README

Fetch and cache files from local filesystem, cloud storage or public webservers in Laravel.

The file cache is specifically designed for use in concurrent processing with multiple parallel queue workers.

Requirements

• PHP ^8.1
• Laravel ^10.0 || ^11.0 || ^12.0

Installation

composer require jackardios/laravel-file-cache

Laravel

The service provider and FileCache facade are auto-discovered by Laravel.

Publishing Configuration

php artisan vendor:publish --provider="Jackardios\FileCache\FileCacheServiceProvider" --tag="config"

Usage

Take a look at the FileCache contract to see the public API of the file cache. Example:

use FileCache;
use Jackardios\FileCache\GenericFile;

// Implements Jackardios\FileCache\Contracts\File.
$file = new GenericFile('https://example.com/images/image.jpg');

FileCache::get($file, function ($file, $path) {
    // do stuff
});

If the file URL specifies another protocol than http or https (e.g. mydisk://images/image.jpg), the file cache looks for the file in the appropriate storage disk configured at filesystems.disks. You can not use a local file path as URL (e.g. /vol/images/image.jpg). Instead, configure a storage disk with the local driver.

Batch Processing

Process multiple files at once while maintaining locks to prevent pruning during processing:

use FileCache;
use Jackardios\FileCache\GenericFile;

$files = [
    new GenericFile('https://example.com/image1.jpg'),
    new GenericFile('https://example.com/image2.jpg'),
];

FileCache::batch($files, function ($files, $paths) {
    // $paths contains cached file paths in the same order as $files
    foreach ($paths as $path) {
        // process each file
    }
});

One-time Files

Use getOnce() or batchOnce() to automatically delete cached files after processing:

FileCache::getOnce($file, function ($file, $path) {
    // File will be deleted after callback completes
});

Check File Existence

$exists = FileCache::exists($file);

For remote files, exists() uses the HTTP retry settings (http_retries, http_retry_delay). HTTP response statuses outside 2xx are treated as "not exists" (false), including when a custom Guzzle client uses http_errors=true. It may throw network/storage exceptions (for example connection errors or unknown disks).

Configuration

The file cache comes with a sensible default configuration. You can override it in the file-cache namespace or with environment variables.

file-cache.max_file_size

Default: -1 (any size) Environment: FILE_CACHE_MAX_FILE_SIZE

Maximum allowed size of a cached file in bytes. Set to -1 to allow any size.

file-cache.max_age

Default: 60 Environment: FILE_CACHE_MAX_AGE

Maximum age in minutes of a file in the cache. Older files are pruned.

file-cache.max_size

Default: 1E+9 (1 GB) Environment: FILE_CACHE_MAX_SIZE

Maximum size (soft limit) of the file cache in bytes. If the cache exceeds this size, old files are pruned.

file-cache.path

Default: 'storage/framework/cache/files'

Directory to use for the file cache.

file-cache.timeout

Default: -1 (indefinitely) Environment: FILE_CACHE_TIMEOUT

Total connection timeout when reading remote files in seconds. If loading the file takes longer than this, it will fail. Set to -1 to wait indefinitely.

file-cache.connect_timeout

Default: 30 (30 seconds) Environment: FILE_CACHE_CONNECT_TIMEOUT

Timeout to initiate a connection to load a remote file in seconds. If it takes longer, it will fail. Set to -1 to wait indefinitely.

file-cache.read_timeout

Default: 30 (30 seconds) Environment: FILE_CACHE_READ_TIMEOUT

Timeout for reading a stream of a remote file in seconds. If it takes longer, it will fail. Set to -1 to wait indefinitely.

file-cache.prune_interval

Default '*/5 * * * *' (every five minutes)

Interval for the scheduled task to prune the file cache.

file-cache.prune_timeout

Default: 300 (5 minutes) Environment: FILE_CACHE_PRUNE_TIMEOUT

Timeout for the prune operation in seconds. If pruning takes longer than this, it will stop early. Set to -1 for no timeout.

file-cache.mime_types

Default: [] (allow all types)

Array of allowed MIME types for cached files. Caching of files with other types will fail.

file-cache.allowed_hosts

Default: null (all hosts allowed) Environment: FILE_CACHE_ALLOWED_HOSTS

Allowed hosts for remote file fetching. This is a security feature to prevent SSRF (Server-Side Request Forgery) attacks. Set to null to allow all hosts, or provide an array of allowed hostnames.

Wildcards (*) are supported at the beginning of hostnames:

'allowed_hosts' => ['example.com', 'cdn.example.com', '*.trusted-domain.com'],

For environment variable, use comma-separated values:

FILE_CACHE_ALLOWED_HOSTS=example.com,cdn.example.com,*.trusted-domain.com

file-cache.http_retries

Default: 0 (no retries) Environment: FILE_CACHE_HTTP_RETRIES

Number of retry attempts for failed HTTP requests. Client errors (4xx except 429) are not retried. This applies to both get() and exists() for remote files.

file-cache.http_retry_delay

Default: 100 (100ms) Environment: FILE_CACHE_HTTP_RETRY_DELAY

Delay between HTTP retry attempts in milliseconds.

file-cache.lifecycle_lock_timeout

Default: 30 (seconds) Environment: FILE_CACHE_LIFECYCLE_LOCK_TIMEOUT

Timeout to wait for lifecycle lock acquisition in seconds when coordinating batch()/batchOnce() with prune()/clear(). Set to -1 to wait indefinitely.

file-cache.batch_chunk_size

Default: 100 Environment: FILE_CACHE_BATCH_CHUNK_SIZE

Maximum number of files to process in a single chunk during batch() and batchOnce(). Lower values reduce the number of simultaneously opened cached file streams and help avoid file descriptor exhaustion. Set to -1 for no limit.

file-cache.lock_max_attempts

Default: 3 Environment: FILE_CACHE_LOCK_MAX_ATTEMPTS

Maximum number of attempts to acquire a lock on a file. Must be at least 1.

file-cache.lock_wait_timeout

Default: -1 (indefinitely) Environment: FILE_CACHE_LOCK_WAIT_TIMEOUT

Timeout to wait for a lock on a file to be released in seconds. Set to -1 to wait indefinitely.

Clearing

The file cache is cleared when you call php artisan cache:clear.

Testing

The FileCache facade provides a fake for easy testing. The fake does not actually fetch and store any files, but only executes the callback function with a faked file path.

use FileCache;
use Jackardios\FileCache\GenericFile;

FileCache::fake();
$file = new GenericFile('https://example.com/image.jpg');
$path = FileCache::get($file, function ($file, $path) {
    return $path;
});

$this->assertFalse($this->app['files']->exists($path));

Exceptions

The following exceptions may be thrown:

• FileIsTooLargeException - File exceeds configured max_file_size
• FileLockedException - File is locked when throwOnLock is true
• HostNotAllowedException - Host is not in the allowed_hosts whitelist
• MimeTypeIsNotAllowedException - MIME type is not in the mime_types whitelist
• SourceResourceIsInvalidException - Could not establish a valid stream resource
• SourceResourceTimedOutException - Stream read operation timed out
• FailedToRetrieveFileException - General file retrieval failure after all retries
• InvalidConfigurationException - Invalid configuration values

License

MIT