#!/usr/bin/php [desc]', $prog); println('Optional arguments:'); println(' -d turn debug output on.'); println(' desc the new description.'); println('Mandatory arguments:'); println(' input the input file, a JPEG or TIFF image.'); println(' output the output file for the changed image.'); exit(1); } /* Any remaining arguments are considered the new description. */ $description = implode(' ', $argv); /* We typically need lots of RAM to parse TIFF images since they tend * to be big and uncompressed. */ ini_set('memory_limit', '32M'); /* The input file is now read into a PelDataWindow object. At this * point we do not know if the file stores JPEG or TIFF data, so * instead of using one of the loadFile methods on PelJpeg or PelTiff * we store the data in a PelDataWindow. */ println('Reading file "%s".', $input); $data = new PelDataWindow(file_get_contents($input)); /* The static isValid methods in PelJpeg and PelTiff will tell us in * an efficient maner which kind of data we are dealing with. */ if (PelJpeg::isValid($data)) { /* The data was recognized as JPEG data, so we create a new empty * PelJpeg object which will hold it. When we want to save the * image again, we need to know which object to same (using the * getBytes method), so we store $jpeg as $file too. */ $jpeg = $file = new PelJpeg(); /* We then load the data from the PelDataWindow into our PelJpeg * object. No copying of data will be done, the PelJpeg object will * simply remember that it is to ask the PelDataWindow for data when * required. */ $jpeg->load($data); /* The PelJpeg object contains a number of sections, one of which * might be our Exif data. The getExif() method is a convenient way * of getting the right section with a minimum of fuzz. */ $exif = $jpeg->getExif(); if ($exif == null) { /* Ups, there is no APP1 section in the JPEG file. This is where * the Exif data should be. */ println('No APP1 section found, added new.'); /* In this case we simply create a new APP1 section (a PelExif * object) and adds it to the PelJpeg object. */ $exif = new PelExif(); $jpeg->setExif($exif); /* We then create an empty TIFF structure in the APP1 section. */ $tiff = new PelTiff(); $exif->setTiff($tiff); } else { /* Surprice, surprice: Exif data is really just TIFF data! So we * extract the PelTiff object for later use. */ println('Found existing APP1 section.'); $tiff = $exif->getTiff(); } } elseif (PelTiff::isValid($data)) { /* The data was recognized as TIFF data. We prepare a PelTiff * object to hold it, and record in $file that the PelTiff object is * the top-most object (the one on which we will call getBytes). */ $tiff = $file = new PelTiff(); /* Now load the data. */ $tiff->load($data); } else { /* The data was not recognized as either JPEG or TIFF data. * Complain loudly, dump the first 16 bytes, and exit. */ println('Unrecognized image format! The first 16 bytes follow:'); PelConvert::bytesToDump($data->getBytes(0, 16)); exit(1); } /* TIFF data has a tree structure much like a file system. There is a * root IFD (Image File Directory) which contains a number of entries * and maybe a link to the next IFD. The IFDs are chained together * like this, but some of them can also contain what is known as * sub-IFDs. For our purpose we only need the first IFD, for this is * where the image description should be stored. */ $ifd0 = $tiff->getIfd(); if ($ifd0 == null) { /* No IFD in the TIFF data? This probably means that the image * didn't have any Exif information to start with, and so an empty * PelTiff object was inserted by the code above. But this is no * problem, we just create and inserts an empty PelIfd object. */ println('No IFD found, adding new.'); $ifd0 = new PelIfd(PelIfd::IFD0); $tiff->setIfd($ifd0); } /* Each entry in an IFD is identified with a tag. This will load the * ImageDescription entry if it is present. If the IFD does not * contain such an entry, null will be returned. */ $desc = $ifd0->getEntry(PelTag::IMAGE_DESCRIPTION); /* We need to check if the image already had a description stored. */ if ($desc == null) { /* The was no description in the image. */ println('Added new IMAGE_DESCRIPTION entry with "%s".', $description); /* In this case we simply create a new PelEntryAscii object to hold * the description. The constructor for PelEntryAscii needs to know * the tag and contents of the new entry. */ $desc = new PelEntryAscii(PelTag::IMAGE_DESCRIPTION, $description); /* This will insert the newly created entry with the description * into the IFD. */ $ifd0->addEntry($desc); } else { /* An old description was found in the image. */ println('Updating IMAGE_DESCRIPTION entry from "%s" to "%s".', $desc->getValue(), $description); /* The description is simply updated with the new description. */ $desc->setValue($description); } /* At this point the image on disk has not been changed, it is only * the object structure in memory which represent the image which has * been altered. This structure can be converted into a string of * bytes with the getBytes method, and saving this in the output file * completes the script. */ println('Writing file "%s".', $output); file_put_contents($output, $file->getBytes()); /* The End. */ ?>