with spaces in them. Filter such empty keywords
$keyList = array_filter(explode(' ', $val));
if (count($keyList) > 0) {
$items['Keywords'] = $keyList;
}
break;
case 'Author':
$items['Artist'] = $val;
break;
case 'Creator':
// Program used to create file.
// Different from program used to convert to pdf.
$items['Software'] = $val;
break;
case 'Producer':
// Conversion program
$items['pdf-Producer'] = $val;
break;
case 'ModTime':
$timestamp = wfTimestamp(TS_EXIF, $val);
if ($timestamp) {
// 'if' is just paranoia
$items['DateTime'] = $timestamp;
}
break;
case 'CreationTime':
$timestamp = wfTimestamp(TS_EXIF, $val);
if ($timestamp) {
$items['DateTimeDigitized'] = $timestamp;
}
break;
// These last two (version and encryption) I was unsure
// if we should include in the table, since they aren't
// all that useful to editors. I leaned on the side
// of including. However not including if file
// is optimized/linearized since that is really useless
// to an editor.
// These last two (version and encryption) I was unsure
// if we should include in the table, since they aren't
// all that useful to editors. I leaned on the side
// of including. However not including if file
// is optimized/linearized since that is really useless
// to an editor.
case 'PDF version':
$items['pdf-Version'] = $val;
break;
case 'Encrypted':
// @todo: The value isn't i18n-ised. The appropriate
// place to do that is in FormatMetadata.php
// should add a hook a there.
// For reference, if encrypted this fields value looks like:
// "yes (print:yes copy:no change:no addNotes:no)"
$items['pdf-Encrypted'] = $val;
break;
// Note 'pages' and 'Pages' are different keys (!)
// Note 'pages' and 'Pages' are different keys (!)
case 'pages':
// A pdf document can have multiple sized pages in it.
// (However 95% of the time, all pages are the same size)
// get a list of all the unique page sizes in document.
// This doesn't do anything with rotation as of yet,
// mostly because I am unsure of what a good way to
// present that information to the user would be.
$pageSizes = array();
foreach ($val as $page) {
if (isset($page['Page size'])) {
$pageSizes[$page['Page size']] = true;
}
}
$pageSizeArray = array_keys($pageSizes);
if (count($pageSizeArray) > 0) {
$items['pdf-PageSize'] = $pageSizeArray;
}
break;
}
}
$meta->addMetadata($items, 'native');
if (isset($data['xmp']) && function_exists('xml_parser_create_ns')) {
// func exists verifies that the xml extension required for XMPReader
// is present (Almost always is present)
// @todo: This only handles generic xmp properties. Would be improved
// by handling pdf xmp properties (pdf and pdfx) via XMPInfo hook.
$xmp = new XMPReader(LoggerFactory::getInstance('XMP'));
$xmp->parse($data['xmp']);
$xmpRes = $xmp->getResults();
foreach ($xmpRes as $type => $xmpSection) {
$meta->addMetadata($xmpSection, $type);
}
}
unset($data['xmp']);
$data['mergedMetadata'] = $meta->getMetadataArray();
return $data;
}
/** Function to extract metadata segments of interest from jpeg files
* based on GIFMetadataExtractor.
*
* we can almost use getimagesize to do this
* but gis doesn't support having multiple app1 segments
* and those can't extract xmp on files containing both exif and xmp data
*
* @param string $filename Name of jpeg file
* @return array Array of interesting segments.
* @throws MWException If given invalid file.
*/
static function segmentSplitter($filename)
{
$showXMP = XMPReader::isSupported();
$segmentCount = 0;
$segments = array('XMP_ext' => array(), 'COM' => array(), 'PSIR' => array());
if (!$filename) {
throw new MWException("No filename specified for " . __METHOD__);
}
if (!file_exists($filename) || is_dir($filename)) {
throw new MWException("Invalid file {$filename} passed to " . __METHOD__);
}
$fh = fopen($filename, "rb");
if (!$fh) {
throw new MWException("Could not open file {$filename}");
}
$buffer = fread($fh, 2);
if ($buffer !== "ÿØ") {
throw new MWException("Not a jpeg, no SOI");
}
while (!feof($fh)) {
$buffer = fread($fh, 1);
$segmentCount++;
if ($segmentCount > self::MAX_JPEG_SEGMENTS) {
// this is just a sanity check
throw new MWException('Too many jpeg segments. Aborting');
}
if ($buffer !== "ÿ") {
throw new MWException("Error reading jpeg file marker. " . "Expected 0xFF but got " . bin2hex($buffer));
}
$buffer = fread($fh, 1);
while ($buffer === "ÿ" && !feof($fh)) {
// Skip through any 0xFF padding bytes.
