With some animated GIFs, for subsequent frames, the only data that is stored, is the difference compared to an earlier frame. Now, I've only just touched the surface of the internal workings of the data structure of (animated) GIFs, but I believe this is done with what is either called compression, frame difference, or disposal (not entirely sure, actually).
With Gmagick, or Imagick, for PHP, how do I extract each frame (write them out) such that not just this minimal data is extracted, but the complete representation of that frame as it is visually intended?
I've tried the following with Gmagick so far, to no avail, unfortunately:
// as per this comment:
// http://php.net/manual/en/imagick.setimagedispose.php#95701
// I've tried values in the range 0-4 here:
$img->setImageDispose( 0 ); // $img is animated GIF instance of Gmagick
$frameIndex = 0;
do
{
$img->setImageIndex( $frameIndex );
$frame = $img->getImage();
$frameName = ++$frameIndex . '.gif';
$frame->writeImage( $frameName, false ); // or:
// $frame->write( $frameName );
}
while( $img->hasNextImage() );
I've tried a few other things as well, such as Gmagick::setImageScene(), etc. But nothing works. I am only able to save the frame difference data, not the data such as visually intended.
I found out on my own how to do it with Imagick:
$frameIndex = 0;
$img = $img->coalesceImages();
foreach( $img as $frame )
{
$frameName = ++$frameIndex . '.gif';
$frame->writeImage( $frameName );
}
In other words, Imagick::coalesceImages() did the trick.
Related
So, I'm testing some of our images with the Imagick PHP library, to see what compression we want to use. But there doesn't seem to be any change to the output file, no matter what I do. This is my basic process:
$original_image = new \Imagick( $image_url );
foreach ( ['original', '92', '80', '60', '40'] as $compression_size )
{
$tester = clone $original_image;
// don't compress the original
if ( 'original' != $compression_size )
{
$tester->setImageCompression(Imagick::COMPRESSION_JPEG);
$tester->setCompressionQuality( (int) $compression_size);
}
$filename: <original base> . "-$compression_size.jpg";
file_put_contents($filename, $tester->getImageBlob() );
$tester = null;
}
The results show that the file sizes between the various compressions don't change, and visually, there is no difference between the original and even the compression = 40 version. What am I doing wrong here?
From the docs (http://php.net/manual/en/imagick.setcompressionquality.php):
This method only works for new images e.g. those created through Imagick::newPseudoImage. For existing images you should use Imagick::setImageCompressionQuality().
So, replace
$tester->setCompressionQuality( (int) $compression_size);
for
$tester->setImageCompressionQuality( (int) $compression_size);
I want download an image from AWS S3 and process it with php. I am using "imagecreatefromjpeg" and "getimagesize" to process my image but it seem that
Storage::disk('s3')->get(imageUrlonS3);
retrieve the image in binary and is giving me errors. This is my code:
function createSlices($imagePath) {
//create transform driver object
$im = imagecreatefromjpeg($imagePath);
$sizeArray = getimagesize($imagePath);
//Set the Image dimensions
$imageWidth = $sizeArray[0];
$imageHeight = $sizeArray[1];
//See how many zoom levels are required for the width and height
$widthLog = ceil(log($imageWidth/256,2));
$heightLog = ceil(log($imageHeight/256,2));
//more code here to slice the image
.
.
.
.
}
// ex: https://s3-us-west-2.amazonaws.com/bucketname/image.jpg
$content = Storage::disk('s3')->get(imageUrlonS3);
createSlices($content);
What am I missing here ?
Thanks
I think you are right in your question what the problem is - the get method returns the source of the image of itself, not the location of the image. When you pass that to createSlices, you're passing the binary data, not its file path. Inside of createSlices you call imagecreatefromjpeg, which expects a file path, not the image itself.
If this indeed the case, you should be able to use createimagefromstring instead of createimagefromjpeg and getimagesizefromstring instead of getimagesize. The functions createimagefromstring and getimagesizefromstring each expects the binary string of the image, which I believe is what you have.
Here's the relevant documentation:
createimagefromstring - http://php.net/manual/en/function.imagecreatefromstring.php
getimagesizefromstring - http://php.net/manual/en/function.getimagesizefromstring.php
Resulting code might look something like this:
function createSlices($imageData) {
$im = imagecreatefromstring($imageData);
$sizeArray = getimagesizefromstring($imageData);
//Everything else can probably be the same
.
.
.
.
}
$contents = Storage::disk('s3')->get($imageUrlOnS3);
createSlices($contents);
Please note I haven't tested this, but I believe from what I can see in your question and what I read in the documentation that this might just do it.
