Download the PHP package megaads/laravel-query-cache without Composer

Laravel Eloquent Query Cache

Laravel Eloquent Query Cache brings back the remember() functionality that has been removed from Laravel a long time ago. It adds caching functionalities directly on the Eloquent level, making use of cache within your database queries.

🤝 Supporting

Renoki Co. on GitHub aims on bringing a lot of open source projects and helpful projects to the world. Developing and maintaining projects everyday is a harsh work and tho, we love it.

If you are using your application in your day-to-day job, on presentation demos, hobby projects or even school projects, spread some kind words about our work or sponsor our work. Kind words will touch our chakras and vibe, while the sponsorships will keep the open source projects alive.

🚀 Installation

Hop into your console and install the package via Composer:

Each model you want cache on should use the Rennokki\QueryCache\Traits\QueryCacheable trait.

🙌 Usage

The package has the ability to track the SQL used and use it as a key in the cache storage, making the caching query-by-query a breeze.

In the above example, both queries have different keys in the cache storage, thus it doesn't matter what query we handle. By default, caching is disabled unless you specify a value for $cacheFor. As long as $cacheFor is existent and is greater than 0, all queries will be cached.

It is also possible to enable caching for specific queries by not specifying $cacheFor and calling cacheFor() within your queries:

Cache Tags & Cache Invalidation

Some caching stores accept tags. This is really useful if you plan on tagging your cached queries and invalidate only some of the queries when needed.

Be careful tho - specifying cache tags does not change the behaviour of key storage. For example, the following two queries, altough the use the same tag, they have different keys stored in the caching database.

Global Cache Invalidation

To invalidate all the cache for a specific model, use the flushQueryCache method without passing the tags.

The package automatically appends a list of tags, called base tags on each query coming from a model. It defaults to the full model class name.

In case you want to change the base tags, you can do so in your model.

Full Automatic Invalidation

To speed up the scaffolding of invalidation within your app, you can specify the model to auto-flush the cache upon any time records gets created, updated or deleted.

When you set up the $flushCacheOnUpdate variable, the package attaches an observer to your model, and any created, updated, deleted, forceDeleted or restored event will trigger the cache invalidation.

In order for auto-flush to work, you will need at least one base tag. Out-of-the-box, the model has a base tag set. In some cases, if you have overwritten the getCacheBaseTags() with an empty array, it might not work.

Partial Automatic Invalidation

In some cases, you might not want to invalidate the whole cache of a specific model. Perhaps you got two queries that run individually and want to invalidate the cache only for one of them.

To do this, overwrite your getCacheTagsToInvalidateOnUpdate() method in your model:

Please keep in mind: Setting $flushCacheOnUpdate to true and not specifying individual tags to invalidate will lead to Full Automatic Invalidation since the default tags to invalidate are the base tags and you need at least one tag to invalidate.

Not specifying a tag to invalidate fallbacks to the set of base tags, thus leading to Full Automatic Invalidation.

Relationship Caching

Relationships are just another queries. They can be intercepted and modified before the database is hit with the query. The following example needs the Order model (or the model associated with the orders relationship) to include the QueryCacheable trait.

Cache Keys

The package automatically generate the keys needed to store the data in the cache store. However, prefixing them might be useful if the cache store is used by other applications and/or models and you want to manage the keys better to avoid collisions.

If no prefix is specified, the string leqc is going to be used.

Cache Drivers

By default, the trait uses the default cache driver. If you want to force a specific one, you can do so by calling cacheDriver():

Disable caching

If you enabled caching (either by model variable or by the cacheFor scope), you can also opt to disable it within your query builder chains:

Equivalent Methods and Variables

You can use the methods provided in this documentation query-by-query, or you can set defaults for each one in the model; using the methods query-by-query will overwrite the defaults. While settings defaults is not mandatory (excepting for $cacheFor that will enable caching on all queries), it can be useful to avoid using the chained methods on each query.

Advanced

Implement the caching method to your own Builder class

Since this package modifies the newBaseQueryBuilder() in the model, having multiple traits that modify this function will lead to an overlap.

This can happen in case you are creating your own Builder class for another database drivers or simply to ease out your app query builder for more flexibility.

To solve this, all you have to do is to add the \Rennokki\QueryCache\Traits\QueryCacheModule trait and the \Rennokki\QueryCache\Contracts\QueryCacheModuleInterface interface to your Builder class. Make sure that the model will no longer use the original QueryCacheable trait.

Generating your own key

This is how the default key generation function looks like:

In some cases, like implementing your own Builder for MongoDB for example, you might not want to use the toSql() and use your own method of generating per-sql key. You can do so by overwriting the MyCustomBuilder class generatePlainCacheKey() with your own one.

It is, however, highly recommended to use the most of the variables provided by the function to avoid cache overlapping issues.

Implementing cache for other functions than get()

Since all of the Laravel Eloquent functions are based on it, the builder that comes with this package replaces only the get() one:

In case that you want to cache your own methods from your custom builder or, for instance, your count() method doesn't rely on get(), you can replace it using this syntax:

In fact, you can also replace any eloquent method within your builder if you use $this->shouldAvoidCache() check and retrieve the cached data using getFromQueryCache() method, passing the method name as string, and, optionally, an array of columns that defaults to ['*'].

Notice that the getFromQueryCache() method accepts a method name and a $columns parameter. If your method doesn't implement the $columns, don't pass it.

Note that some functions like getQueryCacheCallback() may come with an $id parameter. The default behaviour of the package doesn't use it, since the query builder uses ->get() by default that accepts only columns.

However, if your builder replaces functions like find(), $id is needed and you will also have to replace the getQueryCacheCallback() like so:

🐛 Testing

🤝 Contributing

Please see CONTRIBUTING for details.

🔒 Security

If you discover any security related issues, please email [email protected] instead of using the issue tracker.

🎉 Credits

• Alex Renoki
• All Contributors