Some jobs and workflows do not fit in to the linear, asynchronous flow of a Blitline job. To help accomodate these we have added a “pre_process” tag which allows you to define functionality that can happen “before” the Blitline job is executed. This “pre_process” tag can contain any of the following “functions” formatted in the standard Blitline syntax.
Before processing a job, Blitline can move the original “src” file to a particular S3 destination.
We have had a bunch of feedback regarding the ability to process an image and then use that image IN a different Blitline job. Usually this occurs when compositing images, and you want to resize image “A” and the composite it on to a resized image “B”. With the current implementation, this takes 2 separate Blitline jobs. By having jobs in a “pre_process” tag, you can assure that 1 job is done before executing another job, thus providing a “synchronous” queue of image processing.
In other words… pre_processed jobs will be executed BEFORE the main Blitline job. You can also reference those pre_processed images in your main Blitline job.
You can watch a small presentation about this here…
In a nutshell, you can just put a regular “job” inside a “pre_process” tag, this job will be executed before the main Blitline job.
Notice, to reference a “pre_process” job output, just place a “&” in fron of the “image_identifier” from the “pre_process” job.
You can use Blitline to ONLY read metadata about a file without actually downloading the whole file or doing any processing on it. This is referred to as “peeking” at an image.
The resulting postback will contain all the metadata about the file contained in the “src” field.
You can search for an image INSIDE another image. This functionality will take a subimage and search for it inside the jobs primary “src”.
In this example, Blitline will look for “boys_subimage.jpg” inside the “boys.jpeg” image. The results will look like:
Where accuracy represents how certain we are about the image being found in that particular location.
NOTE: It will only return 1 result, the one with the highest accuracy.