First question, so sorry if I do something wrong!
My issue is, I have a product catalog, which displays up to 18 thumbnails, each approx 6kb, in a product catalog. Each of the thumbnails calls a script get_db_image which searches for and returns the image relating to the product. Simple, so far. The issue only arises when approx 3 or 4 requests are made at the same time for a product catalog page, each user is expecting 18 thumbnails and details to be returned, but when they all do it at the same time I get out of memory errors and sometimes the server crashes. I've stripped down the code that retrieves and displays the image, and the hosting people have raised the memory limit to 256M, but it's all to no avail. As far as I can tell I'm destroying the images I've created and the virtual memory goes back to zero split seconds after the requests are made, but at the peak all the memory is being utilised, hence the crashes, so the only thing I can think off doing is getting, displaying and destroying each image before I start the next one, but I don't know how to go about that, but maybe there is a better solution? Please help, pulling my hair out and I don't have a lot to waste!
// executes the query searching for the image
$res = execPDORetRes($query, $vars);
// if there is no image, load a default
if(sizeof($res) == 0)
{
$query_parts = explode(" ", $query);
$query = "select * from ".$query_parts[3]." where id = :id";
$vars = array(':id-int' => 1);
$res = execPDORetRes($query, $vars);
}
$data = $res[0];
// create the image from the DB
$img = imagecreatefromstring($data[$name]);
$type = "image/pjpeg";
Header( "Content-type: image/pjpeg");
$width = imagesx($img);
$height = imagesy($img);
// if the image is too big
if($size_w != $width || $size_h != $height)
{
// set widths and heights
if ($width <= $size_w)
{
$new_w = $width;
$new_h = $height;
}
else
{
$new_w = $size_w;
$new_h = $size_h;
}
// create a new image of the specified width and height
$new_img = imagecreatetruecolor($new_w,$new_h);
// resize the original
imagecopyresized($new_img,$img,0,0,0,0,$new_w,$new_h,$width,$height);
// determine image type and send it to the client
imagejpeg($new_img,"","80");
// clear the image from memory
imagedestroy($new_img);
imagedestroy($img);
unset($width, $height, $new_h, $new_w, $new_img, $img);
}
else
{
// if the image is smaller than or the right size
// determine image type and send it to the client
imagejpeg($img,"","80");
// clear the image from memory
imagedestroy($img);
unset($width, $height, $img);
}
ob_flush();
Thanks for the help.
I think you should generate your thumbnails in advance if you want to this it this way.
This means you should have your normal product images under e.g. "public/images/products/main" and then your thumbs under "public/images/products/thumbs" and then store in the database your paths to the product images and your thumbs.
This approach is better then creating them on the fly.
Otherwise you could just scale them down with css, if bandwidth is not a problem, but I guess it is.
You can even keep your thumbs script and check in the db if a thumbnail exists, if it does not, call your script: generate thumb, save thumb, update db. The next time a user comes and you need to display this thumbnail, it will already be there.
256 megabytes should be plenty. Nothing in your code looks bad and you're even using unset. I recommend debugging with Xdebug and WinCacheGrind, they can produce in-depth logs...
Loads of answers on how to do it for a command line
convert /path/to/file/file.pdf[3] output.jpg
great... but what if I am using in memory processing, I am generating PDF with PDFlib and then output its buffer to a function that I want to generate jpg preview of selected page. How? My code :
[...]
$buf = $pdf->get_buffer();
//$buff is just a PDF stored in a string now.
$im = new Imagick();
$im->readimageblob($buf);
$im->setImageFormat("jpg");
$im->setimagecompressionquality(60);
$len = strlen($im);
header("Content-type: image/jpeg");
header("Content-Length: $len");
header("Content-Disposition: inline; filename=test.jpg");
echo $im;
This creates a jpeg but always returns last page of the PDF. I want to be able to choose which one will be converted. Is it doable without saving temporary files and using command line (exec('convert /path/to/file/file.pdf[3] output.jpg')) syntax?
