Skip to main content

WooCommerce: Controlling an Asset CDN

Continuing on from my last post, I faced a new issue when it came to adding products and the associated images I was putting in (from Cloudinary) was getting uploaded to the WordPress media library. Not only that, using the URL from my site instead of the CDN it had come from. Double up on all of my images, what a waste - and I want to host from the CDN to keep costs of bandwidth down.  So let me show you how I overcame it.

Separating the herd

What was interesting, is that it was keeping a record of the original source location, and I found I could filter these apart from the rest of my media library:


With this in mind, I wrote a function around it so I could use it to give me a true/false if the given attachment was from this source.


Attaching the hook

Next, needed a way that as soon as an image was added, that it would update the attachment (post) pointing to the correct reference, and not to the file on our server. I found the add_attachment hook, which fires only when an image is added.  So I made the following function which not only updates the item to the external URL, but also deletes the file off our server:

It appears we are moving with steam - but somehow the URL persists to be from our server, only this time the file is not found, be we had deleted it.  So we need to alter the logic WordPress is using to resolve the domain and path.  I found the filter wp_get_attachment_url, which provided this cure:


Final cleanse

Short and sweet. But hang on? Large images are now ok, correctly coming from our CDN, but our thumbnails are not?  Generally, with WordPress generating thumbnails, this is a good thing, but in this case it was not desired.  Here, the system attempts to put together the path of the URL, along with the file it would had generated previously when it was uploaded, but our aim now is to keep that clean and only point to the original URL.

The information for these filename and image properties lives in a serialized wp_attachment_metadata key in the wp_postsmeta table.  With time against me, I've taken the quick and nasty (but still clean) method of renaming all sizes for each thumbnail:


As you can see, this pulled me through quite a re-factor, but all problems were solved.

Preventative cure

There is one more thing needed to be implemented to tie things down.  It appears WooCommerce has a background task of attempting to re-generate thumbnails, likely if they are missing.  Fortunately, I quickly came across this small tip:


And that's it!  I hope this helped any of you facing the same hurdles.
Labels:

Comments

Post a Comment

Popular posts from this blog

question2answer Wordpress Integration

 Today I want to journal my implementation of a WordPress site with the package of "question2answer".  It comes as self-promoted as being able to integrate with WordPress "out of the box".  I'm going to vent a small amount of frustration here, because the only integration going on is the simplicity of configuration with using the same database, along with the user authentication of WordPress.  Otherwise they run as two separate sites/themes. This will not do. So let's get to some context.  I have a new hobby project in mind which requires a open source stack-overflow clone.  Enter question2answer .  Now I don't want to come across as completely ungrateful, this package - while old, ticks all the boxes and looks like it was well maintained, but I need every  page to look the same to have a seamless integration.  So, let's go through this step by step. Forum Index Update This step probably  doesn't need to be done, but I just wanted to mak...
Post a Comment
Read more

Machine Learning: Teaching Wisdom of the Crowd

I got lost in an absolute myriad of thoughts the other day, and it essentially wound up wondering if we can teach machines to count, beyond of what it can see in an image, and I've come up with a small experiment that I would absolutely love to collaborate on if anyone (@ Google ?) else is interested. The idea is based on  the concept of the experiments performed using " Wisdom of the Crowd ", commonly in this experiment to use a jar of jelly beans and asking many people to make a guess as to how many is in there.  Machine learning can be used to make predictions from patterns, but it would have nothing to gain looking at one picture of a jelly bean jar to the next and being able to correctly identify that is in fact - a jar of jelly beans. But suppose we feed it several images of jars of jelly beans, along with all of the guesses people have made of how many is in there.  Can we then presume that feeding it a new image, it would be able to give us a fairly accurate c...
Post a Comment
Read more

Running NodeJS Serverless Locally

 So it's been a long time, but I thought this was a neat little trick so I thought I'd share it with the world - as little followers as I have.  In my spare time I've been writing up a new hobby project in Serverless , and while I do maintain a staging and production environment in AWS, it means I need to do a deployment every time I want to test all of the API's I've drafted for it. Not wanting to disturb the yaml configuration for running it locally, I've come up with a simple outline of a server which continues to use the same configuration.  Take the express driven server I first define here: And then put a index.js  in your routes folder to contain this code: Voila! This will take the request from your localhost and interpret the path against your serverless.yml and run the configured function.  Hope this helps someone!
Post a Comment
Read more