$buffer = fread($fh, 1);
}
if ($buffer === "þ") {
// COM section -- file comment
// First see if valid utf-8,
// if not try to convert it to windows-1252.
$com = $oldCom = trim(self::jpegExtractMarker($fh));
UtfNormal::quickIsNFCVerify($com);
// turns $com to valid utf-8.
// thus if no change, its utf-8, otherwise its something else.
if ($com !== $oldCom) {
wfSuppressWarnings();
$com = $oldCom = iconv('windows-1252', 'UTF-8//IGNORE', $oldCom);
wfRestoreWarnings();
}
// Try it again, if its still not a valid string, then probably
// binary junk or some really weird encoding, so don't extract.
UtfNormal::quickIsNFCVerify($com);
if ($com === $oldCom) {
$segments["COM"][] = $oldCom;
} else {
wfDebug(__METHOD__ . " Ignoring JPEG comment as is garbage.\n");
}
} elseif ($buffer === "á") {
// APP1 section (Exif, XMP, and XMP extended)
// only extract if XMP is enabled.
$temp = self::jpegExtractMarker($fh);
// check what type of app segment this is.
if (substr($temp, 0, 29) === "http://ns.adobe.com/xap/1.0/" && $showXMP) {
$segments["XMP"] = substr($temp, 29);
} elseif (substr($temp, 0, 35) === "http://ns.adobe.com/xmp/extension/" && $showXMP) {
$segments["XMP_ext"][] = substr($temp, 35);
} elseif (substr($temp, 0, 29) === "XMP://ns.adobe.com/xap/1.0/" && $showXMP) {
// Some images (especially flickr images) seem to have this.
// I really have no idea what the deal is with them, but
// whatever...
$segments["XMP"] = substr($temp, 29);
wfDebug(__METHOD__ . ' Found XMP section with wrong app identifier ' . "Using anyways.\n");
} elseif (substr($temp, 0, 6) === "Exif") {
// Just need to find out what the byte order is.
// because php's exif plugin sucks...
// This is a II for little Endian, MM for big. Not a unicode BOM.
$byteOrderMarker = substr($temp, 6, 2);
if ($byteOrderMarker === 'MM') {
$segments['byteOrder'] = 'BE';
} elseif ($byteOrderMarker === 'II') {
$segments['byteOrder'] = 'LE';
} else {
wfDebug(__METHOD__ . " Invalid byte ordering?!\n");
}
}
} elseif ($buffer === "í") {
// APP13 - PSIR. IPTC and some photoshop stuff
$temp = self::jpegExtractMarker($fh);
if (substr($temp, 0, 14) === "Photoshop 3.0") {
$segments["PSIR"][] = $temp;
}
} elseif ($buffer === "Ù" || $buffer === "Ú") {
// EOI - end of image or SOS - start of scan. either way we're past any interesting segments
return $segments;
} else {
// segment we don't care about, so skip
$size = wfUnpack("nint", fread($fh, 2), 2);
if ($size['int'] <= 2) {
throw new MWException("invalid marker size in jpeg");
}
fseek($fh, $size['int'] - 2, SEEK_CUR);
}
}
// shouldn't get here.
throw new MWException("Reached end of jpeg file unexpectedly");
}
/** function for gif images.
*
* They don't really have native metadata, so just merges together
* XMP and image comment.
*
* @param string $filename Full path to file
* @return array Metadata array
*/
public static function GIF($filename)
{
$meta = new self();
$baseArray = GIFMetadataExtractor::getMetadata($filename);
if (count($baseArray['comment']) > 0) {
$meta->addMetadata(['GIFFileComment' => $baseArray['comment']], 'native');
}
if ($baseArray['xmp'] !== '' && XMPReader::isSupported()) {
$xmp = new XMPReader(LoggerFactory::getInstance('XMP'));
$xmp->parse($baseArray['xmp']);
$xmpRes = $xmp->getResults();
foreach ($xmpRes as $type => $xmpSection) {
$meta->addMetadata($xmpSection, $type);
}
}
unset($baseArray['comment']);
unset($baseArray['xmp']);
$baseArray['metadata'] = $meta->getMetadataArray();
$baseArray['metadata']['_MW_GIF_VERSION'] = GIFMetadataExtractor::VERSION;
return $baseArray;
}