Let me add that I tried
$im->readimageblob($buf[2]);
and it did not work :)
For the ones who is still searching for solution of reading specific page number from blob, please check this question Creating array populated by images from a PDF using PHP and ImageMagick
$img_array = array();
$im = new imagick();
$im->setResolution(150,150);
$im->readImageBlob($pdf_in);
$num_pages = $im->getNumberImages();
for($i = 0;$i < $num_pages; $i++)
{
$im->setIteratorIndex($i);
$im->setImageFormat('jpeg');
$img_array[$i] = $im->getImageBlob();
}
$im->destroy();
I'm loading the PDF binary into memory from Amazon S3 and then selecting the specific page I want using setIteratorIndex() followed by getImage()
function get_image_from_pdf($pdf_bytes, $page_num){
$im = new \Imagick();
$im->setResolution(150, 150);
$im->readImageBlob($pdf_bytes);
$im->setIteratorIndex($page_num);
$im = $im->getImage();
$im->setImageFormat('png');
return $im->getImageBlob();
}
version 3.0.1
being on the last image or first image in imagic object
$image_obj = new Imagick("test.pdf"); then you on last image in $image_obj
if you use
fp_pdf = fopen("test.pdf", 'rb');
$image_obj = new Imagick();
$image_obj -> readImageFile($fp_pdf);
then you on the first image in $image_obj
In second case to switch to last image you can do
fp_pdf = fopen("test.pdf", 'rb');
$image_obj = new Imagick();
$image_obj -> readImageFile($fp_pdf,2); // 2 can be any positive number?
then you on the last image in $image_obj
echo $image_obj->getNumberImages() // Returns the number of images in the object
then
if ($image_obj->hadPreviousImage)
$image_obj->previousImage() //Switch to the previous image in the object
}
or
if ($image_obj->hasNextImage()) {
$image_obj->nextImage()) //Switch to the next image in the object
}
e.g. if you have 6 images total and you need 4th then do from the end
$image_obj->previousImage()
$image_obj->previousImage()
$image_obj->setImageFormat("png");
header("Content-Type: image/png");
echo $image_obj;
EDIT: Another find is that you can
foreach($image_obj as $slides) {
echo "<br>".$Obj_img->getImageWidth();
//or wehatever you need to do.
}
EDIT 2: Very simple solution would be to use this function $image_obj->setIteratorIndex(4) count starts with zero.
It's not good news unfortunately, but I can definitively say that, as of time of writing, the ImageMagick (and PHP libraries) don't support the page notation that you're trying to use. (For people from the future finding this: I'm checking php-imagick-3.0.1 and imagemagick-6.6.0.4).
I'm trying to do the exact same thing as you, and I've just spent the last few hours trawling through the source, trying to figure out what it does and how it gets the pages, and it looks like it simply won't use it when reading from a stream (ie. the readBlob() call).
As such, I'm just going to be putting it in a temporary file and reading it from there instead. Not as elegant, but it'll work.
I have a table with 150x150 cells which each have a colored background and a small image/symbol in them. Additionally each cell can have zero or more of it's borders set also. The user can change and add borders and change the cell color and cell. This all works pretty well using jquery and it is just a plain html table.
At the end of the user experience I am making a pdf of this table for them to download. This is a big job and takes a while, which is not my main concern now. jQuery gathers the table data in an array and sends it to php to recreate it as an image using the gd library.
So for each cell I was drawing a rectangle of the correct color on a large image, loading the symbol image and resampling it on to the large image, drawing the borders and imagedestroy the symbol image, it worked but took 1 minute. I changed my strategy to make a small colored rectangke impose the image on it and cache it in an array to be quickly used again. That sped my time up and brought it down to 30ish seconds, but now I am exhausting memory.
I am breaking the table down into 50 cell blocks so each fits on a page, each block is made into an image and saved to disk, the next is made and saved, etc. Each time the gd image is destroyed. Then all the blocks are inserted into the pdf.
So after all that, my question is how do I figure out where the memory is being used so I can try to free it up? I have posted the main function I think is causing the issue below. In my test there are up to 30 different symbols/colors images that are 25pxX25px, these are the images that are cached in an array.
I hope I have provided enough info and my problem is clear enough.
Thank you for your time,
Todd
//make one "block" of stitches returns image file name.
//function makeStitchChartBlock($img, $startX, $startY, $endX, $endY, $caption, $brand){
function makeStitchChartBlock($stitchChartArray, $startX, $startY, $endX, $endY, $caption, $brand,$blockNumber){
global $threadColours;
$stitchCache=array();
$saveTo = 'result'.$blockNumber.'.jpeg';
// calculate size of block
$numRows=($endY-$startY);
$numColumns=($endX-$startX);
$heightOfBlock = $numRows*SYMBOL_SIZE; //in pixels --- used to determine size of image to make for block
$widthOfBlock = $numColumns*SYMBOL_SIZE; //in pixels
//----plus any extra for captions grid lines
$heightOfBlock += (($endY-$startY)+1); //each stitch has a grid line before it and the last one also has on after it
$widthOfBlock += (($endX-$startX)+1);
// create image size of block to put stitches in
$newBlockImage = imagecreatetruecolor($widthOfBlock,$heightOfBlock);
$backStitchColor = imagecolorallocate($newBlockImage, 0, 0, 0);
// insert caption????
//draw grid lines
//$newBlockImage = addGridToImage($newBlockImage);
//keep track of where to start next cell top left
$blockX=0;
$blockY=0;
for($y = $startY; $y < $endY; $y++){ //for each pixel in height, move down 1 "row" each iteration
//echo "<tr>";
for($x = $startX; $x < $endX; $x++){ // "draws" a row (for each y pixel)
//rgb(75, 90, 60)
//x and y are column and row #'s
list($r, $g, $b) = getRGBs($stitchChartArray[$y][$x][0]); //get the rgb values for the cell
$stitchColor = imagecolorallocate($newBlockImage, $r, $g, $b);
//calculate x & y start positons
$stitchStartX=($blockX*SYMBOL_SIZE)+$blockX+1; //account for each previous stitch and the grid line, then add one for new grid line
$stitchStartY=($blockY*SYMBOL_SIZE)+$blockY+1;
$stitchEndX=$stitchStartX+(SYMBOL_SIZE);
$stitchEndY=$stitchStartY+(SYMBOL_SIZE);
/* make a symbol cell image with/without color and save it in the cache */
if(!isset($stitchCache[$r][$g][$b]))
{
//create new image
$stitchCache[$r][$g][$b] = imagecreatetruecolor(SYMBOL_SIZE,SYMBOL_SIZE);
$stitchCacheColor = imagecolorallocate($stitchCache[$r][$g][$b], $r, $g, $b);
//draw colored rectangle
imagefilledrectangle($stitchCache[$r][$g][$b], 0, 0, SYMBOL_SIZE-1, SYMBOL_SIZE-1, $stitchCacheColor);
//add the symbol
$symbolFile=$stitchChartArray[$y][$x][1];
if($symbolFile){
$symbolImage = imagecreatefrompng($symbolFile);
imagecopyresampled ($stitchCache[$r][$g][$b],$symbolImage,0,0,0,0,SYMBOL_SIZE-1,SYMBOL_SIZE-1,imagesx($symbolImage), imagesy($symbolImage) );
imagedestroy($symbolImage);
}
}
//add image from cache to the block image
imagecopyresampled ($newBlockImage,$stitchCache[$r][$g][$b],$stitchStartX, $stitchStartY,0,0,SYMBOL_SIZE,SYMBOL_SIZE,SYMBOL_SIZE,SYMBOL_SIZE);
//add the backstitch lines(borders)
if($stitchChartArray[$y][$x][2]>1) //top
{
imagefilledrectangle($newBlockImage, $stitchStartX, $stitchStartY, $stitchEndX, $stitchStartY+1, $backStitchColor);
}
if($stitchChartArray[$y][$x][3]>1) //right
{
imagefilledrectangle($newBlockImage, $stitchEndX-1, $stitchStartY, $stitchEndX, $stitchEndY, $backStitchColor);
}
if($stitchChartArray[$y][$x][4]>1) //bottom
{
imagefilledrectangle($newBlockImage, $stitchStartX, $stitchEndY-1, $stitchEndX, $stitchEndY, $backStitchColor);
}
if($stitchChartArray[$y][$x][5]>1) //left
{
imagefilledrectangle($newBlockImage, $stitchStartX, $stitchStartY, $stitchStartX+1, $stitchEndY, $backStitchColor);
}
//advance x position
$blockX++;
}
//advance y position
//reset x
$blockX=0;
$blockY++;
}
imagejpeg($newBlockImage, $saveTo);
imagedestroy($newBlockImage);
//dump stitch cache
foreach($stitchCache as $r)
{
foreach($r as $g)
{
foreach($g as $b=>$data)
{
imagedestroy($data);
}
}
}
return $saveTo;
}
I would start (if you haven't already) by getting a good IDE and a debugger as these will be invaluable tools. In this instance you may be able to use a profiler to work out where the memory is being used. Failing that some good ole' manual debugging code, say
$memory = memory_get_usage();
as the first line inside your inner loop. Then when you step through using the debugger you'll be able to see where the memory is ramping up.
btw, using global variables is generally not a good idea. You might want to pass in $threadColours as a parameter or look at other ways of getting that data into the function.